This study presents a framework for grouping Japanese professional baseball players. For the analysis, we used 115 different hitting indexes for 327 fielders who played in official games in 2020 provided by DELTA Corporation. We divided these indexes into eight categories defined by DELTA and narrowed 327 players down to seventy-two players with "WAR "of 1.0 or higher. We performed a principal components analysis on each of the itemized indexes to condense the information as variables. Using the principal component scores, we performed a cluster analysis using a mixed distribution model and classified the players into multiple clusters. Furthermore, we used a random forest to identify which indexes contributed to the results of the cluster analysis. This framework suggested that players who belonged to the same cluster more frequently could be substitutes for each other in eight categories, and that players who were farther from each other could be trade personnel.