ADVANCED DATA MINING OF TABLE TENNIS MATCH BASED ON FP-GROWTH DATA ANALYSIS

In the training process for table tennis players, providing strategic information analysis of their opponents is crucial. The application of big data analysis and data mining techniques becomes increasingly important in sports technology. In this study, the table tennis professionals collect information from past match videos. The types of serving and receiving action, the ball landing placement, the number of strokes, and the points gained or lost are recorded. Using statistical analysis and professional examination and description, the relationships between the players and their match strategies are analyzed. In addition, the collected match information is analyzed using advanced algorithm, FP-Growth in the data mining process. The experimental results are based on the analysis of match video of Taiwan's top table tennis player, Lin Yun-Ju. Both the manual and computer analyses show that Lin Yun-Ju has better performance in the “attack-after-serve phase”. Furthermore, the computer analysis reveals that most of the lost points are from the “rally phase”, the rule confidence level is 64.8%. Also, if the serving action is “side under spin” and the lost points are from the “rally phase”, the confidence level is 64.3%. If the points are gained then serving type is “side under spin”, the confidence level is 61.4%. Compared with the professional analysis, data mining provides more important information which has a higher confidence level than 50% in the highly competitive game. The results offer table tennis players clearer training improvement suggestions based on the association rules.