Probabilistic Fuzzy Classification for Strategic Player Selection in T20I Cricket

Abstract

This study presents a knowledge-based intelligent system for T20I cricket team selection using probabilistic fuzzy clustering. Utilizing performance data from 336 international players spanning 2005 to 2024, the analysis classifies players into four functional categories: specialist batters, specialist bowlers, balanced all-rounders, and bowling all-rounders. The Fuzzy C-Means (FCM) clustering algorithm, integrated with Principal Component Analysis (PCA), is employed to handle the multidimensional nature of player statistics while capturing the uncertainty and role overlap typical in T20 cricket. The findings indicate that fuzzy clustering effectively identifies ideal substitutes and balanced team compositions by assigning probabilistic memberships across multiple roles. Eight representative players are selected from each cluster, illustrating the model’s capacity to support strategic, flexible, and data-driven team selection. This probabilistic fuzzy logic framework provides a transparent and practical solution for optimizing squad composition in the dynamic context of T20 cricket..

Keywords:  Fuzzy C Means (FCM), T-20 Cricket, Probabilistic, Player Selection, Membership  Function