Data Science and Management 7 (2024) 79–88

Contents lists available at ScienceDirect

Data Science and Management

journal homepage: www.keaipublishing.com/en/journals/data-science-and-management

Review article

Using multi-criteria decision-making and machine learning for football

player selection and performance prediction: a systematic review
Abdessatar Ati a, Patrick Bouchet b, Roukaya Ben Jeddou a, *
a
Faculty of Legal, Economic, and Management Sciences, University of Jendouba, Jendouba, 8189, Tunisie
b
University of Bourgogne Franche-Comte’, Universit
e de Bourgogne, Dijon, 21078, France

A R T I C L E I N F O A B S T R A C T

Keywords: Evaluating and selecting players to suit football clubs and decision-makers (coaches, managers, technical, and
Multi-criteria decision-making medical staff) is a difficult process from a managerial-financial and sporting perspective. Football is a highly
Machine learning competitive sport where sponsors and fans are attracted by success. The most successful players, based on their
Football player selection
characteristics (criteria and sub-criteria), can influence the outcome of a football game at any given time.
Managerial-financial and sporting performance
Consequently, the D-day of selection should employ a more appropriate approach to human resource manage-
ment. To effectively address this issue, a detailed study and analysis of the available literature are needed to assist
practitioners and professionals in making decisions about football player selection and hiring. Peer-reviewed
journals were selected for collecting published papers between 2018 and 2023. A total of 66 relevant articles
(journal articles, conference articles, book sections, and review articles) were selected for evaluation and analysis.
The purpose of the study is to present a systematic literature review (SLR) on how to solve this problem and
organize the published research papers that answer our four research questions.

1. Introduction decision-makers constantly search for the most efficient technique to

identify outstanding players and build elite teams (Katzenbach and
Soccer has been one of the world’s most influential and popular sport Smith, 1993). One strategic decision decision-makers should made is
for over a century. For the past thirty years, it has also become one of the choosing players for certain football match. The starting line-up com-
most gainful industry, with a significant managerial and financial impact prises 11 players among the 28 available to coaches to play at the
on sponsorship, broadcasting, and transfers of players (Müller et al., beginning of a game. This decision influences the success of a team in
2017; Rohde and Breuer, 2017). Many professional football clubs are achieving financial and sports performance (Carmichael et al., 2011;
complex businesses that exclusively deal with financial issues (Morrow Rohde and Breuer, 2016; Szymanski and Smith, 1997). This strategic
and Howieson, 2014). However, Katzenbach (2009) explored the process is part of a conventional approach that does not refer to a thor-
counter-intuitive features comprising high-performing teams, i.e., ough scientific method; it is carried out mainly by intuition, common
selecting team members based on their skills rather than their compati- sense, and an uncertain methodology (Boon and Sierksma, 2003; Dadelo
bility. Club success is determined by the score on the field, which depend et al., 2014; Purwanto et al., 2018; Salles et al., 2019).
on the line-up of the chosen players. Therefore, the selection should be This process is complex and challenging to manage because it must
based on a more relevant human resource management strategy. consider a large number of qualitative and quantitative criteria and sub-
In a football match, in addition to preparing the team to participate in criteria that describe the performance and position of each player
competitions, the role of decision-makers (managers, coaches, technical, (Arnason et al., 2004; Nasiri et al., 2018; Tavana et al., 2013). To solve
and medical staff) is to hire and select players to form the team line-up by this issue, decision-makers refer to modern methods used in other areas
relying, technical, and tactical factors (Carmichael et al., 2000). Most such as health, software packages, transportation, and environment
importantly, the goal of the player selection process for a particular team which consist of multi-criteria decision-making (MCDM) methods.
is to choose the most suitable player for a particular playing position and MCDM approaches could be beneficial and useful in assisting
role (Dezman et al., 2001; Trninic et al., 2008). Accordingly, decision-makers, providing more objective decisions, especially when

Peer review under responsibility of Xi'an Jiaotong University.

* Corresponding author.
E-mail address: [email protected] (R. Ben Jeddou).

https://doi.org/10.1016/j.dsm.2023.11.001
Received 4 July 2023; Received in revised form 4 November 2023; Accepted 6 November 2023
Available online 10 November 2023
2666-7649/© 2023 Xi'an Jiaotong University. Publishing services by Elsevier B.V. on behalf of KeAi Communications Co. Ltd. This is an open access article under the
CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
A. Ati et al. Data Science and Management 7 (2024) 79–88

combined with machine learning algorithms. Historical research on conference proceedings related to our research topic.
football player selection and performance prediction oscillated between (2) Research questions: After determining the purpose and motivation
the traditional coaching experience and data-driven approaches. MCDM of our research, the SLR proposes a process to answer the following
methods have gained traction in recent years because they provide a research questions (RQs).
multi-criteria perspective for informed decision making (Dadelo et al.,
RQ1. What is the contribution of the literature to the review and study
2014). Simultaneously, machine learning based on large datasets may
of the relationship between managerial-financial and sports perfor-
provide granular predictions of player performance using models, such as
mance? Financial and managerial successes are not the main objective,
neural networks and random forests (Azeman et al., 2021). The potential
and football performance remains the primary goal. These two goals are
fusion of MCDM for selection and machine learning for predictive in-
mutually linked, and each determines the other. Football and financial
sights remains underexplored area of research. Addressing this research
success depend on hiring strategies and players’ lineups.
gap is critical, and motivates this systematic literature review (SLR).
This study provides a research review to improve the evaluation and RQ2. What are the main criteria and sub-criteria for football players? In
selection process of football players. Therefore, this study aims to review addition to preparing the team to participate in competitions, the coach’s
and organize the literature on decision-making analysis and machine role is to select the starting range among the players in the hired stock,
learning by performing an SLR. The remainder of this paper is as orga- based on factors such as physics, technique, tactics and behavior.
nized as follows: The following section presents the research method and
RQ3. The effectiveness of sports clubs is influenced by players’ and
details of the protocol used to plan and conduct the proposed research.
coaches’ talents. The most crucial elements of a sports organization are
Section 3 shows the analysis results to answer the research questions
the athletic resources that constitute its human capital. MCDM ap-
related to the topic. Section 4 discusses and concludes.
proaches could be of interest to support decision makers and provide
more objective decisions. This study investigates the applicability and
2. Research method
utility of MCDM approaches in the selection process for football players.

The present study follows an SLR protocol based on Kitchenham’s RQ4. How can we build a hybrid model combining MCDM approaches
guidelines (Kitchenham and Charters, 2007). Fig. 1 shows three and machine learning algorithms by integrating all technical and
sequential plans reflecting the components of the SLR: planning, con- managerial-financial variables to help decision makers select the best
ducting, and reporting the review. players for a game?
(3) Search process: For planning the SLR process, an appropriate
2.1. Planning the review mechanism must be followed to ensure the retrieval of relevant studies
from the selected digital libraries. A systematic mechanism is used to
The purpose of the planning phase is to specify the following steps: collect the most pertinent articles by formulating a set of specific key-
determine the objectives of a review, set the research questions (four words. These data are then used to search peer-reviewed digital libraries
research questions are selected), perform a search process using key- for research articles. In the proposed SLR, several keywords are related to
words for downloading the relevant papers from the selected digital li- the analysis and applicability of multi-criteria decision models and ma-
braries, and identify the selection criteria. chine learning in sports, mainly football, based on the above research
(1) Objectives: This study examines research out in decision-making questions. However, following a search strategy that identifies the most
and machine learning when evaluating and selecting players, consid- relevant documents is essential. To ensure the quality of the articles, the
ering their playing attributes (physical, technical, tactical and behavioral search focused on various electronic databases: IEEExplore, Spring-
criteria) and how this process influences managerial-financial and erLink, ScienceDirect, Taylor and Francis, and Wiley Online Library.
sporting performance. Player selection is a significant challenge for all These are the most frequently used libraries and are known to publish
clubs, involving multiple criteria that should be evaluated simulta- high-quality materials. The authors use specific keywords to search for
neously. Therefore, the process of hiring and line-ups must respect many relevant articles in these libraries. The keywords are selected based on
sports and financial considerations. We aim to assess the main criteria by the proposed research questions. A search string S is formulated to
which decision-makers set line-ups and predict the impact of their choice interact with the libraries (Table 1). The first attempt in this search
on sport and financial performance. To this end, decision-making process has yielded the following results: IEEE →802, ScienceDirect
methods may help decision-makers recruit and select the best players, →1432, Springer →398, Taylor and Francis →221, and Wiley Online
and machine learning algorithms would be appropriate for predicting Library →137. To refine the results, we assigned a subset of keywords
sporting and financial performances following this selection. Therefore, (S1, S2, S3, S4) to each research question. Given the large amount of
we provide an SLR of books and papers published in journals and information generated by different digital libraries, they cannot be

Fig. 1. Systematic literature review (SLR) general process.

80
A. Ati et al. Data Science and Management 7 (2024) 79–88

included in the proposed SLR. A process for selecting criteria, followed by Table 2
the analysis and assessment of each article, should be developed. Selection criteria of the relevant papers.
(4) Selection criteria: Selection criteria are specified to refine the Inclusion criteria Exclusion criteria
search results, namely, the inclusion and exclusion criteria. Given the
Published between 2018 and 2023 Published outside the 2018-2023 period
large volume of published research, it is crucial to establish boundaries English language Other language than English
for a thorough and detailed literature review. We focus on the period Provide good knowledge of Content of the paper not providing
between 2018 and 2023, sufficiently long to observe critical trends the formulated research questions satisfactory information of the main
related to our research topic. Table 2 summarizes the essential inclusion topic and research questions
Content of the paper provide Informal literature surveys
and exclusion criteria for the proposed SLR. satisfactory information of the
main topic and research questions
- Duplicate papers
2.2. Conducting the review

Once the research protocol is completed, we may proceed with the

review and apply what we have specified in the planning phase. The
process is iterative (Fig. 2), and the first primary study is performed on
search string S. For greater relevance, we use a search string for each
research question (Table 3). Second, the inclusion/exclusion criteria are
applied to a subset of primary studies (Table 4). Third, the remaining
studies are analyzed in depth and filtered by title and paper keywords,
abstracts, and content (Table 5). Each digital library is searched sepa-
rately. Finally, the articles collected from the five databases are imported
into the Zotero reference management software (Zotero_community,
2023) to check for duplicates. Consequently, only 66 relevant documents
(journal articles, conference papers, and book sections) that fit our topics
and answer our research questions are identified. Figs. 3 and 4 show the
research trends in the selected period and the percentage of digital li-
braries, respectively.
(5) Study quality assessment: The study quality phase aims to assess
the quality of the primary studies. Each article is evaluated separately to
ensure that it targets the research objective. We refer to the data analysis,
figures, tables, and appendixes for analyzing and collecting data from
each paper.
(6) Data extraction: The data extraction process extracts items from
the 66 included studies and collects all information required to answer
the research questions related to our topics. An extracted data form
(Table 6) is developed to collect data items from each primary study. This
process is the object of a meticulous examination by the authors to verify
the consistency of the inclusion/exclusion decisions and ensure the Fig. 2. Review process.
quality of the selected studies (Table 7).
We also consider books and book sections on sports management and
decision-making (Dobson and Goddard, 2001; Ishizaka and Nemery,
2013; Katzenbach, 2009; Katzenbach and Smith, 1993; Saaty, 1996; Table 3
Saaty and Vargas, 1982; Tzeng and Huang, 2011). We thoroughly review Search process based on research questions.
articles that answer three or more research questions. However, by
Digital libraries S1 S2 S3 S4 Total
focusing on these articles, we delve into studies with higher levels of
IEEE 52 33 3 13 101
complexity and depth. It is essential to note that our approach does not
ScienceDirect 614 64 29 6 713
diminish the importance of articles that address fewer research ques- Springer 233 94 50 19 396
tions. Such studies can be valuable and insightful and often shed light on Taylor and Francis 147 86 3 33 269
certain aspects of the topic. We also try to find articles on subjects not Wiley Online Library 288 14 2 29 333
from the period selected for the proposed SLR, to avoid omitting any
relevant and useful articles. To this end, we use the Connected Papers
website (Tarnavsky Eitan et al., 2023), a visual tool that helps
researchers.
Table 1 We identify academic articles relevant to our research areas. Some of
List of keywords selected for search process.
Research Keywords for search process
questions Table 4
The whole topic S ¼ (“multi-criteria decision making”) AND (“machine learning”) Search process with inclusion/exclusion criteria.
AND (“selection”)
Digital libraries Journal Conference Book Total
RQ1 S1 ¼ (“football” OR “soccer”) AND (“management” OR
papers papers section
“managerial”) AND (“financial sporting performance”)
RQ2 S2 ¼ (“football” OR “soccer”) AND (“player”) AND (“criteria” OR IEEE 1 21 - 22
“attribute”) ScienceDirect 54 - 3 58
RQ3 S3 ¼ (“multi-criteria decision making”) AND (“selection Springer 36 4 7 47
problems”) Taylor and Francis 32 - - 32
RQ4 S4 ¼ (“multi-criteria decision making”) AND (“prediction”) AND Wiley Online 22 - 1 23
(“machine learning”) AND (“football player selection”) Library

81
A. Ati et al. Data Science and Management 7 (2024) 79–88

Table 5 financial performance and sports performance (RQ1). Section 2 describes

Filtering papers by title and paper keywords, abstract, and contents. the main criteria and sub-criteria related to the football players involved
Digital libraries Title and paper keywords Abstract Content in the selection problem (RQ2). Section 3 presents the decision-making
methods used to assist decision-makers in solving selection problems,
IEEE 30 28 22
ScienceDirect 31 28 20 referring to player criteria and sub-criteria (RQ3). Section 4 highlights
Springer 10 8 7 hybrid models that integrate all technical and economic variables to help
Taylor and Francis 11 11 9 decision makers select the best players for a game (RQ4).
Wiley Online Library 10 10 8 (1) RQ1: What is the contribution of the literature to the review and
Total 92 85 66
study of the relationship between managerial-financial and sporting
performance? A team’s success may be defined by its results and how
managers organize and manage a football club. Several parameters must
be considered during the decision-making analysis. According to Rohde
and Breuer (2018), the managerial-financial process is an investment
activity in a club organization where footballers represent a major asset.
Furthermore, Flegl et al. (2018) highlight that selecting personnel in an
organizations is usually tricky. The process becomes even more compli-
cated when selection occurs in complex organizations where different
areas can work toward multiple conflicting objectives. Contradictions
between a higher income, promoting sports, and providing a good show
can be targets. As many areas may be involved in the selection process, it
can become complex and challenging. Therefore, Samur (2018) suggests
that football organizations can be broken down into critical processes
that can be managed easily. The study underlines the importance of
Fig. 3. Research trends (2018–2023). financial transfers generated by sporting success to all club activity areas
in a balanced manner.
Gonçalves et al. (2020) measured economic-financial performance
using net revenue and other revenue sources such as broadcasting,
matchday, commercial, and player sales. Using panel data analysis, they
find that team performance positively and significantly affects the clubs’
economic and financial performance in the subsequent year. Similarly,
Barros (2006) examined the relationship between football success,
financial management, and performance of clubs. The study concludes
that the financial crisis in the Portuguese first division may be imputed to
mismanagement. Ribeiro and Lima (2012) analyze the relationship be-
tween the distribution of players’ salaries and the effectiveness of clubs,
and how sports performance and financial performance can mutually
impact each other. Clubs should structure salaries to achieve optimal
salary ranking to induce player effort. Alaminos et al. (2020) emphasized
that the financial success of football clubs has become essential in
ensuring their long-term solvency and viability. Therefore, theoretical
and practical considerations reveal that financial, sporting, and business
Fig. 4. Contribution share of each database. elements must be included to properly manage the club’s finances and
analyze potential revenues. They use neural networks to study the
financial success of European football clubs. Their results show that a
these documents are considered in the reporting phase even if they do not
club’s financial performance is influenced by its liquidity, leverage, and
fall within the selected period.
sports performance.
Seungbum and Ross (2012) identified the decision-making factors
3. Results: reporting the literature review
affecting sports sponsorship in a global market context. Sports teams with
star players or coaches increase the club’s attractiveness. Trequattrini
The reporting phase is designed to analyze and implement a literature
et al. (2012) propose a method to determine the economic value of
review that addresses the research questions related to our topics. This
professional football players in the event of a purchase or sale. Dimi-
phase addresses 66 studies from the conducting process and other works
tropoulos and Scafarto (2021) examined the relationship between the
related to the main papers. Section 1 reports the contribution of the
athletic and financial performance of Greek football clubs. They deduced
literature to the study of the relationship between managerial and
that sports and financial performance are linked through investments in
players. Frick and Simmons (2008) established a positive correlation
between spending on team payroll (managerial compensation) and team
Table 6
Data extraction form.
performance. Using stochastic analysis, they demonstrated that coaching
and playing inputs contribute to team success in the league. Boon and
Data item Description
Sierksma (2003) discussed in their model how to support the design of
Ref. Reference ID/number of the paper under year of publication optimal teams and assesses the value-added of a team’s potential.
Study year Year of publication New team members in football and volleyball support the idea that
RQ1 Does the paper relate to our RQ1?
RQ2 Does the paper relate to our RQ2?
hiring new players to improve a team’s quality is one of the difficulties in
RQ3 Does the paper relate to our RQ3? sports and human resources management.
RQ4 Does the paper address RQ4? Based on these studies, we conclude that financial and sporting per-
Case study Does the study apply any case? formances are two major factors in club success; hence, several playing
Database Does the paper include dataset from any database?
and non-playing criteria should be analyzed to improve the player

82
A. Ati et al. Data Science and Management 7 (2024) 79–88

Table 7
Summary of the reviewed literature.
Authors Year RQ1 RQ2 RQ3 RQ4 Case study Data source N

Xia et al. (2018) 2018 ✓ - - - ✓ National Football League & National Basketball Association 1
Blanco et al. (2018) 2018 ✓ - ✓ ✓ ✓ Spanish ACB Basketball League 2
Lepschy et al. (2018) 2018 ✓ - - - - - 3
Flegl et al. (2018) 2018 ✓ ✓ ✓ ✓ ✓ Sofifa 4
Danisik et al. (2018) 2018 ✓ ✓ ✓ - - - 5
Purwanto et al. (2018) 2018 - ✓ ✓ ✓ ✓ Interviews 6
Parida et al. (2022) 2018 ✓ ✓ - - - - 7
Rohde and Breuer (2018) 2018 ✓ - - - ✓ Big five clubs owners from 2003 to 2014 8
Samur (2018) 2018 ✓ - - - - - 9
Hervert-Escobar et al. (2018) 2018 ✓ ✓ - - ✓ FIFA world cup 2018 10
Wadjdi et al. (2018) 2018 - - ✓ - - - 11
Buhagiar et al. (2018) 2018 ✓ - - - - - 12
Galariotis et al. (2018) 2018 - - ✓ ✓ - - 13
Chatterjee et al. (2018) 2018 - - ✓ ✓ ✓ ABP (Media company in India) 15
Farshidi et al. (2018) 2018 - - ✓ ✓ ✓ Software companies 16
Nasiri et al. (2018) 2018 ✓ ✓ ✓ ✓ ✓ Questionnaire 17
Chen et al. (2018) 2018 - - ✓ ✓ ✓ Location Selection 18
Lepschy et al. (2018) 2018 - ✓ ✓ ✓ ✓ American/Canadian basketball league NBA 19
Bhat et al. (2019) 2019 ✓ - - ✓ - - 20
Shahriar et al. (2019) 2019 - ✓ ✓ ✓ ✓ Bangladesh Soccer Team 21
Mohanta et al. (2019) 2019 - - ✓ ✓ - - 22
Nsolo et al. (2019) 2019 - - - ✓ ✓ WhoScored - Opta 23
Ouahli and Cherkaoui (2019) 2019 - - ✓ ✓ ✓ - 24
Purwanto et al. (2018) 2019 - - - ✓ - Wysout (soccer-logs) 25
Patnaik et al. (2019) 2019 - ✓ - - ✓ OPTA 26
Saeed et al. (2019) 2019 - - ✓ - - - 27
Gu et al. (2019) 2019 ✓ - - ✓ Major League Soccer from 2004 to 2015 28
Gomez-Gonzalez et al. (2019) 2019 - - ✓ ✓ - - 29
Alaminos et al. (2020) 2020 ✓ - - - - - 30
Cacho-Elizondo (2020) 2020 - ✓ - - - - 31
Gonçalves et al. (2020) 2020 ✓ - - - ✓ Brazilian sports clubs from 2013 to 2017 32
Yang et al. (2020) 2020 - ✓ ✓ ✓ - - 33
Valenti et al. (2020) 2020 ✓ - - - ✓ Data from 55 UEFA members from 2011 to 2017 34
Salih et al. (2020) 2020 - ✓ ✓ ✓ ✓ Data from (Raheela et al., 2016) 35
Di Simone and Zanardi (2020) 2020 ✓ - - - ✓ European football companies (59 firms from 2013 to 2018) 36
Romero et al. (2021) 2021 - ✓ ✓ ✓ ✓ 2018 European Men’s Handball Championship 37
Bai and Bai (2021) 2021 - ✓ - - - - 38
Manish. et al. (2021) 2021 ✓ ✓ - - - - 39
Azeman et al. (2021) 2021 - ✓ ✓ ✓ ✓ English Premier League Season 2005–2006 40
Rahman and Asadujjaman (2021) 2021 - - ✓ ✓ - - 41
Wieckowski and Watr
obski (2021) 2021 - - ✓ ✓ ✓ Formula 1 42
Zhao et al. (2021) 2021 - - - ✓ ✓ Sofifa 43
Aydemir et al. (2021) 2021 ✓ ✓ ✓ - - - 44
Li et al. (2021) 2021 ✓ ✓ - - ✓ National Basketball Association 2012 to 2015 45
Fry et al. (2021) 2021 ✓ - - - - - 46
Kaczynska et al. (2021) 2021 - - ✓ ✓ ✓ ratings.fide.com (chess) 47
Dimitropoulos and Scafarto (2021) 2021 ✓ - - - - - 48
Garcia-delBarrio and Agnese (2022) 2022 ✓ - - - ✓ Premier League, La Liga, Serie A, and Ligue 1, 2010–2019 49
Steve Arrul et al. (2022) 2022 ✓ - - - ✓ FIFA 2019 50
Morciano et al. (2022) 2022 ✓ ✓ - - - - 51
Rose et al. (2022) 2022 ✓ - - - - - 52
Jenifer and Sundarrajan. (2022) 2022 - - ✓ ✓ - - 53
Datta and Rudra (2022) 2022 ✓ ✓ ✓ - - - 54
Parida et al. (2022) 2022 ✓ ✓ ✓ - ✓ Football Club Barcelona 55
Hu and Fu (2022) 2022 - - - ✓ ✓ Premier League & La Liga data 56
Karakaya et al. (2022) 2022 - ✓ ✓ - - - 57
Mahmood et al. (2021) 2022 ✓ - - - ✓ National Basketball Association 59
Wieckowski and Watr
obski (2021) 2022 - - ✓ ✓ ✓ Randomly generated decision matrix 60
Ribeiro et al. (2022) 2022 ✓ ✓ - - ✓ Portuguese First League from 2016 to 2018 61
Manju and Philip (2023) 2023 - - ✓ ✓ ✓ Dataset of cricket matches 2008–2022 62
Peddii and Jain (2023) 2023 - ✓ ✓ ✓ ✓ Historical data 63
Athish V et al. (2023) 2023 - - ✓ ✓ ✓ kaggle.com & Sofifa.com. 64
Joe Anand et al. (2023) 2023 - - ✓ - - - 65
McHale and Holmes (2023) 2023 ✓ ✓ ✓ - ✓ transfermarkt.com, Instat, sofifa.com 66

selection process. provide an exponential number of datasets for big data research related
(2) RQ2: What are the main criteria and sub-criteria for selecting to football analytics, focusing on individual player characteristics
football players? The selection of a football player includes many per- (Cacho-Elizondo, 2020; Raheela et al., 2016). In selecting the best player,
formance attributes, also known as criteria, such as technical, physical, Nikjo et al. (2015) defined six criteria after questioning ten experts and
mental, and behavioral attributes that have emerged from Big Data three decision-makers. These criteria include technical and tactical skills,
technologies. Due to the increasing amount and different types of sports experience of professional play, average number of goals scored per
data, Big data has become a challenge (Bai and Bai, 2021). Specialized game, ability to coordinate with the team, moral and behavioral features,
companies such as OPTA, FIFA, and the websites of the official leagues and social prestige.

83
A. Ati et al. Data Science and Management 7 (2024) 79–88

Ozceylan (2016) provided a set of criteria based on the position of criteria different efficiency indices. According to Ozceylan (2016), the
players within a football team: Goalkeeper, Fullback, Central Defense, top-performing players in Turkish football clubs are selected for inclusion
Central Midfield, Winger and Forward Center. The study describes a set in the team. In the first phase, the AHP attributes weight to each criterion
of specific criteria for goalkeepers, such as one-on-ones and aerial abili- based on the player’s position. In the second phase, a scalar is established
ties, and general criteria for all positions, such as anticipation, agility, from zero to one, and an integer linear programming model is developed
and first touch. Purwanto et al. (2018) discussed a selection method for using the former weights of the player attributes. Dey et al. (2011)
starting a line-up based on the physical, technical, tactical, and mental contented that the performance of fast-bowlers and spinners of Cricket
criteria. Sub-criteria such as agility, dribbling, ball control, and positional Indian Premier League is based on their economy rate, bowling average,
play, have also been studied to develop a system for line-up selection, and bowling strike rate, using AHP-TOPSIS and AHP-COPRAS (complex
where technical ability and physical strength are two most significant proportional assessment) approaches. Bozbura et al. (2008) rank six
factors. Qader et al. (2017) presented a methodology for evaluating and national basketball association (NBA) players according to their prox-
ranking football players based on anthropometric, fitness, and ability imity to the ideal solution given by the TOPSIS method. Dadelo et al.
tests. Tavana et al. (2013) proposed 18 evaluation criteria (e.g. heading, (2014) developed an integrated model of TOPSIS and expert judgment to
jumping, and reading the game) that are usually considered by coaches ensure greater efficiency in the assessment, rating, and selection of 18
when making team selections. The selection process includes four de- basketball players. Nasiri et al. (2018) proposed an integrated approach
fenders, four midfielders, and two forwards. Nasiri et al. (2018) exam- combining MCDM analysis and mathematical programming to support
ined several criteria and sub-criteria considered necessary by experts and decision-makers through the process of building a football team. Nikjo
the literature for evaluating football players. The study also describes the et al. (2015) selected the best players from a sports team. AHP prioritizes
existing relationships between different sub-criteria. For independent each criterion, and extended TOPSIS is applied for weighing
criteria such as passing, dependent criteria exist, such as creativity, decision-makers and ranking alternatives. Purwanto et al. (2018)
reading the game, reaction, speed, and both feet. It is worth noting that developed a system to select the starting line-up of football players based
the talents of players and coaches influence the effectiveness of sports on the AHP method, in which technical ability and physical strength are
clubs. The essential elements for a sports organization are the athletic the two critical factors for line-up selection. Raheela et al. (2016) pro-
resources that constitute its human capital. Therefore, many authors posed a method for evaluating and ranking 24 football players in the
place players at the center of their studies. Therefore, both player same team using MCDM, where players are selected according to several
recruitment and team line-up selection are critical for maximizing profits physical fitness indicators, such as 30 m speed running. Ballı and Kor-
and victory (Carmichael et al., 2000, 2011; Charnes et al., 1978). ukog lu (2014) applied a fuzzy multi-attribute decision-making algorithm
Consequently, sports management requires decision-making based on based on the fuzzy-AHP and TOPSIS methods to develop a decision
scientific research that supports and informs decision-makers of the most support framework for selecting eligible basketball players in Türkiye.
effective strategies to improve team performance and profitability. The findings of this SLR contribute to a deeper understanding of several
(3) RQ3: What are the existing decision-making methods in which MCDM methods used in sports.
sports decision makers compare the usefulness, results, and validity of (4) RQ4: How can we create a hybrid MCDM model that incorporates
several MCDM methods? Several MCDM methods have been reported in all technical and economic variables to help decision-makers select the
the literature. Following Ishizaka and Nemery (2013), these methods are best players for a game? Our model leverages machine learning algo-
classified into three main categories. rithms to predict players’ performances, encompassing both sporting and
financial aspects, by analyzing historical data sourced from official club
 Full aggregation approach: this method assumes compensable scores. websites. Subsequently, a set of criteria and sub-criteria is assessed and
The analytic hierarchy process (AHP) implemented by Saaty and weighted utilizing a MCDM technique like AHP. This weighted evalua-
Vargas (1982) and the analytic network process (ANP) extended from tion is then applied to rank football players for inclusion in the line-ups
the AHP to consider the interdependent relationships between using a suitable MCDM methodology. This hybrid model allows for a
criteria and sub-criteria belong to this approach (Saaty, 1996). complete analysis by integrating the strength of machine learning algo-
 Outranking approach: this method assumes that a good score cannot rithms to predict criteria and sub-criteria for decision-making models.
compensate for a bad score. Methods belonging to this approach This research question focuses on studies (samples) that construct a
include the elimination and selection from the reality (elimination et model to select or rank the best players (Table 9).
choix traduisant la r
ealit
e, ELECTRE) (Roy, 1968) and the preference
ranking organization method for enrichment evaluations (PROM- 4. Discussion and conclusion
ETHEE) (Mareschal et al., 1984).
 Goal or reference level approach: this method defines a goal for each This SLR allowed us to delve deeper into the proposed study area and
criterion, and then identifies the options closest to the ideal reference analyze the literature concerning four research questions. This analysis
level or goal. This approach includes the technique for order prefer- focused on the following.
ences by similarity to an ideal solution (TOPSIS) proposed by Hwang
and Yoon (1981) to identify the best alternative based on the concept  First, we examined the relationship between player performance and
of the compromise solution, and data envelopment analysis (DEA) football clubs’ managerial and financial performance.
formulated by Charnes et al. (1978) to measure the performance of  Second, the main criteria and sub-criteria related to football players’
decision-making units (DMUs) that convert multiple inputs into positions on the pitch were included in the selection process. In this
multiple outputs. step, we noticed a lack of publications on this topic. In addition,
previous articles did not address all criteria, sub-criteria, and the
This research aims to review articles demonstrating the use of mul- correlations except for (Tavana et al., 2013).
tiple criteria decision-making methods in sports, particularly in football  Third, decision-making methods that could be available to decision-
player selection (Table 8). Mavi et al. (2012) measured the sporting makers for solving the player selection problem were identified.
performance of 18 professional football teams in the German Bandesliga  Finally, studies dealing with the construction of models, case studies,
by collecting data for the season 1999/2000. Seungbum and Ross (2012) and databases for solving the selection of football players were
ranked the criteria and sub-criteria to identify the decision making fac- examined.
tors in sports sponsorship in the global market context. Blanco et al.
(2018) applied a multi-criteria outranking methodology to the Spanish The proposed study used the systematic literature protocol and
ACB Basketball League, using as alternatives the potential players and as guidelines by Kitchenham et al. (2007). Data published between 2018

84
A. Ati et al. Data Science and Management 7 (2024) 79–88

Table 8
Summary of studies related to our RQ3.
Methods/references Category Application Goal Criteria (sub-criteria)

Fuzzy-  Goal or reference Predicting the chances of Predicting the chances of overtaking Drivers’ gap; tyres; distance; pit stop; driver
COMET level overtaking during pit stops in experience; team ranking
Wieckowski and Formula 1 races by using fuzzy
Watr
obski (2021) combined measurement
techniques
Fuzzy-AHP  Full aggregation Selecting team performance in Team performance Individual performance (personal factors,
Ouahli and Cherkaoui industrial and safety critical skills, vigilance, motivations); job
(2019) systems attributes (criticality, activity impact,
safety impact); team attruibutes
(communication, pattern of distribution,
team experience, global synergy
distribution)
Fuzzy-ANP-PROMETHEE  Full aggregation An integrated approach that Status of the players owned by the club Technical ability - mental ability - physical
II-DEA  Outranking combines MCDM analysis and (sale, stay) ability (study the relationship between sub-
Nasiri et al. (2018)  Goal or reference mathematical programming to criteria)
level support the decision maker during
the football transfer season
AHP  Full aggregation Selection starting line-up of Starting line-up selection Physical, technical, tactical, mental
Purwanto et al. (2018) football players based on their
position
PROMETHEEI-  Outranking Applying a multi-criteria Ranking basketball players 6 efficiency indices (such as ratio of points
PROMETHEEII outranking methodology on the scored by the player with respect to the
Blanco et al. (2018) Spanish ACB Basketball League number of minutes played)
using as alternatives the potential
players and as criteria different
efficiency indices
AHP-  Full aggregation Selecting football players based on Team selection Several attributes are given by player’s
TOPSIS  Goal or reference Multi-criteria decision analysis position such as finishing, crossing, agility,
Flegl et al. (2018) level with the integration of qualitative ball control, acceleration, vision, sprint,
and quantitative data speed, dribbling …
TOPSIS  Goal or reference Assessing 24 football players Ranking foot-ball players Anthropometric, fitness, skills
Raheela et al. (2016) level belonging in the same team using
a MCDM
AHP-linear programming  Full aggregation Determining top performers and Selection of top performers Criteria for goalkeepers, defensive centres,
model  Goal or reference players for inclusion in the team midfielder centres, wings, forward centres
Ozceylan (2016) level
AHP-  Full aggregation Selecting the best player in a sport Ranking best player 6 criteria among them: the average number
TOPSIS  Goal or reference team of goals scored per game (due to the post)
Nikjo et al. (2015) level and social prestige (popularity)
TOPSIS  Goal or reference Ranking 18 basketball players Players starting the game 23 physical criteria, classified into 4 groups
Dadelo et al. (2014) level from the best to the worst of competences
(Body size and composition, speed and
quickness, power, aerobic endurance)
Fuzzy Inference System  Full aggregation Selecting player for a football Football player selection by position Heading, jumping, shoot, short, passing,
Tavana et al. (2013) team by position: defenders, crossing, ball control, dribbling, finishing,
midfielders, forwards speed, creativity, create goal, scoring,
position, tackling, both feet, stamina,
height, providing through (long) pass,
technical ability, create attacking
opportunities, read the game
AHP-  Full aggregation Mesuring the sporting Best team Players’ annual wages, Coach’s monthly
TOPSIS  Goal or reference performance of 18 professional wage, points, spectators, stadium
Mavi et al. (2012) level football teams utilization, total revenues
AHP  Full aggregation Ranking criteria and sub-criteria Evaluating weights of factors affecting Sport team factors (media exposure
Seungbum and Ross to identify the decision making sport sponsorship decision opportunity, sponsorship fit, team image,
(2012) factors of sport sponsorship in the fan base strength, performance, hospitality
global market context opportunity, facility average attendance),
country factors (interest level in sport,
political and economic state), environment
factors (competitors ambush marketing,
league authority over sponsorship)
AHP-  Full aggregation Measuring the performance of Ranking of faster-bowlers and spinners Economy rate, bowling average, bowling
TOPSIS  Goal or reference fast-bowlers and spinners of strike rate
AHP- level criket.
COPRAS
Dey et al. (2011)
TOPSIS  Goal or reference Ranking the six National Ranking player to transfer Rebounds per game, points per game,
Bozbura et al. (2008) level Basketball Association players blocks per game, age, player salary, assists
according to their proximities to per game
ideal solution given by TOPSIS
method

85
A. Ati et al. Data Science and Management 7 (2024) 79–88

Table 9
A samples of studies related to RQ4.
References Model/system Application

Feng et al. (2010) Multi-objective 0–1 programming model and Proposed method to select the members of a cross-functional team. They considered both the individual
an improved NSGA-II algorithm performance of candidates and the collaborative performance between candidates. They developed a
multi-objective 0–1 programming model and improved the nondominated sorting genetic algorithm II
(INSGA-II) to solve the member selection problem.
Tavana et al. (2013) Two-phase framework Proposed a two-stage framework for player selection and team formation in football: first, a fuzzy
ranking method, and then evaluation of alternative combinations of selected layers using a fuzzy
inference system (FIS).
This framework is illustrated with a case study using real data from a professional football team.
Ballı and Koruko
glu Fuzzy decision support framework port Developed a fuzzy multi-attribute decision-making algorithm for the selection problem of basketball
(2014) framework players. The model used FAHP method to identify the weights of criteria, and the TOPSIS method for the
final ranking alternatives.
Nikjo et al. (2015) WeFA (weighted factors analysis) framework Presented a new model for clubs’ head coaches and managers by considering experts’ votes. This
AHP-extended TOPSIS approach is based on the analytical hierarchy method (AHP) which is used to determine the weight of
each criterion and the extended TOPSIS method for weighting to decision makers (DM) and ranking of
alternatives. A numerical example is given with 10 experts, 3 decision makers, 4 players, and 6 criteria.
Ozceylan (2016) A mathematical model for player selection Study proposes a two-phase approach for selecting football players. In the first phase, the attributes of
each player at each position were prioritized using AHP. In the second phase, a 0–1 integer linear
programming model was developed using the weights of player attributes and the best players were
determined for inclusion in the team. The proposed solution was applied to a Turkish football club to
demonstrate its applicability and performance.
Blanco et al. (2018) Quantitative tool based on the PROMETHEE A quantitative ranking system was constructed to aggregate basketball players using the PROMETHEE
methodology. A case study of 191 players who participated in the Spanish ACB basketball league during
the 2014–2015 season.
Flegl et al. (2018) New methodology for personnel selection Proposed a new methodology for personnel selection based on MCDA and integrated qualitative and
(AHP-TOPSIS) quantitative data to select Mexican football teams for the 2018 World Cup in Russia.
Purwanto et al. (2018) System for line-up selection A system was developed to select the starting line-up of football players based on the AHP method.
Ouahli and Cherkaoui Team performance model using the fuzzy-AHP Three performance teams in industrial and safety critical systems were developed based on individual
(2019) performance, job attributes and team attributes using the fuzzy-AHP method.
Nasiri et al. (2018) A bi-objective integer programming model Proposed an integrated approach that combines a multi-criteria decision-making analysis (FANP-
PROMETHEE II-DEA) and mathematical programming to assist decision-makers in their selection
process.
Nsolo et al. (2019) Machine learning algorithms Compared which attributes and skills best predict the success of individual players in their positions in
five top European football leagues and evaluated different machine learning algorithms in terms of their
predictive performance.
Pappalardo et al. (2019) PlayeRank-machine learning Developed a data-driven framework to help professional football scouts evaluate, search, rank, and
recommend football players. This tool, called PlayeRank, is based on a machine learning approach. It
uses a large dataset of football records consisting of millions of match events from four seasons of 18
well-known football competitions.
Yang et al. (2020) Deep learning algorithms (MCDM) It has been argued that when faced with a set of alternatives, multi-criteria decision making is one of the
most appropriate decision-making tools. Combining deep learning algorithms with multi-criteria
decision making has proven beneficial for big data.
Zhao et al. (2021) Multi-objective optimization-genetic algorithm The team composition issue was formulated as a multi-objective optimization problem, and a variant of
the genetic algorithm was proposed that can automatically output and recommend a football team with a
high win rate by quantifying the players' abilities under a given budget constraint.
Wieckowski and Complex MCDM model based on COMET and Built a complex MCDM model based on the characteristic objects method (COMET), which was used to
Watr
obski (2021) fuzzy COs predict the chances of overtaking during pit stops in Formula 1 races. This model can reduce the number
of characteristic objects (COs) pair comparisons that must be performed by an expert (from 1,111 to
111), while maintaining its efficiency.

and 2023 were collected from digital libraries (IEEE, ScienceDirect, depending on the experience of coaches and managers (Dadelo et al.,
Springer, Taylor and Francis, and Wiley Online Library). Other relevant 2014; Purwanto et al., 2018; Salles et al., 2019). MCDM approaches may
articles were retrieved from the digital library of the University of Bur- assist decision-makers in making more objective decisions. In this
gundy, although they were not included in the selected period of the SLR. context, the current research aims to produce a comprehensive report on
To successfully perform the proposed SLR, all resources listed in the relationship between managerial-financial and sporting performance,
Tables 7–9 were carefully analyzed: title, keywords, abstract, and con- and the application of MCDM and machine learning algorithms in foot-
tents. These findings led us to conclude that of the 66 papers identified by ball team formation.
the SLR process, only a few articles addressed our four research questions Although this highlighted the combined use of MCDM and machine
(Blanco et al., 2018; Flegl et al., 2018; Kizielewicz and Dobryakova, learning for football player selection and performance prediction, future
2020; Nasiri et al., 2018; Purwanto et al., 2018; Romero et al., 2021; studies should incorporate more biometric and psychological data to
Shahriar et al., 2019). We could conclude that line-up selection using refine these predictions. In addition, with rapid advancements in tech-
MCDM methods was a relevant research area and topical subject, espe- nology, the exploration of AI techniques will ensure increased trans-
cially when combined with machine learning algorithms. This study parency in selection decisions. Finally, the proposed methodology
identified new ways to develop a decision support system for player se- deserves to be studied in other sports contexts, paving the way for
lection problems using multiple and conflicting criteria. This study’s broader applications and interdisciplinary results.
results may help the research community in this area to better under-
stand and design models, decision support systems based on several CRediT author statement
criteria, and integrated machine learning algorithms.
Selecting a line-up of teams for a decisive competition is complex Abdessatar Ati: Supervision. Patrick Bouchet: Supervision. Rou-
because many qualitative and quantitative criteria and sub-criteria must kaya Ben Jeddou: Writing-original draft.
be considered. However, this process tends to be subjectively interpreted,

86
A. Ati et al. Data Science and Management 7 (2024) 79–88

Declaration of competing interest Fry, J., Serbera, J.-P., Wilson, R., 2021. Managing performance expectations in
association football. J. Bus. Res. 135 (Oct.), 445–453.
Galariotis, E., Germain, C., Zopounidis, C., 2018. A combined methodology for the
The authors declare that there are no conflicts of interest. concurrent evaluation of the business, financial and sports performance of football
clubs : the case of France. Ann. Oper. Res. 266 (1–2), 589–612.
References Garcia-del-Barrio, P., Agnese, P., 2022. To comply or not to comply? How a UEFA wage-
to-revenue requirement might affect the sport and managerial performance of soccer
clubs. Manag. Decis. Econ. 44 (2), 767–786.
Alaminos, D., Esteban, I., Fern
andez-G
amez, M.A., 2020. Financial performance analysis Gomez-Gonzalez, C., Del Corral, J., Jewell, R.T., et al., 2019. A prospective analysis of
in European football clubs. Entropy 22 (9), 1–16. competitive balance levels in major league soccer. Rev. Ind. Organ. 54 (1), 175–190.
Arnason, A., Sigurdsson, S.B., Gudmundsson, A., et al., 2004. Physical fitness, injuries, Gonçalves, R.S., Mendes, R.C., Henriques, F.M., et al., 2020. The influence of sports
and team performance in soccer. Inside MS 36 (2), 278–285. performance on economic-financial performance : an analysis of Brazilian soccer
Athish V, P., Rajeswari, D., Sree Nandha S, S., 2023. Football prediction system using clubs from 2013 to 2017. Contextus – Revista Contempor^anea de Economia e Gest~ao
Gaussian naïve bayes algorithm. In: 2023 Second International Conference on 18 (11), 239–250.
Electronics and Renewable Systems. IEEE, pp. 1640–1643. Gu, W., Foster, K., Shang, J., et al., 2019. A game-predicting expert system using big data
Aydemir, A.E., Temizel, T.T., Temizel, A., et al., 2021. A dimension reduction approach to and machine learning. Expert Syst. Appl. 130 (9), 293–305.
player rankings in European football. IEEE Access 9 (8), 119503–119519. Hervert-Escobar, L., Matis, T.I., Hernandez-Gress, N., 2018. Prediction learning model for
Azeman, A.A., Mustapha, A., Razali, N., et al., 2021. Prediction of football matches soccer matches outcomes. In: 2018 Seventeenth Mexican International Conference on
results : decision forest against neural networks. In: 2021 18th International Artificial Intelligence (MICAI). IEEE, pp. 63–69.
Conference on Electrical Engineering/Electronics, Computer, Telecommunications Hu, S., Fu, M., 2022. Football match results predicting by machine learning techniques.
and Information Technology (ECTI-CON). IEEE, pp. 1032–1035. In: 2022 International Conference on Data Analytics, Computing and Artificial
Bai, Z., Bai, X., 2021. Sports big data: management, analysis, applications, and challenges. Intelligence (ICDACAI). IEEE, pp. 72–76.
Complexity 2021 (1), 1–11. Hwang, C.-L., Yoon, K., 1981. Methods for multiple attribute decision making. In:
Ballı, S., Koruko glu, S., 2014. Development of a fuzzy decision support framework for Multiple Attribute Decision Making : Methods and Applications A State-Of-The-Art
complex multi-attribute decision problems : a case study for the selection of skilful Survey. Springer, Berlin, pp. 58–191.
basketball players. Expet Syst. 31 (1), 56–69. Ishizaka, A., Nemery, P., 2013. Multi-criteria Decision Analysis, first ed. Wiley, New
Barros, C.P., 2006. Portuguese football. J. Sports Econ. 7 (1), 96–104. Jersey.
Bhat, Z.U.H., Sultana, D., Dar, Q.F., 2019. A comprehensive review of data envelopment Jenifer, Briscilla S., Sundarrajan, R., 2022. The Fuzzy Analytic Hierarchy Process to
analysis (DEA) in sports. JSEM 9 (2), 82–109. measure qualitative weight factor and Fuzzy MCDM methods for ranking the best
Blanco, V., Salmer
on, R., G

omez-Haro, S., 2018. A multicriteria selection system based on alternatives in team selection process. In: 2022 International Conference on Machine
player performance : case study-the Spanish ACB basketball league. Group Decis. Learning, Big Data, Cloud and Parallel Computing. COM-IT-CON), IEEE, pp. 501–510.
Negot. 27 (6), 1029–1046. Joe Anand, M.C., Martin, N., Clementking, A., et al., 2023. Decision making on optimal
Boon, B.H., Sierksma, G., 2003. Team formation : matching quality supply and quality selection of advertising agencies using machine learning. In: 2023 International
demand. Eur. J. Oper. Res. 148 (2), 277–292. Conference on Information Management (ICIM). IEEE, pp. 121–125.
Bozbura, F.T., Beşkese, A., Kaya, T.S., 2008. Topsis method on player selection in mba. In: Kaczynska, A., Kołodziejczyk, J., Sałabun, W., 2021. A new multi-criteria model for
2008 12th International Research/Expert Conference Trends in the Development of ranking chess players. Procedia Comput. Sci. 192 (10), 4290–4299.
Machinery and Associated Technology TMT 2008, Istanbul, Turkey, pp. 401–404. Karakaya, A., Ulu, A., Akleylek, S., 2022. GOALALERT : a novel real-time technical team
Buhagiar, R., Cortis, D., Newall, P.W.S., 2018. Why do some soccer bettors lose more alert approach using machine learning on an IoT-based system in sports.
money than others? J. Behav. Exp. Finance 18 (6), 85–93. Microprocess. Microsyst. 93 (9), 104606.
Cacho-Elizondo, S., 2020. Big data in the decision-making processes of football teams Katzenbach, J., 2009. Discipline of Teams. Business Review Press, Harvard.
integrating a theoretical framework, applications, and reach. JSIS 15 (2), 1–33. Katzenbach, J., Smith, D.K., 1993. The Wisdom of Teams : Creating the High-Performance
Carmichael, F., Thomas, D., Ward, R., 2000. Team performance : the case of English Organization. Business School Press, Harvard.
Premiership football. Manag. Decis. Econ. 21 (1), 31–45. Kitchenham, B., Charters, S., 2007. Guidelines for Performing Systematic Literature
Carmichael, F., McHale, I., Thomas, D., 2011. Maintaining market position : team Reviews in Software Engineering. Keele University and Durham University Joint
performance, revenue and wage expenditure in the English premier league. Bull. Report. Technical Report EBSE 2007-001.
Econ. Res. 63 (4), 464–497. Kizielewicz, B., Dobryakova, L., 2020. MCDA based approach to sports players’
Charnes, A., Cooper, W.W., Rhodes, E., 1978. Measuring the efficiency of decision making evaluation under incomplete knowledge. Procedia Comput. Sci. 176 (2020),
units. Eur. J. Oper. Res. 2 (6), 429–444. 3524–3535.
Chatterjee, A., Mukherjee, S., Kar, S., 2018. A rough approximation of fuzzy soft set-based Lepschy, H., W€asche, H., Woll, A., 2018. How to be successful in football : a systematic
decision-making approach in supplier selection problem. Fuzzy Inf. Eng. 10 (2), review. Open Sports Sci. J. 11 (1), 3–23.
178–195. Li, Y., Wang, L., Li, F., 2021. A data-driven prediction approach for sports team
Chen, J., Wang, J., Balezentis, T., et al., 2018. Multicriteria approach towards the performance and its application to National Basketball Association. Omega 98 (1),
sustainable selection of a teahouse location with sensitivity analysis. Sustainability 10 102123.
(8), 1–17. Mahmood, Z., Daud, A., Abbasi, R.A., 2021. Using machine learning techniques for rising
Dadelo, S., Turskis, Z., Zavadskas, E.K., et al., 2014. Multi-criteria assessment and ranking star prediction in basketball. KBS 211 (1), 106506.
system of sport team formation based on objective-measured values of criteria set. Manish, S., Bhagat, V., Pramila, R., 2021. Prediction of football players, performance
Expert Syst. Appl. 41 (14), 6106–6113. using machine learning and deep learning algorithms. In: 2021 2nd International
Danisik, N., Lacko, P., Farkas, M., 2018. Football match prediction using players Conference for Emerging Technology (INCET). IEEE, pp. 1–5.
attributes. In: 2018 World Symposium on Digital Intelligence for Systems and Manju, M.K., Philip, A.O., 2023. Novel method for ranking batsmen in Indian Premier
Machines (DISA). IEEE, pp. 201–206. League. Data Sci. Manag. 6 (2023), 158–173.
Datta, M., Rudra, B., 2022. An intelligent decision support system for bid prediction of Mareschal, B., Brans, J.P., Vincke, P., 1984. PROMETHEE: a New Family of Outranking
undervalued football players. In: 2022 2nd International Conference on Intelligent Methods in Multicriteria Analysis. Universite Libre de Bruxelles.
Technologies. IEEE, pp. 1–8. Mavi, R.K., Mavi, N.K., Kiani, L., 2012. Ranking football teams with AHP and TOPSIS
Dey, P.K., Ghosh, D.N., Mondal, A.C., 2011. A MCDM approach for evaluating bowlers methods. Int. J. Decis. Sci. Risk Manag. 4 (1/2), 108–126.
performance in IPL. J. Emerg. Trends Comput. Inf. Sci. 2 (11), 2079–8407. McHale, I.G., Holmes, B., 2023. Estimating transfer fees of professional footballers using
Dezman, B., Trnini
c, S., Dizdar, D., 2001. Expert model of decision-making system for advanced performance metrics and machine learning. Eur. J. Oper. Res. 306 (1),
efficient orientation of basketball players to positions and roles in the game 389–399.
-Empirical verification. Coll. Antropol. 25 (7), 141–152. Mohanta, B., Nanda, P., Patnaik, S., 2019. Uncertainty and ambiguity management in
Di Simone, L., Zanardi, D., 2020. On the relationship between sport and financial supplier selection using fuzzy analytical model. In: 2019 3rd International
performances : an empirical investigation. Manag. Finance 47 (6), 812–824. Conference on Data Science and Business Analytics (ICDSBA). IEEE, pp. 171–177.
Dimitropoulos, P., Scafarto, V., 2021. The impact of UEFA financial fair play on player Morciano, G., Zingoni, A., Morachioli, A., et al., 2022. Machine Learning prediction of the
expenditures, sporting success and financial performance : evidence from the Italian expected performance of football player during training. In: 2022 International
top league. ESMQ 21 (1), 20–38. Conference on Metrology for Extended Reality, Artificial Intelligence and Neural
Dobson, S., Goddard, J., 2001. The Economics of Football. CambridgeUniversity Press, Engineering (MetroXRAINE). IEEE, pp. 574–578.
Cambridge. Morrow, S., Howieson, B., 2014. The new business of football : a study of current and
Farshidi, S., Jansen, S., De Jong, R., et al., 2018. A decision support system for software aspirant football club managers. J. Sport Manag. 28 (5), 515–528.
technology selection. J. Decis. Syst. 27 (Suppl. 1), 98–110. Müller, O., Simons, A., Weinmann, M., 2017. Beyond crowd judgments : data-driven
Feng, B., Jiang, Z., Fan, Z., et al., 2010. A method for member selection of cross-functional estimation of market value in association football. Eur. J. Oper. Res. 263 (2),
teams using the individual and collaborative performances. Eur. J. Oper. Res. 203 (3), 611–624.
652–661. Nasiri, M.M., Ranjbar, M., Tavana, M., et al., 2018. A novel hybrid method for selecting
Flegl, M., Jim
enez-Bandala, C.A., Lozano, C., et al., 2018. Personnel selection in complex soccer players during the transfer season. Expet Syst. 36 (1), 1–19.
organizations : a case of Mexican football team for the 2018 World Cup in Russia. Nikjo, B., Rezaeian, J., Javadian, N., 2015. Decision making in best player selection : an
Revista del Centro de Investigaci
on de la Universidad la Salle 13 (49), 43–66. integrated approach with AHP and extended TOPSIS methods based on wefa
Frick, B., Simmons, R., 2008. The impact of managerial quality on organizational freamwork in MAGDM problems. Int. J. Res. Ind. Eng 4 (1), 1–14.
performance : evidence from German soccer. Manag. Decis. Econ. 29 (7), 593–600.

87
A. Ati et al. Data Science and Management 7 (2024) 79–88

Nsolo, E., Lambrix, P., Carlsson, N., 2019. Player Valuation in European Football. Saeed, M., Saqlain, M., Riaz, M., 2019. Application of generalized fuzzy TOPSIS in
Machine Learning and Data Mining for Sports Analytics, vol. 11330. Springer decision making for neutrosophic soft set to predict the champion of FIFA 2018 : a
International Publishing, pp. 42–54. mathematical analysis. Punjab Univ. J. Math. 51 (8), 111–126.
Ouahli, J., Cherkaoui, A., 2019. Team performance in safety critical systems : review and Salih, M.M., Zaidan, B.B., Zaidan, A.A., 2020. Fuzzy decision by opinion score method.
approximation by Fuzzy-AHP. J. Theor. Appl. Inf. Technol. 97 (13), 3767–3782. Appl. Soft Comput. 96 (Nov.), 106595.
Ozceylan, E., 2016. A mathematical model using ahp priorities for soccer player selection: Salles, S.A.F., Hora, H.R.M.D., et al., 2019. Operations research contributions for football
a case study. S. Afr. J. Ind. Eng. 27 (2), 190–205. teams formation: a systematic review. Pesquisa operacional 39, 277–293.
Pappalardo, L., Cintia, P., Ferragina, P., et al., 2019. PlayeRank : data-driven performance Samur, S., 2018. Organization design in football management process. Int. J. Hum. Mov.
evaluation and player ranking in soccer via a machine learning approach. ACM Trans. Sports Sci. 6 (2), 38–46.
Intell. Syst. Technol. 10 (5), 27. Seungbum, L., Ross, S.D., 2012. Sport sponsorship decision making in a global market an
Parida, S., Thilak, K.D., Singh, R., 2022. Enhancing the prediction of growth of footballers approach of Analytic Hierarchy Process (AHP). Sport Bus. Manag. 2 (2), 156–168.
using real-life statistics and machine learning. In: 2022 International Conference on Shahriar, Md T., Islam, Y., Amin, Md N., 2019. Player classification technique based on
Applied Artificial Intelligence and Computing (ICAAIC). IEEE, pp. 471–475. performance for a soccer team using machine learning algorithms. In: 2019
Patnaik, D., Praharaj, H., Prakash, K., et al., 2019. A study of Prediction models for International Conference on Electrical, Computer and Communication Technologies
football player valuations by quantifying statistical and economic attributes for the (ICECCT). IEEE, pp. 1–4.
global transfer market. In: 2019 International Conference on System, Computation, Steve Arrul, V., Subramanian, P., Mafas, R., 2022. Predicting the football players’ market
Automation and Networking (ICSCAN). IEEE, pp. 1–7. value using neural network model: a data-driven approach. In: 2022 International
Peddii, A., Jain, R., 2023. Random forest-based fantasy football team selection. In: 2023 Conference on Distributed Computing and Electrical Circuits and Electronics
9th International Conference on Advanced Computing and Communication Systems (ICDCECE). IEEE, pp. 1–7.
(ICACCS). IEEE, pp. 1822–1826. Szymanski, S., Smith, R., 1997. The English football industry : profit, performance and
Purwanto, I.N., Widodo, A., Handoyo, S., 2018. System for selection starting lineup of a industrial structure. Int. Rev. Appl. Econ. 11 (1), 135–153.
football players by using analytical hierarchy process (AHP). J. Theor. Appl. Inf. Tarnavsky Eitan, A., Smolyansky, E., Knaan Harpaz, I., et al., 2023. Connected Papers |
Technol. 96 (1), 19–31. Find and explore academic papers. Available at: https://www.connectedpapers.c
Qader, M.A., Zaidan, B.B., Zaidan, A.A., et al., 2017. A methodology for football players om/.
selection problem based on multi-measurements criteria analysis. Measurement 111 Tavana, M., Azizi, F., Azizi, F., et al., 2013. A fuzzy inference system with application to
(2017), 38–50. player selection and team formation in multi-player sports. Sport Manag. Rev. 16 (1),
Raheela, A., Muhammad, Z.T., Saba, H.I., et al., 2016. Football (soccer) analytics : a case 97–110.
study on the availability and limitations of data for football analytics research. Int. J. Trequattrini, R., Lombardi, R., Nappo, F., 2012. The evaluation of the economic value of
Comput. Sci. Inf. Technol. 14 (11), 516–518. long lasting professional football player performance rights. WSEAS Trans. Bus. Econ.
Rahman, M., Asadujjaman, Md, 2021. Multi-criteria decision making for job selection. In: 9 (4), 199–218.
2021 International Conference on Decision Aid Sciences and Application (DASA). Trninic, S., Papic, V., Trninic, V., et al., 2008. Player selection procedures in team sports
IEEE, pp. 152–156. games. Acta Kinesiol. 2 (2), 24–28.
Ribeiro, A.S., Lima, F., 2012. Portuguese football league efficiency and players’ wages. Tzeng, G.-H., Huang, J.-J., 2011. Multiple Attribute Decision Making : Methods and
Appl. Econ. Lett. 19 (6), 599–602. Applications. Taylor and Francis Group, LLC, Oxford.
Ribeiro, A.S., Lima, F., Kraus, S., et al., 2022. Tournaments within football teams : players’ Valenti, M., Scelles, N., Morrow, S., 2020. Elite sport policies and international sporting
performance and wages. ECON RES-EKON ISTRAZ 35 (1), 4884–4901. success : a panel data analysis of European women’s national football team
Rohde, M., Breuer, C., 2016. Europe’s elite football : financial growth, sporting success, performance. ESMQ 20 (3), 300–320.
transfer investment, and private majority investors. Int. J. Financ. Stud. 4 (12), 1–20. Wadjdi, A.F., Sianturi, E.M., Ruslinawaty, N., 2018. Design of data collection form to
Rohde, M., Breuer, C., 2017. The market for football club investors : a review of theory ensure consistency in AHP. In: 2018 10th International Conference on Information
and empirical evidence from professional European football. ESMQ 17 (3), 265–289. Technology and Electrical Engineering: Smart Technology for Better Society
Rohde, M., Breuer, C., 2018. Competing by investments or efficiency ? Exploring financial (ICITEE). IEEE, pp. 529–533.
and sporting efficiency of club ownership structures in European football. Sport Wieckowski, J., Watr
obski, J., 2021. How to determine complex MCDM model in the
Manag. Rev. 21 (5), 563–581. COMET method ? Automotive sport measurement case study. Procedia Comput. Sci.
Romero, F.P., Lozano-Murcia, C., Lopez-Gomez, J.A., et al., 2021. A data-driven approach 192 (2021), 376–386.
to predicting the most valuable player in a game. Comput. math. methods 3 (4), 1–11. Xia, V., Jain, K., Krishna, A., et al., 2018. A network-driven methodology for sports
Rose, J.D., Vijaykumar, M.K., Sakthi, U., et al., 2022. Comparison of football results using ranking and prediction. In: 2018 52nd Annual Conference on Information Sciences
machine learning algorithms. In: 2022 International Conference on Innovative and Systems (CISS). IEEE, pp. 1–6.
Computing, Intelligent Communication and Smart Electrical Systems (ICSES). IEEE, Yang, M., Nazir, S., Xu, Q., et al., 2020. Deep learning algorithms and multicriteria
pp. 1–7. decision-making used in big data : a systematic literature review. Complexity 2020
Roy, B., 1968. Ranking and selection when there are multiple viewpoints. Fr. J. Comput. (8), 1–18.
Sci. Operational Res. 2, 57–75. Zhao, H., Chen, H., Yu, S., et al., 2021. Multi-objective optimization for football team
Saaty, T.L., 1996. Decision Making with Dependence and Feedback : the Analytic member selection. IEEE Access 9 (6), 90475–90487.
Network Process: the Organization and Prioritization of Complexity, first ed. RWS Zotero_community, 2023. Zotero | Your personal research assistant. Available at: http
Publications, Pittsburg. s://www.zotero.org/.
Saaty, T.L., Vargas, L.G., 1982. The Logic of Priorities. Springer Netherlands.

88