Journal of Decision Systems

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/tjds20

Machine learning solutions in the management of

a contemporary business organisation

Leszek Ziora

To cite this article: Leszek Ziora (2020): Machine learning solutions in the management
of a contemporary business organisation, Journal of Decision Systems, DOI:
10.1080/12460125.2020.1848378

To link to this article: https://doi.org/10.1080/12460125.2020.1848378

Published online: 14 Dec 2020.

Submit your article to this journal

Article views: 4

View related articles

View Crossmark data

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

https://www.tandfonline.com/action/journalInformation?journalCode=tjds20
JOURNAL OF DECISION SYSTEMS
https://doi.org/10.1080/12460125.2020.1848378

Machine learning solutions in the management of a

contemporary business organisation
Leszek Ziora
Faculty of Management, Czestochowa University of Technology, Czestochowa, Poland

ABSTRACT ARTICLE HISTORY

The goal of the paper is to present machine learning (ML) solutions Received 11 March 2020
applicability in the management of contemporary business organi­ Accepted 9 June 2020
sation primarily focusing on the improvement of decision-making KEYWORDS
processes and the role of supervised and unsupervised learning in Machine learning; neural
its support. The paper defines and characterises fundamental com­ networks; artificial
ponents of machine learning, provides benefits resulting from its intelligence solutions;
application mainly in business domains as well as reviews the management of a business
applications, case studies and practical examples. The research organisation
study presented in the paper underlines the advantages and dis­
advantages of the mentioned solution.

Introduction
Nowadays, contemporary business organisations put into practice multiple machine (ML)
learning and artificial intelligence solutions. Machine learning and also reinforcement
learning with such frameworks as Cafe, Google tensor flow might play a crucial role in the
improvement of the functionality of enterprise, and it perfects and increase the business
value of the entire decision-making process. What is more such innovative solutions have
a positive impact on the optimisation of business processes and contribute to leveraging
competitive advantage of a particular enterprise in the global market. The vital issue is the
quality of input data, especially big data sets which may constitute the basis for model
training. What is more, the hybrid solutions combine machine learning with data analytics
such as Business Intelligence systems including managerial dashboards, increases the
efficiency of the whole solution.

The notion and development of machine learning concept

An essential characteristic of machine learning is the feature that it „is based on algo­
rithms that can learn from data without relying on rules-based programming” (Pyle & San
Jose, 2015). There can be seen the combining of „AI-based machine learning into analy­
tical systems to create hybrid analysis” (Teich, 2018).
The development and design of artificial intelligence (AI) and later, machine learning is
dated back to 1950s. P. Langley mentions that “in the middle 1960s the importance of

CONTACT Leszek Ziora [email protected] Faculty of Management, Czestochowa University of Technology,

Czestochowa, Poland
© 2020 Informa UK Limited, trading as Taylor & Francis Group
2 L. ZIORA

domain knowledge led to the construction of the first knowledge-intensive systems and
many researchers in pattern recognition began to emphasise algorithmic, numerical
methods contrasting with the heuristic, symbolic methods” (Langley, 1998, pp. 2–3) and
the faster growth could be observed since 1980s. Samudrala provides the definition of
machine learning stating that it “is a set of tools and techniques, in which intelligence is
learned by algorithms and not induced where historical data patterns to extract intelli­
gence is analysed and converted into business rules” (Samudrala, 2018).
E. Alpaydin states that machine learning “is programming computers to optimise
a performance criterion using example data or past experience and also uses the theory
of statistics in building mathematical models where the core task is making inferences
from a sample” (Alpaydin, 2020).
The machine learning solutions are subdivided into supervised learning, semi-
supervised and unsupervised learning and embrace such key components as classifica­
tion, regression, or clustering (Figure 1).
A. Burkov providing a definition of supervised learning states that “in supervised
learning, the dataset is the collection of labelled examples and each element x is called
a feature vector what denotes that it contains a value that describes the example in each
dimension” (Burkov, 2019, pp. 3–4). R. E. Neapolitan and X. Jiang underline the fact that
“supervised learning involves learning a function from a training set and the function
learned is called a model of the underlying system generating the data” (Neapolitan &
Jiang, 2018, p. 89). A. Burkov further states that in case of unsupervised learning, the
dataset is a collection of unlabelled examples where x is a feature vector and claims that

Machine learning (ML)

Supervised learning: Unsupervised learning:

- regression (linear, local, least - cluster analysis (hierarchical, k-

squares), means, neural networks, hiden

- classification (decision trees, Markov model)

Benefits resulting from ML application

in business organization:

- improvement of decision making and

business processes,

- increase of decisions’ business value,

Figure 1. Machine learning application in business organisation.

Source: Author’s study
 JOURNAL OF DECISION SYSTEMS 3

The advantages resulting from machine learning

application in business organization

ability to self-improvement

decision optimization

improvement of business analyses

automations of tasks

increased work efficiency

increased accuracy and efficiency of

decision making process

0% 10% 20% 30% 40% 50% 60% 70% 80%

Figure 2. Advantages resulting from the application of machine learning in a business organisation.
Source: Author’s study

Disadvantages resulting from application of

machine learning in business organization

huge computer power requirement

error susceptibility

neccesity to ensure good data quality

vulnerability to security breach

0% 10% 20% 30% 40% 50% 60% 70%

Figure 3. Disadvantages resulting from the application of machine learning in a business organisation.
Source: Author’s study

the goal of an unsupervised learning algorithm is to create a model that takes a feature
vector x as input and transforms into another vector or into a value that can be used to
solve a practical problem. The same author distinguishes the third category, which is
semi-supervised learning, where “the dataset contains both labelled and unlabelled
examples”. He also identified the fourth category, which is reinforcement learning
where its goal is to learn a policy and is capable of perceiving the state of the environment
as a vector of features”. M. Ganiyu confirms that “regression (with its subdivision, e.g., into
linear, local and least squares) is used to predict continuous numerical data and classifica­
tion is used to predict categorical data sets including k-nearest neighbours” that uses the
4 L. ZIORA

closest data points to that specific data points to predict what category that data point
should belong to (Ganiyu, 2020).
The crucial issue is to ensure appropriate data quality based on which the machine
learning model is trained, and also the size of such training set is significant as well as
whether we deal with qualitative or quantitative data. The huge amount of data (big data)
combined with appropriate selection of machine learning tool and ensuring proper data
quality can have a direct impact on the quality of decisions undertaken by managers of
any business organisation. As far as big data is considered it worth noticing its influence
on the improvement of decisions, in the area of increasing efficiency and efficacy of the
decision-making process (Jelonek et al., 2019, pp. 361–368).
J. G. Carbonell, R. S. Michalski and T.M. Mitchell present objectives of machine learning
and claim that „machine learning is organised around the primary research fields such as
task-oriented studies (called engineering approach), cognitive simulation and theoretical
analysis” (Carbonell et al., 2015). The same authors provide three dimension’s taxonomy of
machine learning research enumerating: classification on the basis of the underlying
learning strategies used, classification on the basis of the representation of knowledge
or skill acquired by the learner, classification in terms of the application domain of the
performance system for which knowledge is acquired. The first classification includes rote
learning, learning from instruction, learning by analogy, from examples and from obser­
vation and discovery (unsupervised learning). The classification according to the type of
knowledge acquired embraces decision trees, parameters in algebraic expressions, formal
grammars, production rules formal logic-based expressions, graphs and networks, taxo­
nomies. The classification by domain of application includes, i.e. natural language proces­
sing, speech recognition, sequence prediction, robotics, medical diagnosis, expert
systems, education, gaming” (Carbonell et al., 2015, p. 13) and many other domains.
Machine learning „has contributed to optimisation, driving the development of new
optimisation approaches” (Sra et al., 2012).
The decision at all level of management can be supported by machine learning
solutions where the crucial is to ensure appropriate data quality. J. de Hoog et al. under­
line the importance of the fact that data quality „can be evaluated based on KPI (Key
Performance Indicators) such as the accuracy of the data, the representation consistency
and completeness (De Hoog et al., 2019).
Y. Roh, G. Heo and S.E. Whang underline the fact that the „major bottleneck in machine
learning is constituted by data collection, which is a critical issue in machine learning
functionality. The cited authors pinpoint the reason for this situation which is mainly the
emergence of new applications which do not possess enough labelled data” (Roh et al., 2019).
In technological terms, the most popular frameworks widely used in deploying machine
learning solutions are Google’s Tensor Flow which is an open-source solution, and it applies
so-called data flow graphs and the second is Caffe developed by Berkeley Vision and Learning
Centre where the user can define the layers with appropriate parameters.

The benefits resulting from a machine learning application

The machine learning (ML) solutions find its application mainly in such domains as decision
support and its improvement, business processes optimisation as well as image recogni­
tion, self-driving cars development, logistics area, image classification in diagnoses in
 JOURNAL OF DECISION SYSTEMS 5

medicine and healthcare area and many more. There exist a possibility to apply a hybrid
approach such as „combination of statistical methods with neural networks, machine
learning, big data solutions, Business Intelligence and Management Information Systems,
sentiment analysis and so on” (Jelonek et al., 2020, pp. 319–327). The sentiment analysis or
opinions mining deploys a wide range of machine learning methods, techniques and tools
especially supervised learning solutions such as decision trees, support vector machines,
neural networks, Bayesian Networks (Ziora, 2016, pp. 234–241).
There can be distinguished multiple benefits resulting from the application of machine
learning solutions in the management of a contemporary business organisation. R.C.
Pozen and J. Ruane indicate three such ML features which are especially applicable to
assets management as: „finding new patterns in existing data sets, making new forms of
data analysable and reduction of the negative effects of human biases on investment
decisions” (Pozen & Ruane, 2019). The AI solutions find its application in contracts
management where in case of “huge volume of procurement contracts the data can be
extracted and clarified what allows to review contracts more rapidly (. . .) and decrease the
potential for contract disputes” (Rich, 2018).
The automated machine learning enables processes automation and finds its applica­
tion mainly in decision making and forecasting areas. It allows for „time savings by
automating data extraction and algorithms, improved accuracy by the elimination of
errors” (AutoML case study, 2020).
ML can be applied in many domains in practice depending on whether it is supervised,
semi-supervised or unsupervised learning B. Marr underlines the most important applica­
tion areas of ML such as data security and especially learning models for malware
detection cloud data access patterns; financial trading in predicting and fast executing
of trades; healthcare domain especially in spotting patterns in case of computer-assisted
diagnosis and understanding risk factors for disease in large populations; marketing
personalisation with recommendations functions; fraud detection; online search; natural
language processing (Marr 2016). Another worth mentioning the machine learning area
of application are early phases of product development supporting the knowledge
extraction in the development processes (Shabestari et al., 2019).

Review of case studies and practical examples

Ai multiple included the list of 22 case studies exemplifying application of automated
machine learning in different industries/companies. The selected benefits may include: in
case of Lenovo which is Technology company in Brazil the DataRobot automated machine
learning solution was used in sales and manufacturing operations and allowed to „increased
accurate predictions from 80% to 87.5% and reduced model creation time from 4 weeks to
3 days. In case of Paypal which is a company providing financial services with the usage of
H2O.ai solution in fraud detection area brought such benefits as improved accuracy by
94.8%, and reduced model training time to under 2 hours. In the case of Imagia company
from Canada, the application of Google Cloud AutoML in healthcare domain allowed for
reduced test processing time from 16 hours to 1 h as well as improved diagnosis results. In
the case of Meredith Cooperation from media and entertainment branch, the usage of
Google Cloud AutoML in content classification improved awareness for future trends and
customer experience. In case of One Marketing company from Denmark, the usage of
6 L. ZIORA

DataRobot solution in the email marketing reduced spam for customers, improved mail
open rate by 14%, improved mail click rate by 24% and increased ticket sales by 83%
(Shabestari et al., 2019). P. Dar characterises the most popular machine learning applications
and uses cases in daily life such as voice assistants recognising speech available in smart­
phones, including Google Assistant, Apple’s Siri. Amazon’s Alexa, Microsoft’s Cortana or
Samsung’s Bixby. These solutions belong to the domain of natural language processing.
Another popular ML applications are connected with object detection, image-enhancing in
cameras, then, e.g., face unlock, identification of people in photographs. This author also
mentions dynamic pricing in transportation in planning routes, then the other use cases in
Google maps, Search and Email (email filtering), Google translate recommendations in
streaming services, e-commerce sites, movie and book review sites. Other applications
embrace customer support queries and chatbots, video surveillance and cybersecurity or
self-driving cars (Dar, 2019). Machine learning is applied in assessing service quality and
performance improvement what is indicated by B. Marr in his case study on Verizon
telecommunication company in the US which is, i.e. a huge telecommunication company
and the Internet provider. The applied machine learning algorithms and big data solutions
which drive business decisions allowed for improvement of existing customer service and to
‘detect manufacturing or production defects in the hardware and software provided by
third parties which the network is built on’ (Marr 2020). As a technological platform, Verizon
uses Hadoop, Spark and Kafka solutions enabling real-time data processing.

Research results
In order to get to know the trends connected with the development of machine learning
solutions, there was conducted opinion survey on the research sample of 50 intramural
students of the Czestochowa University of Technology. The respondents had to indicate
the advantages resulting from the application of a machine learning solution in the area of
management and primarily managerial decisions improvement as well as to show its
disadvantages.
The most frequent answers indicated such benefits as continuous improvement and
increase of accuracy and efficiency of the decision-making process (75%), increased work
efficiency (70%), automation of tasks (65%), improvement of business analyses (e.g.,
market, financial analyses) (60%) emphasising fast identification of trends and patterns
related to sales of products and services, the possibility to optimise business decisions
(e.g., in case of regression application) (40%). The other responses involved the ability to
improve over time (e.g., algorithm’s self-improvements) (30%) - presented in Figure 2.
The disadvantages which were indicated by the respondents included security issues,
the vulnerability of machine learning solutions to attacks (60%), the necessity to ensure
proper data quality for training sets (55%), susceptibility to errors (45%), the requirement
of enormous computing power (45%) - presented in Figure 3.

Conclusion
Nowadays, contemporary enterprises utilise machine learning to a greater extent than
any time before. They apply supervised and unsupervised solutions in order to support its
business functionality, especially to improve decision making support and everyday’s
 JOURNAL OF DECISION SYSTEMS 7

business operations. Multiple benefits resulting from its application may lead to gaining
competitive advantage in the global market as well as it contributes to increasing
business value. To sum up, in the management of modern business organisation the
most important benefits of ML application embrace increased accuracy and efficiency of
the decision-making process, increased work efficiency as well as improvement of con­
ducting multiple business analyses, optimisation of decisions and automation of tasks. As
the future development determinants of AI and especially machine learning solutions can
be indicated and observed a growing meaning of hybrid solutions combining different
methods, techniques and tools of data processing, utilising big data analyses and proces­
sing algorithms.

Disclosure statement
No potential conflict of interest was reported by the author.

References
Alpaydin, E. (2020). Introduction to machine learning. The MIT Press.
Automated Machine Learning. (2020). 22 AutoML case studies/examples: In-depthguide. https://blog.
aimultiple.com/automl-case-studies/
Burkov, A. (2019). The hundred-page machine learning book (pp. 3–4). Amazon.
Carbonell, J.G., Michalski, R.S., & Mitchell, T.M. (2015). Machine learning. An artificial intelligence
approach. Morgan Kaufman Publishers.
Dar, P. (2019, July 15) Popular machine learning applications and use cases in our daily life. https://
www.analyticsvidhya.com/blog/2019/07/ultimate-list-popular-machine-learning-use-cases/
De Hoog, J., Mercelis, S., & Hellinckx, P. (2019). Improving machine learning-based decision-making
through inclusion of data quality. CEUR-WS.org/Vol-2491/short14.pdf
Ganiyu, M. (2020). How machine learning works. An introduction into machine learning terminologies
centred around regression and classification. https://towardsdatascience.com/how-machine-
learning-works-b801303b1a19
Jelonek, D., Mesjasz-Lech, A., Stępniak, C., Turek, T., & Ziora, L. (2020). The artificial intelligence
application in the management of contemporary organization: Theoretical assumptions, current
practices and research review. Lecture Notes in Networks and Systems 69 (pp. 319–327). Springer.
Jelonek, D., Stępniak, C., & Ziora, L. (2019). The meaning of big data in the support of managerial
decisions in contemporary organisations: Review of selected research. In K. Arai, S. Kapoor, &
R. Bhatia (Eds.), Future of Information and Communication Conference (FICC 2018, Singapore),
Advances in Information and Communication Networks (pp. 361–368). Springer.
Langley, P. (1998). Elements of machine learning (pp. 2–3). Morgan Kaufman Publishers.
Marr, B. (2016). The top 10 AI and machine learning use cases everyone should know about. https://
www.forbes.com/sites/bernardmarr/2016/09/30/what-are-the-top-10-use-cases-for-machine-
learning-and-ai/#3f2e73c394c92017
Marr, B. (2020). The amazing ways Verizon uses AI and machine learning to improve performance.
https://bernardmarr.com/default.asp?contentID=1511
Neapolitan, R.E., & Jiang, X. (2018). Artificial intelligence with an introduction to machine learning. CRC
Press.
Pozen, R.C., & Ruane, J. (2019, December) What machine learning will mean for asset managers.
Harvard Business Review. https://hbr.org/2019/12/what-machine-learning-will-mean-for-asset-
managers
Pyle, D., & San Jose, C. (2015, June) An executive’s guide to machine learning. McKinsey Quarterly.
Rich, B. (2018, February) How AI is changing contracts. Harvard Business Review. https://hbr.org/
2018/02/how-ai-is-changing-contracts
8 L. ZIORA

Roh, Y., Heo, G., & Whang, S.E. (2019, October). A survey on data collection for machine learning:
A big data – AI integration perspective. IEEE Transactions on Knowledge and Data Engineering, PP
(99), 1. https://doi.org/10.1109/TKDE.2019.2946162
Samudrala, S. (2018). Machine intelligence. Demystifying machine learning, neural networks and deep
learning. Notion Press.
Shabestari, S., Herzog, M., & Bender, B. (2019). A survey on the applications of machine learning in
the early phases of product development. Proceedings of the Design Society: International
Conference on Engineering Design, 1(1), 2437–2446. https://doi.org/10.1017/dsi.2019.250
Sra, S., Nowozin, S., & Wright, S.J. (2012). Optimisation for machine learning. The MIT Press.
Teich, D.A. (2018, July 6). Management AI: Types of machine learning systems. https://www.forbes.
com/sites/davidteich/2018/07/06/management-ai-types-of-machine-learning-systems
/#51634c6432fb
Ziora, L. (2016). The sentiment analysis as a tool of business analytics in contemporary organisations.
Economics Studies. University of Economics in Katowice Research Papers, no 281(pp. 234–241).