<![CDATA[The professional basketball industry, including the Indonesian Basketball League (IBL), is increasingly adopting data-driven analytics to support strategic decision-making. However, previous studies on domestic basketball leagues have largely relied on individual performance statistics without applying dimensionality reduction techniques, making traditional player position classifications insufficient for objectively representing the characteristics of modern basketball players. To address this limitation, this study integrates Principal Component Analysis (PCA) with clustering algorithms to analyze 160 IBL players from the 2025 season who met the inclusion criteria (GP >= 5 and MP>= 8 minutes) from a total of 246 registered players. Seven statistical variables were used as input features, including Points Per Game (PPG), Rebounds Per Game (RPG), Assists Per Game (APG), Steals Per Game (SPG), Blocks Per Game (BPG), Field Goal Percentage (FGpercen), and Three-Point Field Goal Percentage (3Ppercen). Principal Component Analysis (PCA) was applied prior to clustering to address multicollinearity among variables and reduce data dimensionality while preserving the essential information contained in the original dataset. The PCA results indicate that two principal components explained seventy-one point six percent of the total data variance. The first principal component (PC1) accounted for forty-eight point six percent of the variance and was primarily influenced by PPG and RPG, whereas the second principal component (PC2) explained twenty-three point zero percent of the variance and was dominated by APG, BPG, and 3P percen. Based on internal validation metrics using the optimal number of clusters (k = 4), four distinct player archetypes were identified: All-Around, Bigman, Role Player, and Three-Point Shooter. The K-Means algorithm produced more compact clusters than Ward's Hierarchical Clustering, achieving a Silhouette Score of 0.25, a Davies–Bouldin Index of 1.36, and a Calinski–Harabasz Index of 74.8, with an agreement rate of 81.9 percent between the two clustering algorithms. The moderate Silhouette Score suggests that the boundaries between player archetypes are gradual rather than sharply defined. This study contributes to the academic literature by proposing a player classification framework that combines dimensionality reduction with comparative clustering validation. The proposed approach provides an empirical basis for professional basketball organizations to support player performance evaluation, talent identification, and recruitment decision-making.]]>
