Economic Research-Ekonomska Istraživanja

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/rero20

Feature selection in credit risk modeling: an

international evidence

Ying Zhou, Mohammad Shamsu Uddin, Tabassum Habib, Guotai Chi &
Kunpeng Yuan

To cite this article: Ying Zhou, Mohammad Shamsu Uddin, Tabassum Habib, Guotai
Chi & Kunpeng Yuan (2021) Feature selection in credit risk modeling: an international
evidence, Economic Research-Ekonomska Istraživanja, 34:1, 3064-3091, DOI:
10.1080/1331677X.2020.1867213

To link to this article: https://doi.org/10.1080/1331677X.2020.1867213

© 2021 The Author(s). Published by Informa

UK Limited, trading as Taylor & Francis
Group.

Published online: 17 Jan 2021.

Submit your article to this journal

Article views: 3159

View related articles

View Crossmark data

Citing articles: 2 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=rero20
ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA
2021, VOL. 34, NO. 1, 3064–3091
https://doi.org/10.1080/1331677X.2020.1867213

Feature selection in credit risk modeling:

an international evidence
Ying Zhoua, Mohammad Shamsu Uddina,b, Tabassum Habiba, Guotai Chia and
Kunpeng Yuana
a
School of Economics and Management, Dalian University of Technology, Dalian, China; bSchool of
Business and Economics, Metropolitan University, Bateshwar, Sylhet, Bangladesh

ABSTRACT ARTICLE HISTORY

This paper aims to discover a suitable combination of contempor- Received 21 May 2020
ary feature selection techniques and robust prediction classifiers. Accepted 16 December 2020
As such, to examine the impact of the feature selection method
KEYWORDS
on classifier performance, we use two Chinese and three other
Credit risk; feature selection;
real-world credit scoring datasets. The utilized feature selection least absolute shrinkage
methods are the least absolute shrinkage and selection operator and selection operator;
(LASSO), multivariate adaptive regression splines (MARS). In con- support vector machines
trast, the examined classifiers are the classification and regression
trees (CART), logistic regression (LR), artificial neural network JEL CODES
(ANN), and support vector machines (SVM). Empirical findings C45; C55; C88; G33
confirm that LASSO’s feature selection method, followed by
robust classifier SVM, demonstrates remarkable improvement and
outperforms other competitive classifiers. Moreover, ANN also
offers improved accuracy with feature selection methods; LR only
can improve classification efficiency through performing feature
selection via LASSO. Nonetheless, CART does not provide any
indication of improvement in any combination. The proposed
credit scoring modeling strategy may use to develop policy, pro-
gressive ideas, operational guidelines for effective credit risk man-
agement of lending, and other financial institutions. The finding
of this study has practical value, as to date, there is no consensus
about the combination of feature selection method and predic-
tion classifiers.

1. Introduction
The term credit or lending and borrowing system is old as human civilization
(Thomas et al., 2002). Therefore it has a long history connected with the history of
trade and commerce. Even though it has a very long history of credit, credit scoring
is not as long as credit. It is assumed that the credit or lending and borrowing system
has started 2000 BC or before, whereas the starting period of credit scoring is about

CONTACT Guotai Chi [email protected] School of Economics and Management, Dalian University of
Technology, Dalian, 116024, China.
ß 2021 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/
licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3065

six decades ago. In the early period, credit scores of potential customers were pre-
pared via lending institutions from their respective applications (Hand & Jacka, 1998;
Lewis, 1992; Thomas et al., 2002). Afterward, the area of credit scoring has been
extended in diverse sectors with more new applications. At the beginning of the 21st
century, the application of credit scoring had been developed more than before. This
remarkable technological development was mainly announcing advanced, sophisti-
cated approaches like artificial intelligent methods and prediction measures, for
instance, GINI and area under the ROC (receiver operating characteristic) curve. In
addition, the massive computational capacity of related technologies makes credit risk
modeling considerably easier and efficient in compared to the earlier period (Chang
et al., 2018; Chi et al., 2019a,b; Jiang & Jones, 2018; Jones et al., 2015, 2017; Jones &
Wang, 2019; Uddin et al., 2020a,b).
Besides, additional and irrelevant features may create computational difficulties in
the credit data modeling process and require extra effort and cost to deal with this
issue. To solve this problem, in pattern recognition and data mining, feature selection
playing a significant role in identifying optimal feature sets, reducing data dimension-
ality, and modeling complexity. It selects a subset of only significant predictors for
use in the model evaluation, enhances reliability, enlarges generalization power, and
reduces overfitting. As such, a lot of recent studies developed on different feature
selection approaches (such as Maldonado et al., 2017; L
opez & Maldonado, 2019;
Kozodoi et al., 2019; Arora & Kaur, 2020; Tian et al., 2015; Tian & Yu, 2017; Ala’raj
& Abbod, 2016a). However, there is no consensus for the feature selection technique;
each study applies a different strategy. Besides, the least absolute shrinkage and selec-
tion operator (LASSO) and multivariate adaptive regression splines (MARS) are the
new feature selection methods used in other study fields. Nevertheless, there is no
comprehensive study in the credit scoring literature to identify suitable combinations
of those methods with potent contemporary classifiers like artificial neural networks
(ANN), support vector machines (SVM). Moreover, these methodologies have never
been tested in different study fields, such as small and mid-size enterprises (SME),
agricultural credit, general credit, or other data dimensions, such as balanced, imbal-
anced, high, and low dimensional.
Against this backdrop, this paper employs two robust feature selection methods
LASSO and MARS, with four trendy statistical and machine learning approaches, the
Classification and regression trees (CART), logistic regression (LR), ANN, and SVM
to evaluates the performance of classifiers with feature selection methods. We have
chosen these methods because in data mining, CART, ANN, and SVM are considered
broadly applied and best-supervised machine learning approaches (Lin et al., 2012;
Wu et al., 2008). On the other hand, we also use LR as a conventional statistical
approach, because Jones et al. (2017) stated that most previous credit risk studies
relied on this technique. This model is still the most used method is an industry-
standard (Lessmann et al., 2015).
Two Chinese datasets are used for model training, and three other public datasets
are also utilized for robustness check and validation. The Chinese SME, Chinese agri-
cultural, and German datasets are imbalanced; the Australian and Japanese datasets
are balanced. Therefore, in the beginning, this study used the balancing technique
3066 Y. ZHOU ET AL.

SMOTE (synthetic minority over-sampling technique) to make datasets balanced.

Moreover, we also divided datasets as high dimensional and low dimensional based on
the number of features. We considered high dimensional when the number of features
is more than 30. Otherwise, the dataset is denoted as low dimensional. Thus, this
paper’s first research objective is to evaluate how baseline classifiers perform with dif-
ferent data dimensions. Afterward, the second objective is to determine the impact of
feature selection approaches on classification performance; in other words, which com-
bination can provide the best classification result in credit risk modeling performance.
Finally, this paper intends to evaluate the degree of improvement and whether the pro-
gress is enough to use the new combined method instead of existing approaches.
Our empirical results confirm that a contemporary feature selection method
LASSO followed by robust classifier SVM offers tremendous performance and outper-
forms other competitive classifiers. However, the classification efficiency is not
equivalent for all types of datasets. In the balanced datasets, SVM with MARS accur-
acy is slightly better than the top classifiers SVM combined with feature selection
method LASSO. Besides, the ANN also offers improved accuracy with feature selec-
tion methods, but the CART does not indicate any development. The industry stand-
ard LR can improve classification efficiency when performing feature selection by
LASSO; however, according to average outcomes, the MARS undervalued LR’s base-
line model accuracy.
This paper provides a comprehensive analysis of using four widely used sophis-
ticated statistical and machine learning approaches with two contemporary, robust
feature selection methods. This paper also validates all classifiers with five differ-
ent types of datasets, like balanced, imbalanced, high dimensional, low dimen-
sional, SME, agricultural, and general credit. Therefore the most suitable approach
for each particular field is identified, and these findings are applicable for the
baseline methods of the future study. This study also allows us to know which
classifiers are more sensitive to feature selection methods. These can provide a
direction for potential credit risk research. This study recommends the LASSO
feature selection method with robust machine learning classifier SVM according
to the examined findings. Using LASSO has some benefits, such as LASSO, which
naturally handle multicollinearity problems; it is also considered stable, efficient,
and easily implemented. Conversely, SVM is regarded as the best algorithm for
classification due to its massive computational capacity, easy construction, and
advanced properties compared to other techniques.
Our paper is somewhat similar to existing studies (Tian et al., 2015; Tian & Yu,
2017), as these two studies also used LASSO and for the practical application.
However, these two studies related to time series bankruptcy analysis. Conversely, our
research is designed for credit risk modeling of Chinese SME and agricultural credit
approval datasets; besides, we also used three public datasets for comparison and
proper validation. Moreover, this study also divided datasets into different dimensions
for a specific indication, such as balanced, imbalanced, and high dimensional, low
dimensional. Finally, this study provides a new gesture for academic research and
practical fields with general and specific findings in each area via comprehensive
data modeling.
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3067

The balance of the paper is summarized as follows. Section 2 presents a related lit-
erature review. Section 3 describes data and methods. Section 4 discusses and presents
empirical findings, and section 5 concludes the paper.

2. Literature review
In credit risk modeling, neural networks and support vector machines are considered
robust and mostly used classifiers. The neural network is a numerical method
provoked via the human brain’s procedures, which is significant in problem-solving
systems. Gately (1996) described neural networks as "an artificial intelligence prob-
lem-solving computer program that learns through a training process of trial and
error." Thus, for a better decision-making outcome, neural networks’ structures
involve a training process, and the linear or non-linear variables in the training pro-
cess support differentiating predictors. In credit risk modeling, intelligent method
neural networks somewhat different compared to other statistical methods. Such as,
Al Amari (2002) distinguished NN from regression models; he mentioned that the
regression model used the "inverse matrix" to construct applicant’s scores, whereas
"applicant’s profiles" were utilized by neural networks to prepare relative applicants
scores. In addition, during the modeling process of neural networks, the process
altered the findings until the most favorable outcome. In recent times neural network
has emerged as a practical technology, with successful applications in default, bank-
ruptcy, and failure prediction of the bank. Gately (1996) also suggested that a neural
network can be simply used in other financial sectors, such as mortgage submission,
option pricing, and others. Some other researchers (such as Bishop, 1995; Masters,
1995) addressed many different types of neural networks: pattern recognition feed-
forward net’s architecture, the multilayer feed-forward neural networks, and the prob-
abilistic neural networks are mostly used. A small number of credit scoring studies
applied probabilistic neural networks (Master, 1995; Zekic-Susac et al., 2004). In con-
trast, most studies have been utilized multilayer feed-forward networks (Bishop, 1995;
Desai et al., 1996; Dimla & Lister, 2000; Reed & Marks, 1999; Trippi &Turban, 1993,
Chi et al., 2019b; West, 2000).
On the other hand, Cortes and Vapnik (1995) developed SVM, which is consid-
ered popular and mostly used machine learning technology in different real-world
study fields. In credit risk modeling it has widely been utilized due to its advanced
classification ability and comparatively easy construction than its close counterpart
ANN and other classifiers (Bao et al., 2019; Danenas & Garsva, 2015). The SVM is
based on statistical learning theory. In contrast, the other traditional algorithm (such
as NN) uses the empirical risk minimization (ERM) principle to minimize the
sample’s error, and it creates over-fitting. On the other hand, statistical learning the-
ory implements structural risk minimization (SRM) theory, which reduces the upper
bound of the classifier’s generational error and the sample’s error. These processes
improve the classifier’s generalization via minimizing structural risk. Due to its
advanced properties, compared to other techniques, SVM considers the best algo-
rithm for classification and regression (Ping & Yongheng, 2011). As such, a lot of
recent studies also developed models based on robust classifier SVM (Al-Hadeethi
3068 Y. ZHOU ET AL.

et al., 2020; Jalal et al., 2020; Jiang et al., 2018; Kouziokas, 2020; Luo et al., 2020; Yu
et al., 2018; Zheng et al., 2020). This paper also used industry-standard LR (Ohlson,
1980) and CART (Breiman et al., 1984) to compare with the above-mentioned trendy
classifiers.
Besides, the original credit datasets may have numerous features; however, all fea-
tures are not equally important. The additional features are reliable for extreme
dimensionality as well as occupied the feature location. There are some benefits; how-
ever, it creates severe difficulties (Yu & Liu, 2003). In the practical field, it is compli-
cated to run classifiers, and hence it is not capable of confine diverse affiliation of
distinctiveness because this dataset might be different in dimension, characteristics,
and inherent values (Ala’raj & Abbod, 2016a). Practically, the high dimensional data-
sets required maximum training time, but it produces minimum accuracy (Liu &
Schumann, 2005).
Moreover, distinct unrelated and redundant characteristics in high-dimensional data
are not beneficial to classification results but meaningfully increase the computational
difficulties (Hu et al., 2018). To solve the challenges mentioned above, a lot of contem-
porary studies developed on feature selection approaches in different study fields (such
as Maldonado et al., 2017; L
opez & Maldonado, 2019; Kozodoi et al., 2019; Arora &
Kaur, 2020; Tian et al., 2015; Tian & Yu, 2017; Ala’raj & Abbod, 2016a). As such, this
paper makes feature selection as a pre-processing step for choosing the most influential
variables and, therefore removing the redundant features. This paper employed the least
absolute shrinking and selection operator (LASSO), multivariate adaptive regression
splines (MARS) methods for feature selection. We utilized these models because recent
findings confirm that these two methods are efficient and can provide superior selec-
tion outcomes compared to other related techniques (Ala’raj & Abbod, 2016a; Tian
et al., 2015; Tian & Yu, 2017). This study selected three intelligent methods because of
their respective efficiency, classification superiority, and application in the previous lit-
erature of credit scoring. Finally, the results of all three approaches are compared with
industry-standard statistical method LR.
The related credit scoring studies used different types of modeling strategies, such
as Huang et al. (2004), applied SVM and multilayer perceptron (MLP) as a bench-
mark model and revealed that these two approaches consistently perform better than
industry standard LR. In other perspectives, Huang et al. (2007) recommend that the
SVM classified default and non-default customers accurately. Invariably, some other
studies (such as Min & Lee, 2005; Kim & Ahn, 2012; Shin et al., 2005) applied SVM
on the Korean dataset, and few studies (Ding et al., 2008; Xie et al., 2011) exploited
on Chinese listed company dataset. Nevertheless, both groups of studies provided
similar ending about SVM’s superiority over other counterparts, such as DA, LR, and
NN. Furthermore, Boyacioglu et al. (2009) mentioned that SVM and NN outper-
formed some other multivariate statistical approaches in the bank credit failure pre-
diction. In addition, Zhong et al. (2014) and Wang et al. (2018) applied SVM, NN,
and other technologies in rating distribution. They concluded that SVM is better in
rating distribution, whereas NN is better than SVM on reliability.
Credit risk is crucial for financial organization; additional and irrelevant features
may create computational difficulties and require extra effort and cost. However, the
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3069

Table 1. Description of databases used in the experiment.

Databases Total cases Nondefault/default cases No. of attributes
Australian credit 690 307/383 14
Chinese SME credit 3111 3040/71 81
Chinese agricultural credit 2036 2012/24 44
Japanese credit 690 307/383 15
German credit 1000 700/300 20
Source: Authors’ own calculations.

studies mentioned above did not use feature selection approaches to identify signifi-
cant variables for model training and minimize error and cost. Therefore low model-
ing accuracy was reported without sufficient explanation. Also, there is no consensus
in the existing literature for the feature selection technique; each study applies a dif-
ferent strategy. Besides, LASSO and MARS are the new feature selection methods
used in other study fields. Still, there is no comprehensive study in the credit scoring
literature to identify suitable combinations of those methods with potent contempor-
ary classifiers like ANN and SVM.

3. Data and methods

3.1. Description of real-world databases
This paper utilizes five credit approval datasets from the real-world credit field to ver-
ify the efficacy and viability of the recommended credit scoring models. Two of our
projects, high dimensional databases, are related to small and mid-size enterprises
(SME), and agricultural credit was gathered from one of the top public Chinese com-
mercial bank. These two historical data samples hold financial, nonfinancial, and
macroeconomic variables from 28 major cities. In addition, three open datasets,
namely, Australian, Japanese, and German credit data are available at the University
of California, Irvine (UCI) Machine Learning Repository; however, the processed
datasets are gathered from (Chi et al., 2017). A review of the five above mentioned
databases is given in Table 1.
Any training set with a jagged allocation between the two classes can be considered
imbalanced in the standard. Though, sample ratios 1:5 (minority samples: majority
samples) or upper has generally been judged in the trail as insignificant datasets (He
& Garcia, 2009). In our study, Australian and Japanese datasets are not imbalanced,
German credit datasets are slightly imbalanced, and Chines SME and agricultural
credit are mostly imbalanced. Therefore, in the beginning, this study employed the
balancing technique SMOTE (synthetic minority over-sampling technique) to make
datasets balanced. The experimental databases, consequently, are an excellent combin-
ation of balanced and imbalanced examples.

3.2. Data randomization

In the modeling, to reduce data redundancy and develop data reliability, it is signifi-
cant to normalize data before construction and to train the models. It is better to
transform data from different forms to common forms to escape bias and train the
3070 Y. ZHOU ET AL.

classifiers via uniform data instances. Some models like SVM and NN required input
instances that scale from 0 to 1 and in a vector of a real number. Therefore, in this
study, the min-max normalization method is used for data normalization. The
new converted highest value is 1 (max_new), and the transformed lowest attribute is
given a value of 0 (min_new). The transformation has done based on the following
equation:

New value ¼ ðoriginal  minÞ=ðmax  minÞ

3.3. Data balancing

In the real-world business credit data, it is frequently found those maximum bor-
rowers are non-default with a small percentage of default. Consequently, imbalanced
ratios harm a model’s evaluation performance. To solve this issue, synthetically
under-sampling the majority class can boost modeling accuracy. Nevertheless, in this
process, some virtual instances may be lost, and model over-fitting might happen.
To get possibly improved modeling efficiency without losing data, SMOTE (syn-
thetic minority over-sampling technique) facilitates a data-miner to oversample the
minority class (Chawla et al., 2002, Bifet et al., 2010). SMOTE produces new instances
by working within the current feature space to avoid overfitting problems while
expanding minority class regions. A new sample brings significance to the underlying
data set as values are derived from interpolation rather than extrapolation. For every
minority class instance, using a k-nearest neighbor technique SMOTE interpolates
values and creates attribute values for new data instances (Drown et al., 2009). For
each minority data, a novel synthetic data instance (I) is made by taking the distinc-
tion among the feature vector of I and its nearest neighbor (J) belonging to the equal
group, multiply it by a random number between 0 and 1, and then adding it to I.
This process generates a random line segment among all fair of existing features from
instances I and J, resulting in forming a new sample contained by the dataset (Bifet
et al., 2010). The procedure is recurring for the other k-1 neighbors of the minority
instance I. As a result, SMOTE makes more general regions from the minority class,
and decision tree classifiers can use for the data treatment for better generalizations.

3.4. Feature selection

Feature selection is an essential step to construct models proficiently. To enhance reliabil-
ity, enlarge generalization power and reduce overfitting, it is the procedure of choosing a
subset of only significant predictors for use in the model evaluation. Several methods
have been applied for feature selection. However, this study used the following two
robust feature selection techniques to provides more clean datasets to classifiers.

3.4.1. Least absolute shrinkage and selection operator (LASSO)

Previous studies (e.g., Altman, 1968; Beaver, 1966; Beaver et al., 2005; Campbell et al.,
2008; Chava & Jarrow, 2004; Shumway, 2001) have presented various accounting
ratios and market-related variables for the accurate modeling of credit and
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3071

bankruptcy data. However, there is no standard variable or consensus regarding pre-

dictors to improve model accuracy. Variable selection is considered an essential and
crucial part of the statistical literature to discover significant predictor variables and
improve prediction accuracy. Contemporary development of feature selection litera-
ture shows the potential role of penalized shrinkage methods (Meier et al., 2008;
Tibshirani, 1996; Zou, 2006); these approaches select relevant variables via shrunken
coefficients under a pre-specified roughness penalty. This study applied the original
method of least absolute shrinkage and selection operator (LASSO), recommend by
Tibshirani (1996), and contained by the shrinkage framework to pick a parsimonious
set of the default set of default predictors. Amendola et al. (2011) have introduced
LASSO in the binary problem for the Italian data accounting variables. The LASSO
can be presented by reducing the negative log-likelihood function subject to weighted
constraints,
n 
X     
0 0
Yi, tþ12 b0 þ b Xi, t þ log 1 þ exp b0 þ b Xi , t (1)
i¼1

Pp
subject to k¼1 j bk j  s, in the equation, n represents the number of instances, and p
is the quantity of the predictors used in the respective model. The frequency of
shrinkage can be control by the roughness penalty tuning parameter of s. It should
be mentioned that the lower value of s generally indicates a more parsimonious set of
the selected predictive variables.
The different penalizing coefficient with various weights in LASSO facilitates some
desirable features. LASSO chooses predictors via zeroing some coefficients and
shrinking others. The subset selection provides more interpretability than other mod-
els, and in the ridge regression, it also offers stability. LASSO assists in selecting sig-
nificant predictors; selected variables can indicate the determinants of default. Some
variables, especially some accounting variables, correlate with each other. LASSO
automatically handle multicollinearity problem among the predictors. The efficiency
of the shrinkage approach to solving multicollinearity problems, for example, ridge
regression, is also mentioned in the earlier literature (Mahajan et al., 1977; Mason
et al., 1991; Vinod, 1978). In addition, the LASSO is also mentionable for its compu-
tational efficiency (Efron et al., 2004).
The general user-friendly conventional feature selection method is the best-subset
because of its interpretability. However, it has some limitations, such as any little
modification in data that may reduce classification accuracy (Breiman, 1995, 1996;
Tibshirani, 1996; Zou, 2006). It also assumes that the subset is not a feasible solution
in the corporate bankruptcy or default prediction problem (Tian et al., 2015). In the
practical field, the most used method is a stepwise subset, but it does not consider
stochastic errors in the variable selection process (Fan & Li, 2001). Because of its
heuristic algorithm, it may yield a local best solution rather than a global solution.

3.4.2. Multivariate adaptive regression splines (MARS)

Jerome H. Friedman develops multivariate adaptive regression splines (MARS) in
1991, which use for non–parametric and non-linear regression analysis. MARS can
3072 Y. ZHOU ET AL.

automatically perform the variable selection and other activities, such as variable
transformation, interaction detection, and self-testing at high speed. The structure of
MARS can be discussed in the following way:

X
k
y ¼ c0 þ ci Bi ðxÞ (2)
i¼1

where co is a constant coefficient, Bi ðXÞ is the basis function, and ci is a coeffi-

cient of the basis function. In basis function, it gets various forms of independent
variable’s connections; the familiar functions used are the hinge functions that are
used to discover variables, which are chosen as knots. Therefore the function obtains
the following form (Friedman, 1991).

max ð0, X  cÞ (3)

or,

max ð0, c –XÞ (4)

where c is a constant, threshold, or knot location, X represents the independent varia-

bles. The purpose of the basis function is to convert the independent variables X into
new variables (e.g., X’). Consistent with the Eq. (3) and (4), X’ will take the value of
X if X is larger than c, and it will take the value of zero if the value of X is less than
c (Briand et al., 2004). For additional discussion about the MARS model, please refer
to (Friedman, 1991 and Hastie et al., 2005).

3.5. Baseline classifiers

Four well-known classifiers, namely logistic regression (LR), classification and regres-
sion trees (CART), artificial neural network (ANN), and support vector machines
(SVM), have been utilized to modeling credit approval data. The LR and CART
methods are too popular to be described here. Therefore, this study only discusses
the robust classifiers ANN and SVM are as follows:

3.5.1. Artificial neural network (ANN)

The structure of an artificial neural network (ANN) is developed based on the bio-
logical neural frameworks. ANN can effectively be applied in credit scoring and sev-
eral other study fields, such as data classification and clustering, time series
prediction, pattern recognition, signal processing, etc. ANN’s training process is
related to altering relations among the neurons. Feedforward neural network (FFNN)
and feedback neural network (FBNN) are the two types of ANNs that depend on the
system topology. In FFNN, the network’s data stream is unidirectional with no
response loops, whereas, in FBNN, the stream is bi-directional with response loops.
Multilayer perceptron (MLP) belongs to FFNN, and it has three layers; input layer,
hidden layer, and output layer. In the input layer, inputs are received, and the num-
bers of neurons are the same as the dataset’s features. The hidden layer is the
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3073

essential part of MLP used to map and transfer function between the input and out-
put layer. Finally, the output layer is used to provide the outcome of the network.
In the MLP networks, neurons in the layer are completely interrelated by numeric
weights; every neuron holds summation and activation functions. The summation
function summarizes the product of inputs, weights, and bias as exposed in Eq. (5).
Where wij is the relationship weight linking Ii to neuron j, bj is a bias term, and n is
the entire quantity of neuron inputs. Activation functions will receive the output of
the summation function as an input. Typically, the S-shapes curved sigmoid function
is used as the non-linear activation function. The sigmoid function is shown in Eq.
(6). Consequently, the outcome of the neuron j can be described as in Eq. (7).

X
n
Sj ¼ wij Ii þ bj (5)
i¼1

1
f ðxÞ ¼ (6)
1 þ ex
!
X
n
yj ¼ fj wij Ii þ bj (7)
i¼1

While the formation of ANN is designed, this approach’s learning procedure is

functional to facilitate the network’s parameters (set of weights). Alternatively, these
weights are curved and modernized to estimate the outcomes and reduced the little
error standard. One of the essential MLP training methods is supervised learning.
The objective of supervised learning is to minimize the error among the expected and
computed outcomes. Backpropagation is regarded as one of the general supervised
learning algorithms based on the Gradient technique. It is a process of discovering
the derivative of ANNs objective function regarding the weights and bias that replace
among layers. This technique is ineffective when the search space is large, and the
method is appropriate only for differentiable objective functions.

3.5.2. Support vector machines (SVM)

Cortes and Vapnik (1995) developed SVM, which is considered popular and mostly
used machine learning technology in different real-world study fields. The credit risk
ground has been widely applied due to its advanced classification ability and com-
paratively easy construction than its close counterpart ANN and other classifiers (Bao
et al., 2019; Danenas & Garsva, 2015). The SVM’s purpose is to minimize the gener-
alization error’s upper bound depending on the structural risk minimization
approach. In the SVM, it is initially necessary to utilize training instances to estimate
a function
Ð for evaluation. The function presented in the following:
: R ! {1,-1}, which are k N-dimensional patterns Xi and class labels Yi, where
N

ðX1 , Y1 Þ, . . . :ðXk , Yk Þ 2 RN X f1,  1g (8)

3074 Y. ZHOU ET AL.

According to Eq (8), the SVM classifier should satisfy the following formulation:

WT u ðXi Þ þ b  1 ifyi ¼ þ 1 (9)

WT u ðXi Þ þ b   1 ifyi ¼ 1 (10)

This is the equivalent of the next equation.

h i
yi wT u ðxi Þ þ b  1, i ¼ 1, 2, . . . :, k (11)

The non-linear function u will map the original space to a high-dimensional fea-
ture space. The hyperplane will be constructed by the mentioned inequalities, which
is defined as

wT u ðxi Þ þ b ¼ 0 (12)

Its main idea is to project the input data into a high-dimensional feature space
and then find a hyperplane supported by the support vectors to separate the two
classes with a maximal margin. Based on the support vector’s feature, the label of the
new input sample can be predicted. Many functions (or called kernels) in SVM can
be chosen to map the input data into the high-dimensional feature space, namely lin-
ear, polynomial, radial basis function (RBF), and sigmoid (Zhou et al., 2010).

3.6. Performance evaluation

Four popular credit scoring evaluation measures are adopted to predict the model’s
accuracy to reach the dependable and hearty ending. These are accuracy, area under
the curve (AUC), type I error, type II error. The evaluation criteria are presented in
the following manner:

Accuracy ¼ ðTP þ TNÞ = ðTP þ FN þ TN þ FPÞ (13)

AUC ¼ ð1=2Þ ðSensitivity þ SpecificityÞ (14)

Type I error ¼ FN = ðTP þ FNÞ (15)

Type II error ¼ FP=ðFP þ TNÞ (16)

TP, TN, FP, and FN are the main components of the confusion matrix are repre-
sent true positive, true negative, false positive, and false negative, respectively.
Generally, accuracy (Eq. 13) is a widespread performance measure that appraises the
model’s general effectiveness by probable results. However, there are some limitations,
like accuracy cannot distinguish default and non-default customers accurately. The
area under the receiver operating characteristic (ROC) curve (Eq. 14) is another
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3075

universal classification accuracy measure in the literature (e.g., Jones, 2017; Jones
et al., 2015; Swets et al., 2000). ROC curves are the alternative suitable option to
evaluate the classifier’s accuracy and are free of any threshold. Two types of perform-
ance measures are widely used to measure the error rates of the classifiers. These are
type I (Eq. 15) and type II error (Eq. 16); when the respected classifier misclassified
non-default customer as default one, it denoted as type I error. On the other hand,
the default customer misclassified as the non-default customer is considered a type II
error. The cost of type II is more than type I error.

3.7. Statistical significance tests

Garc
ıa et al. (2015) recommend that through various classifiers that use unique split-
ting methods to reach a dependable ending, some hypothesis testing should be
applied to determine the best model among the comparative models. Parametric
and non-parametric are two types of statistical tests. However, it is assumed that
the parametric tests hypothetically inappropriate and statistically apprehensive
(Demsar, 2006). Non-parametric tests are also apposite to parametric tests for some
statistical characteristics like the normality of the variance’s data or homogeneity
(Demsar, 2006).
Our paper used twelve classifiers, so the statistical significance test is essential to
discover the top model from comparative classifiers. In this regard, we used a non-
parametric test, known as Friedman’s test, to ranks all the related approaches.
Friedman’s (1940) test evaluates classifier performances separately; therefore, it pro-
vides rank according to classifier performances, like 1 for the best model, 2 for a close
competitor of the best model, and so on. The Friedman statistics x2F is distributed
based on x2 by K-1 degrees of freedom when N (number of databases) and K (num-
ber of approaches). In this paper, the following hypothesis is tested: H0: there is no
difference between classifiers.
In case when the null hypothesis of the Friedman test is rejected, therefore we
utilized a posthoc Bonferroni-Dunn’s test (Dunn, 1961) to discover the specific pair-
wise evaluations that create significant variances and to make the decision for
comparative models (Demsar, 2006; Marqu
es et al., 2012a,b). According to this test,
the outcomes of two or more approaches are considerably different if their normal
position is different by at minimum the critical difference (CD), as follows:

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
kðK þ 1Þ
CD ¼ qa
6N

where qa is measured
pffiffiffi as a standardized assortment value by confidence point,
a/(k-1) divided by 2 . In addition, K¼ figure of approaches contrasted to the best
model and N¼ figure of datasets used in a particular study. Some recent studies
(such as Ala’raj & Abbod, 2016a,b) also used Friedman and Bonferroni-Dunn’s test
to compare the classifiers and determine the best model.
3076 Y. ZHOU ET AL.

4. Empirical results
Tables 2–6 present four classifiers’ evaluation results from the base and two robust
feature selection approaches on five real-world credit scoring datasets. Without any
variable selection, the method is considered as a baseline. The outcomes which
secured top position in all the respective criteria are used in bold font. The significant
statistical rank and significant differences are measured by Friedman’s and
Bonferroni-Dunn’s test. The modeling classifiers are optimized by using the default

Table 2. Performance of classifiers over Australian data.

Name LR CART ANN SVM
Accuracy Base 0.7652 0.8333 0.8362 0.8565
MARS 0.8737(þ14.18) 0.8725 (þ4.70) 0.8768 (þ4.86) 0.8797 (þ2.71)
LASSO 0.8710 (þ13.83) 0.8551 (þ2.62) 0.8710 (þ4.16) 0.8797 (þ2.71)
AUC Base 0.7625 0.8333 0.8340 0.8567
MARS 0.8771(þ15.03) 0.8738 (þ4.86) 0.9361 (þ 12.24) 0.9537 (111.32)
LASSO 0.8741 (þ14.64) 0.8620 (þ 3.44) 0.9378 (þ12.45) 0.9396 (þ9.68)
Type I error Base 0.2575 0.1667 0.1930 0.1994
MARS 0.0912 (–0.17) 0.1140 (–0.05) 0.1384 (–0.05) 0.1619 (–0.08)
LASSO 0.0977 (–2.06) 0.0749 (–55.07) 0.1305 (–32.38) 0.1383 (–30.64)
Type II error Base 0.2174 0.1667 0.1390 0.0872
MARS 0.1545 (–28.93) 0.1384 (–16.98) 0.1042 (–25.04) 0.0684 (–21.56)
LASSO 0.1540 (–29.16) 0.2010 (20.58) 0.1270 (–8.63) 0.0977 (12.04)
Source: Authors’ own calculations.

Table 3. Performance of classifiers over Chinese SME data.

Name LR CART ANN SVM
Base Accuracy 0.8729 0.9895 0.9814 0.9975
MARS 0.8790 (þ0.70) 0.9865 (–0.30) 0.9806(–0.08) 0.9896 (–0.79)
LASSO 0.9868 (þ13.05) 0.9906 (þ0.11) 0.9957 (þ1.46) 0.9983 (10.08)
Base AUC 0.8729 0.9895 0.9814 0.9975
MARS 0.8790 (þ0.70) 0.9865 (–0.30) 0.9937 (þ1.25) 0.9966 (–0.09)
LASSO 0.9868 (þ13.05) 0.9906 (þ0.11) 0.9998 (þ1.87) 0.9999 (10.24)
Base Type I error 0.0437 0.0085 0.0241 0.0026
MARS 0.1453 (þ232.49) 0.0138 (þ62.35) 0.0000 (–100.00) 0.0000(–100)
LASSO 0.0260 (–40.50) 0.0109 (þ28.24) 0.0007(–97.10) 0.0013 (–50)
Base Type II error 0.2105 0.0125 0.0132 0.0024
MARS 0.0967 (–54.06) 0.0132 (þ5.60) 0.0388 (þ193.94) 0.0207 (þ762.50)
LASSO 0.0003(–99.86) 0.0079 (–36.80) 0.0079 (–40.15) 0.0020(–16.67)
Source: Authors’ own calculations.

Table 4. Performance of classifiers over Chinese agricultural data.

Name LR CART ANN SVM
Accuracy Base 0.9569 0.9886 0.9846 0.9960
MARS 0.8074 (–15.62) 0.9744 (–1.44) 0.9679 (–1.70) 0.9945 (–0.15)
LASSO 0.9396 (–1.81) 0.9838 (–0.49) 0.9945 (þ1.01) 0.9995 (10.35)
AUC Base 0.9584 0.9886 0.9846 0.9960
MARS 0.8074 (–15.76) 0.9744 (–1.44) 0.9918 (þ0.73) 0.9999 (10.39)
LASSO 0.9396 (–1.96) 0.9838 (–0.49) 0.9994 (þ1.50) 0.9999 (þ0.39)
Type I error Base 0.0831 0.0134 0.0119 0.0043
MARS 0.2202 (þ164.98) 0.0283 (þ111.19) 0.0601(þ405.04) 0.0099 (þ130.23)
LASSO 0.0954 (þ14.80) 0.0219 (þ63.43) 0.0109 (–8.40) 0.0001 (–97.67)
Type II error Base 0.0000 0.0094 0.0189 0.0037
MARS 0.1650 (þ100) 0.0229 (þ143.62) 0.0040 (–78.84) 0.0001 (–97.30)
LASSO 0.0253(þ2.53) 0.0104 (þ10.64) 0.0000 (–100.00) 0.0000 (–100)
Source: Authors’ own calculations.
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3077

Table 5. Performance of classifiers over German data.

Name LR CART ANN SVM
Accuracy Base 0.7603 0.7796 0.7380 0.7700
MARS 0.7347(–3.37) 0.7753 (–0.55) 0.7332 (–0.65) 0.8417 (þ9.31)
LASSO 0.7689 (þ1.13) 0.7767 (–0.37) 0.7532 (þ2.06) 0.9979 (129.60)
AUC Base 0.7660 0.7796 0.6845 0.7299
MARS 0.7347 (–4.09) 0.7753 (–0.55) 0.7963 (þ16.33) 0.9094 (þ24.59)
LASSO 0.7689 (þ0.38) 0.7768 (–0.36) 0.8286 (þ21.05) 0.9997 (136.96)
Type I error Base 0.2582 0.2183 0.1997 0.1992
MARS 0.2796 (þ8.29) 0.2068 (–5.27) 0.2553 (þ27.84) 0.1969 (–1.15)
LASSO 0.2553 (–1.12) 0.2211 (þ1.28) 0.3281 (þ64.30) 0.0014 (–99.30)
Type II error Base 0.2097 0.2225 0.4312 0.3410
MARS 0.2511 (þ83.51) 0.2425 (þ91.75) 0.2781(–155.05) 0.2825 (þ120.71)
LASSO 0.2068 (–1.38) 0.2254 (þ1.30) 0.1655 (–61.62) 0.0029 (–99.15)
Source: Authors’ own calculations.

Table 6. Performance of classifiers over Japanese data.

Name LR CART ANN SVM
Accuracy Base 0.8623 0.8551 0.8377 0.8551
MARS 0.8536 (–1.01) 0.8551(þ0.00) 0.8623 (þ2.94) 0.8884 (13.89)
LASSO 0.8609 (–0.16) 0.8652 (þ1.18) 0.8710 (þ3.98) 0.8841 (þ3.39)
AUC Base 0.8676 0.8620 0.8360 0.8584
MARS 0.8601(–0.86) 0.8620 (þ0.00) 0.9328 (þ11.58) 0.9584 (111.65)
LASSO 0.8663 (–0.15) 0.8670 (þ0.58) 0.9323 (þ11.52) 0.9406 (þ9.58)
Type I error Base 0.0847 0.0749 0.1782 0.2133
MARS 0.0814 (–3.90) 0.0749 (10.00) 0.0945 (–46.97) 0.0749 (–64.89)
LASSO 0.0847 (þ0) 0.1173 (þ56.61) 0.1075 (–39.67) 0.1140 (–46.55)
Type II error Base 0.1802 0.2010 0.1499 0.0699
MARS 0.1984 (þ10.10) 0.2010 (þ0.00) 0.1723 (þ14.94) 0.1410 (þ101.72)
LASSO 0.1828 (þ1.44) 0.1488 (–25.97) 0.1462 (–2.47) 0.1175(þ68.10)
Source: Authors’ own calculations.

setting of programming software Python 3.5. The parameters of SVM based on

Python 3.5 are: {kernel¼‘Linear’; penalty¼L2 regularization; loss function ¼ squared
hinge loss; loss tolerance ¼ 0.0001; punishment slack coefficient C ¼ 1.0; max iter-
ation ¼ 1000}. The parameters of ANN (also called Multi-Layer Perceptron, MLP)
based on Python 3.5 are: {hidden layer sizes ¼ 100; activation function¼‘relu’, which
is f(x)¼max(0, x); slover¼‘sgd’, which is Stochastic Gradient Descent algorithm;
learning rate ¼ 0.001; momentum ¼ 0.9; L2 regularization parameter ¼ 0.0001; loss
tolerance ¼ 0.0001; max iteration ¼ 200}. The overall results confirm that LASSO
based classifiers outperformed other models. However, to provide a specific scenario,
we present our findings in the following different dimensions.

4.1. Classification outcomes based on credit scoring datasets

Australian credit data: In the Australian dataset (Table 2), all classifiers show con-
sistent improvement with two feature selection approaches; performing feature selec-
tion outperformed baseline classifiers in all four measures. In the accuracy and AUC,
MARS improved the classification accuracy for LR by about 14% to 15%. From the
perspective of type I and type II error LASSO can minimize classification error for
the same classifier about 62% and 29%. However, MARS with SVM offered top scores
in accuracy and AUC. The base classifier SVM performance is better than all other
3078 Y. ZHOU ET AL.

baseline models. In type I error, LASSO is followed by CART, and in type II error,
SVM with MARS produces the minimum error.
Chinese SME credit data: The Chinese SME data classification results (Table 3)
are quite different from the Australian dataset. For the accuracy, feature selec-
tion method MARS undervalued classification performance of baseline models
CART, ANN, and SVM. In the case of AUC, the results are the same for the
CART and SVM. SVM followed by LASSO is outperformed other classifiers in
accuracy and AUC. LASSO based SVM also minimizes type I and II error from
baseline SVM; however, in type I error, MARS-based ANN and SVM offered
minimum results.
Chinese agricultural credit data: In Table 4, the agricultural dataset’s classification
performances revealed that two classifiers, feature selection methods followed by LR
and CART, do not indicate improvement in any case. ANN followed by LASSO,
presents improvement in all the circumstances. However, SVM with LASSO outper-
formed all other classifiers in all the criteria.
German credit data: In the German data in Table 5, the classification outcomes
confirm LASSO’s supremacy, followed by SVM in all the performance measures.
LASSO based LR and ANN also provide better results in maximum criteria. The fea-
ture selection based CART models undervalued its performance in all the criteria
from baseline classifiers.
Japanese credit data: Looking for Japanese data in Table 6, ANN and SVM with
feature selection methods outperform baseline models in accuracy, AUC, and type I
error. The findings are unusual because feature selection improves classification per-
formances in maximum cases; however, in the perspective of classification error, espe-
cially in type II error, SVM baselines outperform all other classifiers.

4.2. Best combination between feature selection methods and prediction

classifiers
Balanced datasets: In our analysis, the Australian and Japanese datasets are balanced,
low dimensional credit scoring datasets. The experimental outcomes confirm that
MARS followed by SVM offers the utmost accuracy in the maximum criteria, whereas
LASSO with SVM also provides competitive results. On the other hand, ANN, with
two feature selection approaches, even present good outcomes.
Imbalanced datasets: In this study, the Chinese SME, Chinese agricultural credit,
and German datasets are imbalanced. The SME and agricultural datasets are high
dimensional, but the German dataset is low dimensional. The classification perform-
ances of imbalanced data sets are quite different from balanced datasets. The empir-
ical findings demonstrate that LASSO, followed by SVM, significantly outperforms
other classifiers. In the agricultural and German datasets, the accuracy is better than
other counterparts, while in the SME dataset, it offers maximum accuracy in three
out of four performance measures. Alternatively, MARS based approaches provide a
minimal indication of improvement over the analysis.
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3079

4.3. The effect of feature selection on average outcomes of classifiers

According to the average results of five credit scoring datasets, the effect of feature
selection on model classification accuracy is discussed in the following.
LR: The industry-standard LR with feature selection results are not consistent with
all the models. The average effect of MARS based LR models does not provide any
indication of improvement. It undervalues the base classifier outcomes, about 1.64%,
1.63% in the accuracy and AUC. It also increases classification error by 12.44% and
5.86% in type I and type II errors. However, LASSO followed by LR classifier presents
significant improvement; it increases classification accuracy by 4.97%, 6.56% in accur-
acy, and AUC. It also minimizes about 23% and 30% in type I and type II error from
the baselines model.
CART: There are no differences between the performances of the baseline and fea-
tures selection based classifiers. The findings are consistent with the Liang et al.
(2015) recommendations; they mentioned that the feature selection technique is
already used in the CART construction process. Therefore, feature selection may not
be supportive of the development of performance for the CART.
ANN: Performing feature selection for ANN is helpful for the improvement of the
classification performances. ANN improves its classification accuracy by 0.97%, 7.64%
with MARS, and 2.45%, 8.73% with LASSO from the perspective of accuracy and
AUC. It minimizes the type I error from baseline models at 9.65%, 4.81%, using
MARS and LASSO. ANN also reduces type II error at 20.58%, 40.63% from the base-
line model using MARS and LASSO. The findings were also supported by a previous
study (Liang et al., 2015).
SVM: Combining a feature selection method with SVM significantly improves
classification efficiency. Mainly, SVM combined with LASSO achieves an average
of 95.19% and 97.59% scores for accuracy and AUC, beating by 6.35% and 9.94%
of baseline SVM. It also reduces classification error by about 58% for type I and
56% for type II error. Our findings are slightly different from previous results in
SVM’s case (Liang et al., 2015). However, they used different feature selection
approaches. Therefore, we can conclude that the LASSO variable selection method
is more efficient than other traditional techniques, such as filter or wrap-
per methods.
In this paper, SVM and ANN are two robust classifiers. Empirical findings demon-
strate that the SVM based model outperforms other models. The SVM model could
obtain the global minimum solution as the SVM objective function can be transferred
into convex quadratic programming to attain the global minimum solution. Different
solution algorithms do not affect the optimal solution but only influence the conver-
gence rate, which means the optimal solution’s time cost.
Conversely, the ANN model’s objective optimization function is non-convex
and non-smooth, so there is a risk of obtaining a local minimum. To minimize
this risk, we conduct the stochastic gradient descent (SGD) algorithm, which has
been proved better than other solving algorithms (like batch gradient descent
algorithm) and could attain almost zero training loss to find a global minimum
(Du et al., 2019).
3080 Y. ZHOU ET AL.

Table 7. The result of statistical significance tests (Friedman and Bonferroni-Dunn’s test).
Accuracy AUC
Hypothesis Hypothesis Hypothesis Hypothesis
Method P-value (a ¼ 0.1) (a ¼ 0.05) P-value (a ¼ 0.1) (a ¼ 0.05)
LR/Base 0.0004 Rejected Rejected 0.0004 Rejected Rejected
LR/MARS 0.0002 Rejected Rejected 0.0004 Rejected Rejected
LR/LASSO 0.0057 Rejected Rejected 0.0044 Rejected Rejected
CART/Base 0.0179 Rejected Rejected 0.0085 Rejected Rejected
CART/MARS 0.0141 Rejected Rejected 0.0038 Rejected Rejected
CART/LASSO 0.0655 Rejected Rejected 0.0201 Rejected Rejected
ANN/Base 0.0005 Rejected Rejected 0.0001 Rejected Rejected
ANN/MARS 0.0044 Rejected Rejected 0.2364 Not Rejected Not Rejected
ANN/LASSO 0.1144 Rejected Rejected 0.5107 Not Rejected Not Rejected
SVM/Base 0.0794 Rejected Not Rejected 0.0075 Rejected Rejected
SVM/MARS 0.6295 Not Rejected Not Rejected 0.8608 Not Rejected Not Rejected
Source: Authors’ own calculations.

Figure 1. Average performances of classifiers on accuracy. Source: Authors’ own calculations.

4.4. Statistical significance test results

According to Garc
ıa et al. (2015), as the different approaches utilized distinct splitting
methods, it is not sufficient to validate one model achieves results more significant to
others. To evaluate performance thoroughly, it would appear appropriate to use a few
hypotheses testing to highlight that the result’s investigational differences are statistic-
ally significant.
This study measured the value of Friedman’s ranks for accuracy and AUC. We
have put only two statistical significance test results here. With a ¼ 0.1and 0.05 sig-
nificance levels, it has shown that LASSO based SVM is the best model according to
accuracy and AUC. The post-hoc Bonferroni-Dunn’s test in Table 7 has expressed
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3081

Figure 2. Average performances of classifiers on AUC. Source: Authors’ own calculations.

Figure 3. Average performances of classifiers on type I error. Source: Authors’ own calculations.

the level of differences with the best model to other models. If the test’s P-value is
lower than 5% or 10%, then the null hypothesis is rejected. Therefore, LASSO fol-
lowed by SVM, is significantly better than corresponding models regarding accuracy.
However, it has no statistically significant difference with SVM/MARS at 10% level
and SVM/MARS and SVM baseline models with a 5% level. In the case of AUC with
a 5% and 10% significant level, the results indicate no significant difference with
ANN/MARS, ANN/LASSO, and SVM/MARS; nonetheless, there is a significant dif-
ference with other remaining models (Figure 5).
3082 Y. ZHOU ET AL.

Figure 4. Average performances of classifiers on type II error. Source: Authors’ own calculations.

Figure 5. Statistical rank of classifiers (Friedman test) on accuracy. Source: Authors’ own calculations.

4.5. Discussion
To achieve the major objectives described in Section 1, Figure 1–4 show the average
performance of classifiers with feature selection methods on five real-world credit
datasets. In addition, statistical significance test results are also presented to show dif-
ferences among the classifiers (Figure 6).
From the perspective of classification accuracy, on average, LASSO feature selec-
tion with SVM has the potential to perform better than other combinations from the
viewpoint of classification accuracy (i.e., 95% for LASSO þ SVM, 92% for
MARS þ SVM, and 90% for baseline SVM (without feature selection). In addition,
LASSO with ANN also slightly improves classification accuracy from the baseline
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3083

Figure 6. Statistical rank of classifiers (Friedman test) on AUC. Source: Authors’ own calculations.

ANN model (i.e.90% for LASSO þ ANN, 88% for baseline ANN). Furthermore,
LASSO feature selection with LR can improve classification accuracy from the base-
line model (i.e.89% for LASSO þ LR, 84% for baseline LR). In the case of MARS with
feature selection methods, there is no improvement from the baseline model. In a 5%
significant level, according to the statistical significance test (Friedman and
Bonferroni-Dunn’s test), there is a significant difference in the classification perform-
ances of other models with LASSO þ SVM, except for MARS þ SVM. Therefore,
LASSO þ SVM would be a better choice than any other combination.
On the other hand, from the viewpoint of classification AUC on average, LASSO
feature selection with SVM significantly outperforms baseline and other combina-
tions (i.e., 98% for LASSO þ SVM, 96% for MARS þ SVM, and 89% for baseline
SVM (without feature selection). Moreover, LASSO with ANN also improves classi-
fication accuracy from the baseline ANN model (i.e.94% for LASSO þ ANN, 86%
for baseline ANN). LASSO feature selection with LR also improves classification
accuracy from the baseline model (i.e.89% for LASSO þ LR, 85% for baseline LR).
In MARS with feature selection methods, there is a minimal improvement from the
baseline model (i.e.90% for LASSO þ CART, 89% for baseline CART). From the
viewpoint of the statistical significance test (Friedman and Bonferroni-Dunn’s test)
at a 5% significant level, there is a significant difference in other models’ classifica-
tion performances with LASSO þ SVM, except MARS þ SVM, LASSO þ ANN, and
MARS þ ANN. Thus, LASSO þ SVM would be a better option compared to other
combinations.
Furthermore, if we consider type I error and type II, the LASSO feature selection
method with SVM also significantly reduces classification error from the baseline
model, and it outperforms other combinations. Therefore, according to four perform-
ance measures, the LASSO þ SVM method is recommended. This is because it offers
comparative classification accuracy; at the same time, it decreases the largest percent-
age of classification error.
3084 Y. ZHOU ET AL.

5. Conclusions
This paper examines the impact of feature selection methods on the classifier’s per-
formances using several real-world credit datasets. More especially, we attempt to
evaluate the sensitivity of the robust classifiers on different data dimensions. This
issue has been studied extensively. However, there is still no consensus about the
combination of feature selection method and prediction classifiers, which may be for
data dimensionality and diversity. This paper aims to recommend a suitable combin-
ation by using robust methods on different datasets in this background.

5.1. Concluding remarks and relationship with previous findings

This paper has three major objectives; the first research objective is to evaluate how
baseline modeling classifiers perform with different data dimensions. The second
objective is to determine the impact of feature selection approaches on classification
performance; alternatively, which combination can provide the best classification
result in credit risk modeling. Finally, this paper intends to evaluate the degree of
improvement and whether the progress is enough to use the new combined method
instead of existing approaches.
Against these backdrops, this paper executes modeling classifiers LR, CART, ANN,
and SVM as a baseline model on the original data samples of all five different data-
sets. Afterward, we shortlisted features of five respective datasets by using two feature
selection approaches, LASSO and MARS. Then, we used four above-mentioned robust
classifiers on selected features. Empirical findings are compared with the baseline
classifiers to evaluate classification accuracy by feature selection methods. The pre-
dictive performance is evaluated against four performance measures: accuracy, AUC,
type I error, and type II error. Finally, the average results of all datasets and statistical
significance tests are carried out for every single model to compare and determine
the most reliable model.
The overall empirical results and statistical ranking confirm that feature selection
methods LASSO, followed by robust classifier SVM demonstrates remarkable
improvement and outperforms other competitive classifiers. This combined top classi-
fier offers 6.35% and 9.94% prediction improvement on the accuracy and AUC per-
formance measures from the base classifiers. However, the classification efficiency is
not equivalent for all types of datasets. In the imbalanced datasets, the recommended
model significantly outperforms other classifiers. On the other hand, in the balanced
datasets, the SVM with MARS accuracy is slightly better than the top classifiers SVM
performing feature selection with LASSO. Besides, among the other classifiers, the
ANN also offers improved accuracy with feature selection methods. Nevertheless, the
CART does not provide any indication of improvement combined with LASSO and
MARS. The industry standard LR can improve classification efficiency when perform-
ing feature selection by LASSO; however, according to average outcomes, the MARS
undervalues the baseline model accuracy of LR.
This paper offers a comprehensive analysis of four widely used trendy statistical
and machine learning approaches with two contemporary and robust feature selection
methods. The appropriate combination of feature selection methods and prediction
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3085

classifiers is recommended after proper validation of classifiers via different types of

datasets (like balanced, imbalanced, high dimensional, low dimensional, SME, agricul-
tural, and general credit). This mechanism can significantly improve prediction per-
formances (such as accuracy and AUC) and minimize classification errors (type I and
type II error). The outcomes of this paper also provide information about the sensi-
tivity of classifiers with feature selection methods.
This paper’s finding differs from existing similar studies (such as Tian et al., 2015;
Tian & Yu, 2017), as this paper developed for Chinese SME and agricultural credit
risk modeling. On the other hand, the above mentioned two papers related to time
series bankruptcy analysis. In addition, for comparison and additional validation, our
paper also used three public datasets. Moreover, for the particular assessment, the
paper’s findings are also divided based on datasets into different dimensions, such as
balanced and imbalanced, as well as high dimensional and low dimensional.
Therefore, this study provides a novel indication for the dynamic financial world and
academic researchers via comprehensive research and general and specific findings on
each application area.

5.2. Implications
Generally, this study has an important implication on the current credit risk litera-
ture. The accuracy of the prediction classifier is significant for lending institutions; it
is also crucial for the country’s overall economic health. As risk modeling is a natural
real-world problem, it is considered that any little improvement can generate huge
earning and minimize significant losses as well. Our analytical findings indicate that
the proposed combination of classifier and feature selection technique can signifi-
cantly improve classification accuracy and reduce prediction error. There is consider-
able potential to develop a sophisticated credit-scoring model by this combination.
Therefore this study has practical significances for financial institutions, management,
employees, investors, and government authority to minimize risk and maximize effi-
ciency in the decision-making process.
Furthermore, this paper has three specific implications. Such as, first in the model-
ing, we used balanced and imbalanced datasets. The findings are discussed in those
dimensions and for high dimensional and low dimensional datasets. Therefore our
conclusions have a specific implication on above mentioned all types of datasets.
Second, we assess the sensitivity of classifiers with feature selection methods on dif-
ferent problem areas. In the future, it would be applicable as baseline information to
use robust prediction classifiers. Moreover, the generability of our modeling approach
is tested on SME credit, agricultural credit, and general credit as well. Therefore the
recommended combined approach can be used in other possible business domains,
such as customer churn prediction or fraud detection.

5.3. Limitations and future research

Our paper’s limitation is that we cannot develop any feature selection or prediction
algorithm; instead, we used some existing methods to find the proper combination of
3086 Y. ZHOU ET AL.

feature selection and classification approaches. However, some other current methods
could determine the appropriate combination of feature selection methods and classi-
fiers. Besides, we have employed five real-world credit datasets to authenticate and
validation of our proposed model. However, due to space constraints, we cannot dis-
cuss the importance of features, such as which feature is essential and significantly
related to default prediction.
In the future, we want to use some newly developed feature selection methods,
such as Butterfly Optimization Algorithm (BOA) (Arora & Anand, 2019), Dynamic
Feature Importance (DFI) (Wei et al., 2020), ensemble feature selection techniques
(Tsai & Sung, 2020). It would be interesting to use the Bayesian quantile regression
and survival curve to model data with novel robust feature selection methods.
Moreover, in future research, we will try to improve the interpretability of modeling
outcomes by providing more information about the model’s success and failure. On
the other hand, recently, some features such as textual data, social media information,
and profit-driven features generate more accuracy in the prediction process. Future
work would also include new features to justify whether they continue to improve
prediction accuracy.

Disclosure statement
The authors reported no potential conflict of interest.

Funding
This work has been supported by the Key Programs of National Natural Science Foundation
of China (the grant number 71731003), the General Programs of National Natural Science
Foundation of China (the grant number 72071026, 71873103, 71971051, and 71971034), the
Youth Programs of National Natural Science Foundation of China (the grant number
71901055, 71903019), the Major Projects of National Social Science Foundation of China
(18ZDA095). The project has also been supported by the Bank of Dalian and Postal Savings
Bank of China. We thank the organizations mentioned above.

References
Al Amari, A. (2002). The credit evaluation process and the role of credit scoring: A case study of
Qatar [Ph.D. Thesis]. University College Dublin.
Ala’raj, M., & Abbod, M. F. (2016a). A new hybrid ensemble credit scoring model based on
classifiers consensus system approach. Expert Systems with Applications, 64, 36–55. https://
doi.org/10.1016/j.eswa.2016.07.017
Ala’raj, M., & Abbod, M. F. (2016b). Classifier consensus system approach for credit scoring.
Knowledge-Based Systems, 104, 89–105. https://doi.org/10.1016/j.knosys.2016.04.013
Al-Hadeethi, H., Abdulla, S., Diykh, M., Deo, R. C., & Green, J. H. (2020). Adaptive boost LS-
SVM classification approach for time-series signal classification in epileptic seizure diagnosis
application. Expert Systems with Applications, 161, 113676. https://doi.org/10.1016/j.eswa.
2020.113676
Altman, E. I. (1968). Financial ratios, discriminant analysis and the prediction of corporate bankruptcy.
The Journal of Finance, 23 (4), 589–609. https://doi.org/10.1111/j.1540-6261.1968.tb00843.x http://
links.jstor.org/sici?sici=00221082%28196809%2923%3A4%3C589%3AFRDAAT%3E2.0.CO%3B2-R.
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3087

Amendola, A., Restaino, M., & Sensini, L. (2011). Variable selection in default risk models.
The Journal of Risk Model Validation, 5 (1), 3–19. https://doi.org/10.21314/JRMV.2011.066
Arora, S., & Anand, P. (2019). Binary butterfly optimization approaches for feature selection.
Expert Systems with Applications, 116, 147–160. https://doi.org/10.1016/j.eswa.2018.08.051
Arora, N., & Kaur, P. D. (2020). A Bolasso based consistent feature selection enabled random
forest classification algorithm: An application to credit risk assessment. Applied Soft
Computing Journal, 86, 105936. https://doi.org/10.1016/j.asoc.2019.105936
Bao, W., Lianju, N., & Yue, K. (2019). Integration of unsupervised and supervised machine
learning algorithms for credit risk assessment. Expert Systems with Applications , 128,
301–315. https://doi.org/10.1016/j.eswa.2019.02.033
Beaver, W. H. (1966). Financial ratios as predictors of failure. Journal of Accounting Research,
4, 71–111. https://doi.org/10.2307/2490171
Beaver, W. H., McNichols, M. F., & Rhie, J. (2005). Have financial statements become less
informative? Evidence from the ability of financial ratios to predict bankruptcy. Review of
Accounting Studies, 10(1), 93–122. https://doi.org/10.2139/ssrn.634921
Bifet, A., Holmes, G., Kirkby, R., & Pfahringer, B. (2010). MOA: Massive Online Analysis.
Journal of Machine Learning Research, 11, 1601–1604. https://dl.acm.org/doi/10.5555/
1756006.1859903
Bishop, C. M. (1995). Neural Networks for Pattern Recognition (Advanced Texts in
Econometrics (Paperback). Oxford University Press.
Boyacioglu, M. A., Kara, Y., & Baykan, O. K. (2009). Predicting bank financial failures using
neural networks, support vector machines and multivariate statistical methods: a compara-
tive analysis in the sample of savings deposit insurance fund (SDIF) transferred banks in
Turkey. Expert Systems with Applications, 36(2), 3355–3366. https://doi.org/10.1016/j.eswa.
2008.01.003
Breiman, L. (1995). Better subset regression using the nonnegative garotte. Technometrics,
37(4), 373–384. https://doi.org/10.2307/1269730
Breiman, L. (1996). Heuristics of instability and stabilization in model selection. Annals of
Statistics, 24, 2297–2778. https://doi.org/10.1214/aos/1032181158
Breiman, L., Friedman, J. H., Olshen, R. A., & Stone, C. J. (1984). Classification and regression
trees. The Wadsworth.
Briand, L. C., Freimut, B., & Vollei, F. (2004). Using multiple adaptive regression splines to
support decision making in code inspections. Journal of Systems and Software, 73 (2),
205–217. https://doi.org/10.1016/j.jss.2004.01.015
Campbell, J., Hilscher, J., & Szilagyi, J. (2008). In search of distress risk. The Journal of
Finance, 63(6), 2899–2939. https://doi.org/10.1111/j.1540-6261.2008.01416.x
Chang, Y.-C., Chang, K.-H., & Wu, G.-J. (2018). Application of eXtreme gradient boosting
trees in the construction of credit risk assessment models for financial institutions. Applied
Soft Computing Journal, 73, 914–920. https://doi.org/10.1016/j.asoc.2018.09.029
Chava, S., & Jarrow, R. A. (2004). Bankruptcy prediction with industry effects. Review of
Finance, 8(4), 537–569. https://doi.org/10.2139/ssrn.287474
Chawla, N., Bowyer, K., Hall, L., & Kegelmeyer, W. (2002). SMOTE: Synthetic Minority Over-
sampling Technique. Journal of Artificial Intelligence Research, 16, 321–378. https://doi.org/
10.1613/jair.953
Chi, G., Abedin, M. Z., & Moula, F. E. (2017). Modeling credit approval data with neural net-
works: an experimental investigation and optimization. Journal of Business Economics and
Management, 18 (2), 224–240. https://doi.org/10.3846/16111699.2017.1280844
Chi, G., Uddin, M. S., Abedin, M. Z., & Yuan, k. (2019). Hybrid model for credit risk predic-
tion: An application of neural network approaches. International Journal on Artificial
Intelligence Tools, 28(05), 1–33. https://doi.org/10.1142/S0218213019500179
Chi, G., Yu, S., & Zhou, Y. (2019a). A novel credit evaluation model based on the maximum
discrimination of evaluation results. Emerging Markets Finance and Trade, 56(11),
2543–2562. https://doi.org/10.1080/1540496X.2019.1643717
3088 Y. ZHOU ET AL.

Cortes, C., & Vapnik, V. (1995). Support-Vector Networks. Machine Learning, 20(3), 273–297.
https://doi.org/10.1007/BF00994018
Danenas, P., & Garsva, G. (2015). Selection of support vector machines based classifiers for
credit risk domain. Expert Systems with Applications, 42(6), 3194–3204. https://doi.org/10.
1016/j.eswa.2014.12.001
Demsar, J. (2006). Statistical comparisons of classifiers over multiple datasets. The Journal of
Machine Learning Research, 7, 1–30.
Desai, V. S., Crook, J. N., & Overstreet, G. A. A (1996). A comparison of neural networks and
linear scoring models in the credit union environment. European Journal of Operational
Research, 95 (1), 24–37. https://doi.org/10.1016/0377-2217(95)00246-4
Dimla, D. E., & Lister, P. M. (2000). On-line metal cutting tool condition monitoring. II: tool-
state classification using multilayer perceptron neural networks. International Journal of
Machine Tools and Manufacture, 40 (5), 769–781. https://doi.org/10.1016/S0890-
6955(99)00085-1 https://doi.org/10.1016/S0890-6955(99)00085-1
Ding, Y., Song, X., & Zen, Y. (2008). Forecasting financial condition of Chinese listed compa-
nies based on support vector machine. Expert Systems with Applications, 34(4), 3081–3089.
https://doi.org/10.1016/j.eswa.2007.06.037
Drown, D. J., Khoshgoftaar, T. M., & Seliya, N. (2009). Evolutionary sampling and software
quality modeling of high-assurance systems. IEEE Transactions on Systems, Man, and
Cybernetics - Part A: Systems and Humans, 39(5), 1097–1107. DOI: 10.1109/
TSMCA.2009.2020804 https://doi.org/10.1109/TSMCA.2009.2020804
Du, S. S., Zhai, X., Poczos, B., & Singh, A. (2019). Gradient Descent Provably Optimizes Over-
parameterized Neural Networks. International Conference on Learning Representations
(ICLR).
Dunn, O. J. (1961). Multiple Comparisons among Means. Journal of the American Statistical
Association, 56(293), 52–64. https://doi.org/10.1080/01621459.1961.10482090
Efron, B., Hastie, T., Johnstone, I., & Tibshirani, R. (2004). Least angle regression. The Annals
of Statistics, 32(2), 407–499. https://projecteuclid.org/euclid.aos/1083178935 https://doi.org/
10.1214/009053604000000067
Fan, J., & Li, R. (2001). Variable selection via nonconcave penalized likelihood and its oracle
properties. Journal of the American Statistical Association, 96(456), 1348–1360. https://doi.
org/10.1198/016214501753382273
Friedman, J. H. (1991). Multivariate adaptive regression splines. The Annals of Statistics, 19
(1), 1–67. https://doi.org/10.1214/aos/1176347963
Friedman, M. (1940). A comparison of alternative tests of significance for the problem of rank-
ings. The Annals of Mathematical Statistics, 11(1), 86–92. https://doi.org/10.1214/aoms/
1177731944
Garc
ıa, V., Marqu
es, A. I., & S
anchez, J. S. (2015). An insight into the experimental design for
credit risk and corporate bankruptcy prediction systems. Journal of Intelligent Information
Systems, 44(1), 159–189. https://doi.org/10.1007/s10844-014-0333-4
Gately, E. (1996). Neural Networks for Financial Forecasting: Top Techniques for Designing and
Applying the Latest Trading Systems. John Wiley & Sons, Inc.
Hand, D. J., & Jacka, S. D. (1998). Statistics in Finance. Arnold Applications of Statistics:
London.
Hastie, T., Tibshirani, R., Friedman, J., & Franklin, J. (2005). The elements of statistical learn-
ing: data mining, inference, and prediction. Mathematical Intelligencer, 27 (2), 83–85.
He, H., & Garcia, E. (2009). Learning from imbalanced data. IEEE Transactions on Knowledge
and Data Engineering, 21, 1263–1284. https://doi.org/10.1109/TKDE.2008.239
Hu, L., Gao, W., Zhao, K., Zhang, P., & Wang, F. (2018). Feature selection considering two
types of feature relevancy and feature interdependency. Expert Systems with Applications, 93,
423–434. https://doi.org/10.1016/j.eswa.2017.10.016
Huang, Z., Chen, H., Hsu, C. J., Chen, W. H., & Wu, S. (2004). Credit rating analysis with
support vector machines and neural networks a market comparative study. Decision Support
Systems, 37(4), 543–558. https://doi.org/10.1016/S0167-9236(03)00086-1
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3089

Huang, C. L., Chen, M. C., & Wang, C. J. (2007). Credit scoring with a data mining approach
based on support vector machines. Expert Systems with Applications, 33(4), 847–856. https://
doi.org/10.1016/j.eswa.2006.07.007
Jalal, M., Arabali, P., Grasley, Z., Bullard, J. W., & Jalal, H. (2020). Behavior assessment,
regression analysis and support vector machine (SVM) modeling of waste tire rubberized
concrete. Journal of Cleaner Production, 273, 122960. https://doi.org/10.1016/j.jclepro.2020.
122960
Jiang, H., Ching, W., Yiu, K. F. C., & Qiu, Y. (2018). Stationary Mahalanobis kernel SVM for
credit risk evaluation. Applied Soft Computing, 71, 407–417. https://doi.org/10.1016/j.asoc.
2018.07.005
Jiang, Y., & Jones, S. (2018). Corporate distress prediction in China: A machine learning
approach. Accounting & Finance, 58 (4), 1063–1109. https://doi.org/10.1111/acfi.12432
Jones, S., Johnstone, D., & Wilson, R. (2017). Predicting corporate bankruptcy: An evaluation
of alternative statistical models. Journal of Business Finance & Accounting, 44 (1–2), 3–34.
https://doi.org/10.1111/jbfa.12218
Jones, S. (2017). Corporate bankruptcy prediction: a high dimensional analysis. Review of
Accounting Studies, 22 (3), 1366–1422. https://doi.org/10.1007/s11142-017-9407-1
Jones, S., Johnstone, D., & Wilson, R. (2015). An empirical evaluation of the performance of
binary classifiers in the prediction of credit ratings changes. Journal of Banking & Finance,
56, 72–85. https://doi.org/10.1016/j.jbankfin.2015.02.006
Jones, S., & Wang, T. (2019). Predicting private company failure: A multi-class analysis.
Journal of International Financial Markets, Institutions & Money, 61, 161–188. https://doi.
org/10.1016/j.intfin.2019.03.004.
Kim, K. J., & Ahn, H. (2012). A corporate credit rating model using multi-class support vector
machines with an ordinal pairwise partitioning approach. Computers & Operations Research,
39, 1800–1811. https://doi.org/10.1016/j.cor.2011.06.023
Kouziokas, G. N. (2020). A new W-SVM kernel combining PSO-neural network transformed
vector and Bayesian optimized SVM in GDP forecasting. Engineering Applications of
Artificial Intelligence, 92, 103650. https://doi.org/10.1016/j.engappai.2020.103650
Kozodoi, N., Lessmann, S., Papakonstantinou, K., Gatsoulis, Y., & Baesens, B. (2019). A multi-
objective approach for profit-driven feature selection in credit scoring. Decision Support
Systems, 120, 106–117. https://doi.org/10.1016/j.dss.2019.03.011
Lessmann, S., Baesens, B., Seow, H. V., & Thomas, L. C. (2015). Benchmarking state-of-the-art
classification algorithms for credit scoring: An update of research. European Journal of
Operational Research, 247(1), 124–136. https://doi.org/10.1016/j.ejor.2015.05.030
Lewis, E. M. (1992). An Introduction to Credit Scoring. Fair, Isaac & Co., Inc.
Liang, D., Tsai, C.-F., & Wu, H.-T. (2015). The effect of feature selection on financial distress
prediction. Knowledge-Based Systems, 73, 289–297. https://doi.org/10.1016/j.knosys.2014.10.
010
Lin, W.-Y., Hu, Y.-H., & Tsai, C.-F. (2012). Machine learning in financial crisis prediction: A
survey. IEEE Transactions on Systems, Man and Cybernetics –Part C: Applications and
Reviews, 42(4), 421–436. https://doi.org/10.1109/TSMCC.2011.2170420
Liu, Y., & Schumann, M. (2005). Data mining feature selection for credit-scoring models.
Journal of the Operational Research Society, 56(9), 1099–1108. https://doi.org/10.1057/pal-
grave.jors.2601976
L
opez, J., & Maldonado, S. (2019). Profit-based credit scoring based on robust optimization
and feature selection. Information Sciences , 500, 190–202. https://doi.org/10.1016/j.ins.2019.
05.093
Luo, J., Yan, X., & Tian, Y. (2020). Unsupervised quadratic surface support vector machine
with application to credit risk assessment. European Journal of Operational Research, 280(3),
1008–1017. https://doi.org/10.1016/j.ejor.2019.08.010
Mahajan, V., Jain, A. K., & Bergier, M. (1977). Parameter estimation in marketing models in
the presence of multicollinearity: an application of ridge regression. Journal of Marketing
Research, 14 (4), 586–591. https://doi.org/10.1177/002224377701400419
3090 Y. ZHOU ET AL.

Maldonado, S., Bravo, C., L
opez, J., & P
erez, J. (2017). Integrated framework for profit-based
feature selection and SVM classification in credit scoring. Decision Support Systems, 104,
113–121. https://doi.org/10.1016/j.dss.2017.10.007
Marqu
es, A. I., Garc
ıa, V., & S
anchez, J. S. (2012a). Exploring the behaviour of base classifiers
in credit scoring ensembles. Expert Systems with Applications, 39 (11), 10244–10250. https://
doi.org/10.1016/j.eswa.2012.02.092
Marqu
es, A. I., Garc
ıa, V., & S
anchez, J. S. (2012b). Two-level classifier ensembles for credit
risk assessment. Expert Systems with Applications, 39 (12), 10916–10922. https://doi.org/10.
1016/j.eswa.2012.03.033[Mismatch]
Mason, C. H., William, D., & Perreault, J. R. (1991). Collinearity, power, and interpretation of
multiple regression analysis. Journal of Marketing Research, 28 (3), 268–280. https://doi.org/
10.2307/3172863 https://doi.org/10.1177/002224379102800302
Masters, T. (1995). Advanced Algorithms for Neural Networks: A Cþþ Sourcebook. John Wiley
& Sons, Inc.
Meier, L., Geer, S., & B€uhlmann, P. (2008). The group lasso for logistic regression. Journal of the
Royal Statistical Society: Series B (Statistical Methodology)), 70(1), 53–71. https://doi.org/10.
1111/j.1467-9868.2007.00627.x
Min, J. H., & Lee, Y. C. (2005). Bankruptcy prediction using support vector machine with
optimal choice of kernel function Parameters. Expert Systems with Applications, 28(4),
603–614. https://doi.org/10.1016/j.eswa.2004.12.008
Ohlson, J. (1980). Financial ratios and the probabilistic prediction of bankruptcy. Journal of
Accounting Research, 18 (1), 109–131. https://doi.org/10.2307/2490395
Ping, Y., & Yongheng, L. (2011). Neighborhood rough set and SVM based hybrid credit scor-
ing classifier. Expert Systems with Applications, 38(9), 11300–11304. https://doi.org/10.1016/j.
eswa.2011.02.179
Reed, R. D., & Marks, R. J. (1999). Neural Smithing: Supervised Learning in Feedforward
Artificial Neural Networks. The MIT Press.
Shin, K. S., Lee, T. S., & Kim, H. J. (2005). An application of support vector machines in
bankruptcy prediction model. Expert Systems with Applications, 28(1), 127–135. https://doi.
org/10.1016/j.eswa.2004.08.009
Shumway, T. (2001). Forecasting bankruptcy more accurately: a simple hazard model. The
Journal of Business, 74(1), 101–124. https://doi.org/10.2139/ssrn.171436
Swets, J. A., Dawes, R. M., & Monahan, J. (2000). Better decisions through science. Scientific
American, 283 (4), 82–87. https://doi.org/10.1038/scientificamerican1000-82
Thomas, L. C., Edelman, D. B., & Crook, L. N. (2002). Credit Scoring and Its Applications.
Philadelphia. Society for Industrial and Applied Mathematics.
Tian, S., & Yu, Y. (2017). Financial ratios and bankruptcy predictions: An international evi-
dence. International Review of Economics & Finance ,51, 510–526. https://doi.org/10.1016/j.
iref.2017.07.025
Tian, S., Yu, Y., & Guo, H. (2015). Variable selection and corporate bankruptcy forecasts.
Journal of Banking & Finance, 52, 89–100. https://doi.org/10.1016/j.jbankfin.2014.12.003
Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal
Statistical Society: Series B (Methodological), 58(1), 267–288. https://doi.org/10.1111/j.2517-
6161.1996.tb02080.x
Trippi, R. R., & Turban, E. (1993). Neural Networks in Finance and Investing: Using Artificial
Intelligence to Improve Real-World Performance. IRWIN.
Tsai, C.-F., & Sung, Y.-T. (2020). Ensemble feature selection in high dimension, low sample
size datasets: Parallel and serial combination approaches. Knowledge-Based Systems, 203,
106097. https://doi.org/10.1016/j.knosys.2020.106097
Uddin, M. S., Chi, G., Al Janabi, M. A. M., & Habib, T. (2020b). Leveraging random forest in
micro-enterprises credit risk modelling for accuracy and interpretability. International
Journal of Finance & Economics, 1–17. https://doi.org/10.1002/ijfe.2346
 ECONOMIC RESEARCH-EKONOMSKA ISTRAŽIVANJA 3091

Uddin, M. S., Chi, G., Habib, T., & Zhou, Y. (2020a). An alternative statistical framework for
credit default prediction. Journal of Risk Model Validation, 14 (2), 1–36. https://doi.org/10.
21314/JRMV.2020.220
Vinod, H. D. (1978). A survey of ridge regression and related techniques for improvements
over ordinary least squares. The Review of Economics and Statistics, 60 (1), 121–131.
https://ssrn.com/abstract=1750091 https://doi.org/10.2307/1924340
Wang, D., Zhang, Z., Bai, R., & Mao, Y. (2018). A hybrid system with filter approach and
multiple population genetic algorithm for feature selection in credit scoring. Journal of
Computational and Applied Mathematics, 329, 307–321. https://doi.org/10.1016/j.cam.2017.
04.036
Wei, G., Zhao, J., Feng, Y., He, A., & Yu, J. (2020). A novel hybrid feature selection method
based on dynamic feature importance. Applied Soft Computing Journal, 93, 106337. https://
doi.org/10.1016/j.asoc.2020.106337
West, D. (2000). Neural network credit scoring models. Computers & Operations Research, 27
(11–12), 1131–1152. https://doi.org/10.1016/S0305-0548(99)00149-5. https://doi.org/10.1016/
S0305-0548(99)00149-5
Wu, X., Kumar, V., Ross Quinlan, J., Ghosh, J., Yang, Q., Motoda, H., McLachlan, G. J., Ng,
A., Liu, B., Yu, P. S., Zhou, Z.-H., Steinbach, M., Hand, D. J., & Steinberg, D. (2008). Top
10 algorithms in data mining. Knowledge and Information Systems, 14(1), 1–37. https://doi.
org/10.1007/s10115-007-0114-2
Xie, C., Luo, C., & Yu, X. (2011). Financial distress prediction on SVM and MDA methods:
the case of Chinese listed companies. Quality & Quantity, 45, 671–686. https://doi.org/10.
1007/s11135-010-9376-y
Yu, L., & Liu, H. (2003). Feature selection for high-dimensional data: A fast correlation-based
filter solution. International Conference on Machine Learning, 2, 856–863.
Yu, L., Zhou, R., Tang, L., & Chen, R. (2018). A DBN-based resampling SVM ensemble learn-
ing paradigm for credit classification with imbalanced data. Applied Soft Computing, 69,
192–202. https://doi.org/10.1016/j.asoc.2018.04.049
Zekic-Susac, M., Sarlija, N., & Bensic, M. (2004). Small Business Credit Scoring: A
Comparison of Logistic Regression, Neural Networks, and Decision Tree Models. 26th
International Conference on Information Technology Interfaces, Croatia. https://doi.org/10.
1109/ITI.2004.241696
Zheng, K., Chen, Y., Jiang, Y., & Qiao, S. (2020). A SVM based ship collision risk assessment
algorithm. Ocean Engineering, 202, 107062. https://doi.org/10.1016/j.oceaneng.2020.107062
Zhong, H., Miao, C., Shen, Z., & Feng, Y. (2014). Comparing the learning effectiveness of BP,
ELM, I-ELM, and SVM for corporate credit ratings. Neurocomputing , 128, 285–295. https://
doi.org/10.1016/j.neucom.2013.02.054
Zhou, L., Lai, K. K., & Yu, L. (2010). Least squares support vector machines ensemble models
for credit scoring. Expert Systems with Applications, 37(1), 127–133. https://doi.org/10.1016/
j.eswa.2009.05.024
Zou, H. (2006). The adaptive lasso and its oracle properties. Journal of the American Statistical
Association, 101(476), 1418–1429. https://doi.org/10.1198/016214506000000735