See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/351169516

Stock Market Prediction using Supervised Machine Learning Techniques: An

Overview

Conference Paper · December 2020

DOI: 10.1109/CSDE50874.2020.9411609

CITATIONS READS

19 1,437

3 authors, including:

Zaharaddeen Karami Lawal Rufai Zakari Yusuf

Universiti Brunei Darussalam University of Electronic Science and Technology of China
15 PUBLICATIONS 127 CITATIONS 23 PUBLICATIONS 180 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Zaharaddeen Karami Lawal on 13 May 2022.

The user has requested enhancement of the downloaded file.

 Stock Market Prediction using Supervised
Machine Learning Techniques: An Overview
Zaharaddeen Karami Lawal Hayati Yassin Rufai Yusuf Zakari
Department of Computer Science Faultyof Integrated Technologies Department of Engineering
Federal University Dutse Universiti Brunei Darussalam University of Electronic Science and
2020 IEEE Asia-Pacific Conference on Computer Science and Data Engineering (CSDE) | 978-1-6654-1974-1/20/$31.00 ©2020 IEEE | DOI: 10.1109/CSDE50874.2020.9411609

[email protected] [email protected] Technology of China

[email protected]

Abstract— Stock price prediction is one of the most [6]. Financial organizations and traders have made different
extensively studied and challenging glitches, which is acting so exclusive models to attempt to beat the market for themselves
many academicians and industries experts from many fields or their customers, however seldom has anybody
comprising of economics, and business, arithmetic, and accomplished reliably higher-than-normal degrees of
computational science. Predicting the stock market is not a profitability [7].
simple task, mainly as a magnitude of the close to random-walk
behavior of a stock time series. Millions of people across the Machine learning has the potential predict stock market by
globe are investing in stock market daily. A good stock price training and testing the models with the historical datasets,
prediction model will help investors, management and decision social media data, crawled financial news or trends.
makers in making correct and effective decisions. In this paper,
we review studies on supervised machine learning models in This paper discussed the usage of machine learning
stock market predictions. The study discussed how supervised models in stock market predictions where it details out how
machine learning techniques are applied to improve accuracy of supervised machine learning algorithms/techniques
stock market predictions. Support Vector Machine (SVM) was (Classification and Regression) are applied. The paper is
found to be the most frequently used technique for stock price organize as follows: Section II elucidates the methodology
prediction due to its good performance and accuracy. Other used to conduct this review; Section III explains the findings
techniques like Artificial Neural Network (ANN), K-Nearest and discussion. Conclusions are drawn in Section IV.
Neighbor (KNN), Naïve Bayes, Random Forest, Linear
Regression and Support Vector Regression (SVR) also showed II. SYSTEMATIC REVIEW METHODOLOGY
a promising prediction result. Research method that has been adopted in this paper is
Systematic Literature Review (SLR). A systematic review can
Keywords— Stock Market Prediction, Supervised Machine
be define as a research technique and process for identifying
Learning, Classification, Regression, Support Vector Machine
and critically appraising relevant research, as well as for
(SVM), Artificial Neural Network (ANN)
collecting and analyzing data from said research[8]. Execution
I. In t r o d u c t io n of a systematic review can be grouped into three main stages:
planning, conducting the review and reporting the review[9].
Stock price prediction is one of the most extensively
studied and challenging glitches, which attracting several
researchers from many fields comprising of economics, and
business, arithmetic, and computational science. [1]. Stock
price prediction has been at focus for years since it can yield
significant profits. Predicting the stock market is not a simple
task, mainly as a consequence of the close to random-walk
behavior of a stock time series[2]. Stock market prediction is
one of the most pains taking tasks due to its volatility. The
challenge of stock market prediction is so productive that even
a small increase in prediction by the new model can bring
about huge profits. Stock prices are an essential part of the
prediction[3]. In recent years, the fast growing financial
markets opened new horizons for investors and the same time
bringing new challenges for financial analysts in their efforts
to make effective decisions and reduce the investment risks.
Stock market is a highly dynamic and complex system since
there are a great number of interacting factors that affect the Fig. 1. Systematic Literature Review Process[9].
future prices[4]. Researchers have worked very tough to
Literature review from all the journal publications and
ascertain the point that financial markets are predictable. With
conference articles gathered and used to answer the research
the advancement and availability of technology, stock markets
questions mentioned as follows:
are now more accessible to investors. Various models have
been proposed, both in industry and academia, for stock R1: What are the Supervised Learning Algorithms that are
market prediction ranging from machine learning, to data commonly used in Stock Market Prediction?
mining, to statistical models[5].
R2: How Supervised Machine Learning techniques are
A few speculations with respect to markets exchanges applied in Stock Market Prediction?
have been conceptualized throughout the long term. They
either attempt to clarify the idea of market exchanges or R3: Which technique is considered the best or frequently
attempt to clarify whether the business sectors can be beaten used for Stock Market Prediction among the Supervised
Machine Learning techniques?

978-1-6654-1974-1/21/$31.00 ©2021 IEEE

Authorized licensed use limited to: California State University Fresno. Downloaded on June 21,2021 at 18:25:10 UTC from IEEE Xplore. Restrictions apply.
R4: What are the directions of future research on Stock Market TABLE I. SUPERVISED MACHINE LEAARNING TECHNIQUES

Prediction? Journal/Conference Stock Market Algorithm Type

Paper Prediction Models
A. Supervised Machine Learning Techniques in Stock
Robert P. Shumaker et SVM Regression
Market Prediction al [10]
Stock-market prediction using machine-learning Han Lock Slew et al Linear Regression (LR) Regression
technique which aims at developing effective and efficient [11]
Khalid Alkhatib et al KNN and None Linear Classification and
models that can provide a better and higher rate of prediction [12] Regression Regression
accuracy [40]. Numerous classification and regression models Carol Hargreaves et al Logistic Regression Regression
have been used in stock market predictions for many years. [13]
Several supervised machine learning techniques applied in Usman Hegazy et al [2] LS-SVM Classification
stock market prediction yielded a better outcome. For more Abubakar S. Magaji et SVM-SMO Classification
than a decade, a number of supervised machine models and al [14]
Saahil Madge et al[1] SVM Classification
techniques has been proposed or implemented for stock
market predictions. Plethora of researchers and technology Shashaank D.S et al Hybrid future selection Classification
[15]
practitioners tried their level best in order to come up with an
Hakob Grigorian et al SVM Classification
algorithm that w ill give the best stock market prediction. [4]
Khan, W. Ghazanfar et SVM, KNN and Naïve Classification
The majority of work published on supervised machine
al [16] Bayes classifiers
learning techniques used in stock market predictions are Makram Zaidi et al ANN Regression
categorized as follows: are not included in the study: [17]
Aparna Nayak et al Decision Tree, Logistic Classification and
• The comparison of classification algorithms [18] Regression and SVM Regression
• The comparison of regression algorithms Mustansar Ali SVM Classification
Ghazanfar et al [5]
• The comparison of classification and regression Mr. Pramod Mali et al LR and Multiple Regression
algorithms of [19] Regression
Hakan Gunduz et al SVM and ANN Classification
B. Systemtic Literature Review Approach [20]
D. A Puspitasari et al SVM and KNN Classification
We collected journals and conference papers via electronic
[21]
sources using credible and relevant databases such as IEEE Gareja Pradip et al [22] Naïve Bayes Classification
Xplore, ScienceDirect, Google Scholar, Semantic Scholar, Meizhen Liu et al [23] SVM Classification
Open Access and Springer. The search keywords used to R.Seethalakshmi et al Linear Regression (LR) Regression
search publications are “ Stock Market Predictions using [24]
Machine Learning” , “ Stock Market Prediction using Bruno Miranda Support Vector Regression
Henrique et al [25] Regression (SVR)
Supervised Machine Learning Techniques/Algorithms” and “
V Kranthi Sai Reddy et SVM Classification
Stock Market Prediction using Classification/ Regression al [26]
Algorithms” . K. Hiba Sadia et al [27] SVM and Random Classification
Forest
The study has address issues of Supervised Machine Lakshminarayanan et LSTM and SVM Classification
Learning and Stock Market Predictions therefore they were al [3]
excluded. The articles has reduced the number from 84 to 38. Neha Bhardwaj et al Random forest and KNN Classification
[28]
The following criteria has been used in eliminating the Mariam Moukalled et RNN, SVM and SVR Classification and
papers that are not included in the study: al [29] Regression
Dev Shah et al [6] ANN, SVM and Naïve Classification
• Exclude if the paper do not emphasis on Machine Bayes
Learning Predictions. Amit Gupta et al [30] ANN and SVM Regression
Nitin Nand Kumar LR and SVR Regression
• Exclude if the paper do not discuss about Supervised Sakhare et al [31]
Machine Learning in Stock Market Predictions or Poornima S P et al [32] KNN and LR Classification and
Stock Market Predictions Regression
Vaishnavi Guruj et al LR and SVM Classification and
III. FINDINGS AND DISCUSSION [33] Regression
Alaa F. Sheta et al [34] ANNs, SVM and Classification and
In this section, the answers to research questions as Multiple Linear Regression
presented in Section II we have provided in details where the regression (MLR)
result of first research question is discussed in Section 4.1 N P Samarth et al [35] Random Forest Classification
while the answers to the remaining research questions are Sikki setti SVM and KNN Classification
discussed in Section III.A, III.B and III.C respectively. The Jyothirmayee et al [7]
research questions are provided with details answers based on Chi-Cheng Chen et al ANN and SVM Classification
[36]
the chosen case studies. Wasiat Khan et al [37] Random Forest Classification
A. Supervised Machine Learning (SML) Algorithm Sidra Mehtab et al [38] ANN, LR, SVM, KNN Classification and
and Decision Tree Regression
Investigation in Stock Market Prediction Mehtabhom Obthong ANN, SVM, SVR, Classification and
Table I illustrated the most widely used supervised et al [39] KNN, LR, etc. Regression
machine learning techniques in stock market prediction. Some Isaac Kofi Nti et al Decision tree, SVM and Classification and
studies used a single while others used more than 1. [4°] NN Regression

Authorized licensed use limited to: California State University Fresno. Downloaded on June 21,2021 at 18:25:10 UTC from IEEE Xplore. Restrictions apply.
 Based on 38 case studies reviewed, the supervised the most frequently used supervised machine learning
machine learning techniques that are used for stock market algorithm among the classification and regression categories
predictions summarized in Table II as following. Majority of that are used in stock market predictions from the 38 case
the case studies showed that classification is the most widely studies.
used for stock market /price predictions.

TABLE II. SUPERVISED MACHINE LEAARNING ALGORITHMS

Supervised Machine Learning Algorithms Case Studies

Classification 50%
Regression 24%
Comparison between Classification and Regression 24%

Table II showed that classification is the most frequently

used algorithm for stock market predictions. To answer the
first research question clearly from the aforementioned
research questions. The chart below shows the publications
based on technique(s) used in ten years. Although the chart is
static due to availability of publications we have for this study.
Fig. 2 illustrated journals/ conference papers published in
2006 and 2012 used regression algorithms, while in 2013
published articles uses classification, regression and
comparison between classification and regression. In 2014
and 2015 the authors used classification technique in their
publications. From 2016 to 2020 comparison between both
algorithms rose from a single publication in a year to 4 and 3
D. Direction of future research on Stock Market Prediction
publications in 2019 and 2020 respectively.
To overcome the challenges in the stock market analysis,
B. SML Techniques in Stock Market Prediction several computational models based on soft-computing and
Supervised machine learning algorithms can be applied in machine learning paradigms have been used in the stock­
stock market predictions depending on how they are designed market analysis, prediction, and trading[40]. Despite the
to perform. Fig. 2 and Table III summarized the answers to better accurate results recorded in stock market predictions
the second research question of this paper, in which it by using techniques like Support Vector Machine (SVM), K-
clarifies how supervised machine learning algorithms are Nearest Neighbor (KNN), Support Vector Regression (SVR),
used or applied in stock market predictions. Linear Regression and Artificial Neural Network (ANN).

Fig. 2. Publications on Stock Market prediction using supervised ML in the There is need for an efficient technique(s) that w ill overcome
last ten years. the challenges faced by using the existing techniques. In spite
of the high accuracy, robustness and ability to handle noisy
C. SML Models in Stock Market Prediction and missing data. Over fitting remains the challenge of using
Many algorithms have shown a promising prediction ANN in stock market prediction. SVM is also sensitive to
result, depending on the type of data used for training models. outliners and parameter selection. RNN has also shown a
Some algorithms perform better with small amount of data good prediction result, but training its models is a difficult
while some predict better with large amount data. Support task. SVR is also sensitive to user’s defined free parameters.
Vector Machine (SVM) are the most commonly used Addressing the aforementioned problems of the existing
algorithm found in this study. Fig. 3 summarized the answer techniques are the future research direction directions of
to third research question of this study in which it explains stock market prediction using supervised machine learning.

Authorized licensed use limited to: California State University Fresno. Downloaded on June 21,2021 at 18:25:10 UTC from IEEE Xplore. Restrictions apply.
 TABLE III. SUPERVISED MACHINE LEAARNING TECHNIQUE IN STOCK MARKET PREDICTION

Journal/ Stock Market Prediction Models research data and then applies the SVM technique to
Conference perform time series prediction.
Paper
Robert P. The authors developed four different models and varied Khan, W. The author(s) o f this paper presents a detailed study on
Shumaker et al the data given to them. The first model, Regress, was a Ghazanfar et al data of London, New York, and Karachi stock exchange
[10] simple linear regression estimate of the +20-minute [16] markets to predict the future trend in these stock
stock price. While they acknowledge the obvious exchange markets. In this research, they have applied
violation of Random Walk Theory, within such a machine learning classifiers before and after applying
compressed amount o f time weak predictive ability principle component analysis (PCA) and reported errors
remains. Next, the three models use the supervised and accuracy of the algorithms before and after applying
learning o f SVM regression to compute their +20- PCA.
minute predictions. Makram Zaidi The dataset used for this research covered the trading
Han Lock Slew The authors o f this paper examine the theory and et al [17] days from 5th April 2007 to 1st January 2015. With
et al [11] practice o f regression techniques for the prediction o f logistic regression, it may be observed that four
stock price trends by using a transformed data set in an variables i.e. open price, higher price, lower price, and
ordinal data format. The original pre-transformed data oil can classify up to 81.55% into two categories up and
source contains data of heterogeneous data types used down. Although with neural networks The prediction
for handling of currency values and financial ratios. accuracy o f the model is high both for the training data
Khalid Alkhatib In this paper, the authors applied the K-Nearest (84.12%) and test data (81.84%).
et al [12] Neighbor algorithm and Non-Linear Regression Aparna Nayak In this paper, an effort has been made to the prediction
approach in order to predict stock prices for a sample o f et al [18] o f the stock market trend. Two models were built one
6 main companies itemized on the Jordanian stock for daily prediction and the other one is for monthly
exchange to support investors, management, decision­ prediction. Supervised machine learning algorithms
makers, and users in making correct and informed were used to build the models. As part of the daily
investments decisions. According to their results, the prediction model, historical prices are combined with
KNN algorithm is robust with a small error ratio; sentiments. Up to 70% o f accuracy is witnessed using
consequently, the results were coherent and also supervised machine learning algorithms on the day-to­
reasonable. day prediction model. The monthly prediction model
Carol In this paper, the authors construct a framework that tries to calculate whether there is any match between
Hargreaves et al enables them to make class predictions about industrial any two months trend.
[13] stock performances. In order to have a systemized Mustansar Ali In this research, authors applied and compared salient
approach for the selection of stocks and a high Ghazanfar et al machine learning algorithms to predict stock exchange
likelihood o f the performance o f the stock price [5] volume. The performance of these algorithms has been
increasing, they applied several analytical techniques. A compared using accuracy metrics on the dataset,
trading strategy is also designed and the performance o f collected over the period of six months, by crawling the
the stocks evaluated. KSE and SSE website.
Usman Hegazy This paper proposes a machine learning model to Mr. Pramod The development and implementation o f a stock price
et al [2] predict stock market prices. The authors proposed an Mali et al [19] prediction was explained in this project and for this
algorithm in order to integrate Particle swarm purpose regression algorithm and object-oriented
optimization (PSO) and least square support vector approach of software development is used.
machine (LS-SVM). The proposed model was applied Hakan Gunduz This paper aimed at the stock daily return prediction
and evaluated using thirteen benchmark financials et al [20] problem, the set o f features is extended to contain
datasets and compared with artificial neural networks indicators not only for the stock to be predicted itself but
with the Levenberg-Marquardt (LM) algorithm. also a set o f other stocks and currencies. Afterwards,
Abubakar S. The authors presented the Naive Bayes and SVM-SMO Different feature selection and classification methods
Magaji et al algorithms as a tool for predicting the Nigerian Stock are utilized for prediction.
[14] Exchange Market; subsequently, they used the same D. A Puspitasari In this paper, Support Vector Machines (SVM) with K
transformed data of the NSEM and explored the et al [21] Nearest Neighbor (KNN) approach applied to forecast
implementation o f the Logistic function on the Back­ stock prices of a listed companies in Indonesia Stock
propagation algorithm on the WEKA platform, and Exchange (IDX). The stock data are collected from
results obtained, made them to also conclude that the January 2013 to December 2016.
Back-propagation model o f Artificial Neural Network Gareja Pradip et In this research, the authors explain, the development
(ANN) performed very well and thus it is another al [22] and implementation of a stock market price prediction
algorithm that can effectively and efficiently be used for application using a machine learning algorithm.
predicting the Nigerian Stock Exchange Market. Meizhen Liu et The authors developed Support vector machine (SVM)
Saahil Madge et This study uses day-to-day ultimate prices for 34 al [23] based on statistical learning theory new method, its
al[1] technology stocks to compute price volatility and training algorithm is basically a problem of solving the
momentum for individual stocks and for the whole quadratic programming.
sector. Their model tries to predict whether a stock price R.Seethalakshm This paper concentrates on best independent variables to
erstwhile in the future will be higher or lower than it is i et al [24] predict the closing value of the stock market.
on a given day. Bruno Miranda In this stud, the authors used a machine learning
Shashaank D.S This paper use Weka to try is to explore and rate the Henrique et al technique called Support Vector Regression (SVR) to
et al [15] performance o f classifiers based on the features selected [25] predict stock prices for big and minor capitalizations and
by Hybrid Feature Selection. The authorized dataset for in three different markets, using prices with both daily
predicting the turnover was taken from www.bsc.com and up-to-the-minute frequencies.
and included the stock market values o f various V Kranthi Sai In this paper, machine learning technique called Support
companies over the past 10 years. Reddy et al [26] Vector Machine (SVM) was used to predict stock prices
Hakob This research emphasizes on financial time series for the large and small capitalizations and in the three
Grigorian et al prediction problem. The authors combined prediction distinct markets, employing prices with both every day
[4] model based on support vector machines (SVM) with and up-to-the-minute occurrences.
independent component analysis which they called K. Hiba Sadia et The author o f this paper reviewed a more feasible
(ICA) (called SVM-ICA) .They first uses the ICA al [27] method to predict the stock movement with higher
technique to remove important features from the accuracy. The paper also presents a machine-learning

Authorized licensed use limited to: California State University Fresno. Downloaded on June 21,2021 at 18:25:10 UTC from IEEE Xplore. Restrictions apply.
 model to predict the longevity of stock in a competitive models outperform benchmark market index. (2) By
market. risk-adjusted measures, the empirical results suggest
Lakshminaraya This study presents a comparative study of the that Random Forest generates the best performance,
nan et al [3] performance of Long Short-Term Memory (LSTM) followed by SVM and ANN.
neural network models with Support Vector Machine Wasiat Khan et In this study, the authors compare results o f different
(SVM) regression models. The framework built as a part al [37] algorithms to find a consistent classifier. Finally, for
of this study comprises of eight models. achieving maximum prediction accuracy, deep learning
Neha Bhardwaj In this paper, the authors presented a comparison of is used and some classifiers are ensemble. The
et al [28] machine learning aided algorithms to evaluate the stock experimental results show that highest prediction
prices in the future to analyze market behavior. Our accuracies of 80.53% and 75.16% are achieved using
method is able to correctly analyze supervised social media and financial news, respectively. They also
algorithms and compare which algorithm performs the show that New Y ork and Red Hat stock markets are hard
best to predict the future stock market prices in the to predict, New York and IBM stocks are more
market. influenced by social media, while London and
Mariam In this work, the authors proposed an automated trading Microsoft stocks by financial news. Random forest
Moukalled et al system that integrates mathematical functions, machine classifier is found to be consistent and highest accuracy
[29] learning, and other external factors such as news’ of 83.22% is achieved by its ensemble.
sentiments for the purpose of attaining enhanced stock Sidra Mehtab et In this paper, the authors presented a very robust and
prediction accuracy and allotting gainful trades. al [38] accurate framework o f stock price prediction that
Dev Shah et al This paper reviewed stock markets and taxonomy of consists o f an agglomeration o f statistical, machine
[6] stock market prediction methods. The study the focus on learning and deep learning models. They also use the
some of the research achievements in stock analysis and daily stock price data, collected at five minutes interval
prediction. The paper also discussed technical, of time, of a very famous company that is listed in the
fundamental, short- and long-term methods used meant National Stock Exchange (NSE) of India. They built
for stock analysis. eight classification and eight regression models based
Amit Gupta et The authors of this paper tend to utilize car backward on statistical and machine learning approaches.
al [30] model to predict the more drawn out term estimation of Furthermore, a deep learning regression model using a
a stock. The model is extremely trendy and that they long-and-short-term memory (LSTM) network was also
examine in predicting the stock costs in all respects built.
accurately. Mehtabhom This study, reviewed and compared the state-of the-art
Nitin Nand In this paper, the authors compared three different Obthong et al o f ML algorithms and techniques that have been used in
Kumar Sakhare machine learning algorithms, namely, Linear [39] finance, specifically the stock value prediction. The
et al [31] Regression, Polynomial Regression and Support Vector number o f ML algorithms and techniques. Have been
Regression. They have applied stated techniques on data discussed in terms o f types o f input, purposes,
consisted of index and stock prices of S&P 500. advantages and disadvantages.
Poornima S P et In this study, the authors applied KNN method and Isaac Kofi Nti et In this research, the authors performed an extensive
al [32] linear regression aimed at predicting the stocks. The al [40] comparative analysis o f ensemble techniques such as
performance of linear Regression model on the certain boosting, bagging, blending, and super learners
data set is better when compared to KNN algorithm (stacking). Using Decision Trees (DT), Support Vector
technique. The stock holders can invest assuredly based Machine (SVM) and Neural Network (NN), they have
on the results obtained from the model. constructed twenty-five (25) different ensemble
Vaishnavi This survey compared two algorithms, Linear regressors and classifiers. They have also compared
Guruj et al [33] Regression (LR) and Support Vector Machines (SVMs) their execution times, accuracy, and error metrics over
to predict stock market values. Advantages and stock data from Ghana Stock Exchange (GSE),
disadvantages of using both algorithms were also Johannesburg Stock Exchange (JSE), Bombay Stock
discussed. Exchange (BSE-SENSEX), and New York Stock
Alaa F. Sheta et In this article, the authors explore the use of Artificial Exchange (NYSE), from January 2012 to December
al [34] Neural Networks (ANNs) and Support Vector 2018.
Machines (SVM) to build prediction models for the
S&P 500 stock index. They have also show how IV. CONCLUSION
traditional models such as multiple linear regression
(MLR) behave in this case. The developed models will
The systematic review was organized based on the works
be assessed and compared based on a number of published in Stock Market Prediction. In this paper, we
evaluation criteria. discussed the usage of machine learning models in stock
N P Samarth et In this paper, the authors proposed an approach that uses market predictions where we details out how supervised
al [35] machine learning algorithms and will be trained on the
historical stock dataset that are available and gain
machine learning algorithms/techniques Classification and
intelligence, later it uses the knowledge acquired for Regression are applied in stock market for predictions and
predicting the stock prices accurately. Using Random we reviewed the literatures of the concepts and applications,
Forest Regression algorithm. and examined them using dimensions related to ongoing and
Sikkisetti This paper is predominantly concerned with the best
emerging issues in Stock price prediction. This study is
Jyothirmayee et model to predict the stock market value. For the period
al [7] of the mechanism of contemplating the various based on a literature review of Stock Market Prediction using
techniques and variables that can be taken into a keyword Index and article title search on IEEE Xplore,
consideration, the authors discovered five models Science Direct, Google Scholar, Semantic Scholar, Open
Which are based on supervised learning techniques i.e..,
Access and Springer.
Support Vector Machine (SVM), Random Forest, K-
Nearest Neighbor (KNN), Bernoulli Naïve Bayes. The
experimental results show that SVC performs the best
REFERENCES
for large datasets and Random Forest, Naïve Bayes is [1] Madge, Saahil, and Swati Bhatt. "Predicting stock price direction using
the best for small datasets. support vector machines." Independent work report spring, 2015.
Chi-Cheng This study compared three machine learning models: [2] Hegazy, Osman, Omar S. Soliman, and Mustafa Abdul Salam. "A
Chen et al [36] Artificial neural networks (ANN), support vector machine learning model for stock market prediction." arXiv preprint
machines (SVM), and random forest. With a arXiv:1402.7351,2014.
performance evaluation of S&P 500 index historical
data spanning from 2014 to 2018, the authors found that:
(1) By overall performance measures, machine learning

Authorized licensed use limited to: California State University Fresno. Downloaded on June 21,2021 at 18:25:10 UTC from IEEE Xplore. Restrictions apply.
[3] Lakshminarayanan, Sai Krishna, and John McCrae. "A Comparative [22] G. Pradip, C. Bari, and J. S. Nandhini, “ Stock market prediction using
Study of SVM and LSTM Deep Learning Algorithms for Stock Market machine learning,” vol. 3, no. 10.
Prediction." In AICS, pp. 446-457. 2019. [23] M. Liu, C. Duan, and S. Price, “ A Review of Using Support Vector
[4] E. Studies, “ A Stock Market Prediction Method Based on Support Machine Theory to Do Stock Forecasting,” vol. 147, no. Ncce 2018,
Vector Machines (SVM) and Independent Component Analysis pp. 2013-2015.
(ICA),” pp. 12-21, 2000. [24] A. Mathematics, “ Analysis of stock market predictor variables using
[5] M. A. Ghazanfar, S. A. Alahmari, Y. Fahad, A. Mustaqeem, and M. A. Linear Regression,” vol. 119, no. 15, pp. 369-378.
Azam, “Using Machine Learning Classifiers to Predict Stock Exchange [25] B. M. Henrique, V. A. Sobreiro, and H. Kimura, “ Stock Price
Index,” vol. 7, no. 2, pp. 24-29, doi: 10.18178/ijmlc.2017.7.2.614. Prediction Using Support Vector Regression on Daily and Up to the
[6] Shah, Dev, Haruna Isah, and Farhana Zulkernine. "Stock market Minute Prices,” J. Financ. Data Sci., no. 2018, doi:
analysis: A review and taxonomy o f prediction techniques." 10.1016/j .jfds.2018.04.003.
International Journal of Financial Studies vol 7, no. 2 pp 26., 2019 [26] K. Sai and R. Vanukuru, “ Stock Market Prediction Using Machine
[7] S. Jyothirmayee, V. D. Kumar, C. S. Rao, and R. S. Shankar, Learning,” doi: 10.13140/RG.2.2.12300.77448.
“ Predicting Stock Exchange using Supervised Learning Algorithms,” [27] K. H. Sadia, A. Sharma, A. Paul, and S. Sanyal, “ Stock Market
no. 1, pp. 4081-4090, doi: 10.35940/ijitee.A4144.119119. Prediction Using Machine Learning Algorithms,” no. 4, pp. 25-31.
[8] H. Snyder, “ Literature review as a research methodology: An overview [28] N. Bhardwaj and A. Ansari, “ Prediction of Stock Market using
and guidelines,” J. Bus. Res., vol. 104, no. 2019, pp. 333-339, doi: Machine Learning Algorithms,” pp. 5994-6005, 2008.
10.1016/j.jbusres.2019.07.039.
[29] M. Moukalled and W. E. Mohamad, “ Automated Stock Price
[9] N. S. Selamat, “ An Overview of Big Data Usage in Disaster Prediction Using Machine Learning.”
Management,” vol. 11, no. 1, pp. 35-40.
[30] A. Gupta and T. J. Nagalakshmi, “ Stock Price Prediction using Linear
[10] R. P. Schumaker and H. Chen, “ Textual Analysis of Stock Market Regression in Machine Learning,” no. 12, pp. 1382-1385, doi:
Prediction Using Breaking Financial News : The AZFinText System,” 10.35940/ijitee.L3932.1081219.
no. 2006, pp. 1-29.
[31] N. Sakhare and S. S. Imambi, “ Performance Analysis of Regression
[11] H. L. Siew, “Regression Techniques for the Prediction of Stock Price Based Machine Learning Techniques for Prediction of Stock Market
Trend,” doi: 10.1109/ICSSBE.2012.6396535. Movement,” no. 6, pp. 206-213.
[12] M. K. A. Shatnawi, “ Stock Price Prediction Using K -Nearest Neighbor [32] S. P. Poornima, C. N. Priyanka, P. Reshma, S. K. Jaiswal, and K. N.
( k NN ) Algorithm,” vol. 3, no. 3, pp. 32-44, 2009. Surendrababu, “ Stock Price Prediction using KNN and Linear
[13] C. Hargreaves, “ Prediction of Stock Performance Using Analytical Regression,” no. 5, pp. 142-145.
Techniques,” vol. 5, no. 2, pp. 136-142, doi: 10.4304/jetwi.5.2.136- [33] V. Gururaj, V. R. Shriya, and K. Ashwini, “ Stock Market Prediction
142. using Linear Regression and Support Vector Machines,” vol. 14, no. 8,
[14] Magaji, A.S,” An intense Nigerian stock exchange market prediction pp. 1931-1934, 1996.
using logistic with back-propagation ANN model”, Science World [34] S. E. M. Ahmed, “ A Comparison between Regression , Artificial
Journal, 9(2), pp.8-13, 2014. Neural Networks and Support Vector Machines for Predicting Stock
[15] Shashaank, D. S., V. Sruthi, M. L. Vijayalakshimi, and Jacob Shomona Market Index,” vol. 4, no. 7, pp. 55-63.
Garcia. "Turnover prediction of shares using data mining techniques: a [35] N. P. Samarth, G. V Bhat, and N. Hema, “ Stock Price Prediction,” no.
case study." arXiv preprint arXiv:1508.00088 ,2015. 2, pp. 425-429, 2019, doi: 10.35940/ijitee.B1042.1292S19.
[16] Khan, W., M. A. Ghazanfar, M. Asam, A. Iqbal, S. Ahmad, and Javed [36] C. Chen, C. Chen, and T. Liu, “ Investment Performance of Machine
Ali Khan. "PREDICTING TREND IN STOCK MARKET Learning: Analysis of S & P 500 Index,” vol. 10, no. 1, pp. 59-66,
EXCHANGE USING MACHINE LEARNING CLASSIFIERS." 2000.
Science International 28, no. 2 ,2016.
[37] W. Khan, M. Ali, G. Muhammad, A. Azam, A. Karami, and K. H.
[17] Zaidi, M., and A. Amirat. "Forecasting stock market trends by logistic Alyoubi, “ Stock market prediction using machine learning classifiers
regression and neural networks: Evidence from KSA stock market." and social media , news,” J. Ambient Intell. Humaniz. Comput., no.
Int. J. Econ. Commer. Manag 4,vol. IV, no. 6, pp. 220-234, 2016 0123456789, doi: 10.1007/s12652-020-01839-w.
[18] A. Nayak, M. M. M. Pai, and R. M. Pai, “ Prediction Models for Indian [38] S. Mehtab and J. Sen, “A Time Series Analysis-Based Stock Price
Stock Market,” Procedia - Procedia Comput. Sci., vol. 89, pp. 441-449, Prediction Using Machine Learning and Deep Learning Models A
1877, doi: 10.1016/j.procs.2016.06.096. Time Series Analysis-Based Stock Price Prediction Using Machine
[19] H. Karchalkar, A. Jain, A. Singh, and V. Kumar, “ Open Price Learning and Deep Learning Models,” doi:
Prediction of Stock Market using Regression Analysis,” vol. 6, no. 5, 10.13140/RG.2.2.14022.22085/2.
pp. 418-421, 2007, doi: 10.17148/IJARCCE.2017.6578. [39] M. Obthong and N. Tantisantiwong, “ A Survey on Machine Learning
[20] Y. Yaslan, “ Stock daily return prediction using expanded features and for Stock Price Prediction : Algorithms and Techniques,” 2014.
feature selection,” no. 2017, pp. 4829-4840, doi: 10.3906/elk-1704- [40] I. K. Nti, A. F. Adekoya, and B. A. Weyori, “ A comprehensive
256. evaluation of ensemble learning for stock - market prediction,” J. Big
[21] A. You, M. A. Y. Be, and I. In, “ Application of SVM-KNN using SVR Data, 2020, doi: 10.1186/s40537-020-00299-5.
as feature selection on stock analysis for Indonesia stock exchange
Application of SVM-KNN Using SVR as Feature Selection on Stock
Analysis for Indonesia Stock Exchange,” vol. 020207, no. 2018.

Authorized licensed use limited to: California State University Fresno. Downloaded on June 21,2021 at 18:25:10 UTC from IEEE Xplore. Restrictions apply.
View publication stats