MCDA based approach to sports players’ evaluation under incomplete knowledge

Abstract

Basketball is a constantly developing sport due to the dynamics of changes in the significant characteristics of a basketball player. The acquisition of new basketball players’ skills strongly influences the development of new technologies, as well as the game itself. In addition, it plays an important role in the personal development of young people seeking new achievements. However, it is difficult to determine which basketball player is the best. Creating a proper ranking of the best basketball players in the NBA league is extremely important. There are many rankings of players in the NBA league, but it is difficult to determine if they are correct due to lack of relevant data in the assessment process. The NBA has many talented basketball players in different positions. Their different predispositions make the correct ranking very difficult. Using only one attribute of a given basketball player may prove to be wrong when assessing basketball players playing on different positions. In addition, some basketball players may be missing data, because it was not collected during the player’s playing time. Therefore, the survey was carried out in the absence of some statistics in several basketball players. The aim of such a survey is to prove that creating the correct ranking is possible for incomplete data. For the evaluation of selected basketball players from the NBA league a technique from the family of multi-criteria decision making methods (MCDA) called COMET was used. The COMET method works on the basis of fuzzy logic, and its distinguishing feature compared to other methods is its resistance to the paradox of reversal of rankings. Resistance to the paradox of reversal of rankings is ensured by the fact that the assessment of alternatives does not take place on their own, but on characteristic objects. An expert has been involved in their proper evaluation. This article presents research against NBA basketball players, which shows that correct ranking of basketball players even with partial lack of data for them is possible.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 38 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.