Elite male table tennis matches diagnosis using SHAP and a hybrid LSTM–BPNN algorithm

Abstract

This study adopts a new approach, SHapley Additive exPlanation (SHAP), to diagnose the table tennis matches based on a hybrid algorithm, namely Long Short-Term Memory–Back Propagation Neural Network (LSTM–BPNN). 100 male singles competitions (8535 rallies) from 2019 to 2022 are analyzed by a hybrid technical–tactical analysis theory, which hybridizes the double three-phase and four-phase evaluation theories. A k-means cluster analysis is conducted to classify 59 players’ winning rates into three levels (high, medium, and low). The results show that LSTM–BPNN has excellent performance (MSE = 0.000355, MAE = 0.014237, RMSE = 0.018853, and $${\mathrm{R}}^{2}$$ R 2 = 0.988311) compared with six typical artificial intelligence algorithms. Using LSTM–BPNN to calculate the SHAP value of each feature, the global results find that the receive-attack and serve-attack phases of the ending match have essential impacts on the mutual winning probabilities. Finally, case applications show that the SHAP can directly obtain each feature importance on one or more matches, which is more objective and reliable than the traditional simulation method. This research explores an innovative way to understand and analyze matches, and these results have implications for the performance analysis of table tennis and related racket sports.

Exploring the relationship among soccer-related knowledge, attitude, practice, and self-health in Chinese campus soccer education

“…Then, this study selected the best algorithm to explain the model globally as well as locally using SHAP, which is consistent with previous studies. 78 , 79 , 80 The local interpretations have two perspectives: 1) gender (male, female) and 2) city vs. village. All data processing and modeling were performed via Python 3.8.…”

Long-term gridded land evapotranspiration reconstruction using Deep Forest with high generalizability

“…In this study, we exploited the SHAP method 32 and ALE plots 33 to provide an interpretation of the model, leveraging their solid theoretical foundations and wide applicability across various fields 39 – 44 . SHAP values are calculated for each sample in the dataset to quantitatively capture the impacts of each feature within that sample.…”

Using SHAP to analyze technical and tactical performance in table tennis based on hybrid Catboost-TPE with four-phase evaluation theory

“…The defined scope of each phase follows the following rules: 1) the serve-attack phase: the points scored and lost by the serving player on the first stroke and the third stroke, and the points lost on the fifth stroke; 2) the receive-attack phase: the points scored and lost by the receiving player on the second and fourth strokes; 3) The stalemate I phase: the points scored by the serving player on the fifth stroke and the points scored and lost on the odd strokes after the fifth stroke. 4) The stalemate II phase: the points scored and lost by the receiving player on the sixth stroke and the points scored and lost on the even-numbered strokes after the sixth stroke [29]. The rules can be shown in Table 1.…”

“…SR 4 × UR 4 × Z 2 equals the (phase score in the stalemate II phase of the four-phase evaluation theory) × 100% × 100%. According to Table 1, the score in the stalemate phase of the new three-phase evaluation theory equals the score in the stalemate I phase, and the stalemate II phase of the four-phase evaluation theory [29]. Therefore, SR 3 × UR 3 × Z 1 + SR 4 × UR 4 × Z 2 equals to (phase score + phase points lost) × 100% × 100%.…”

“…Fortunately, there have been many approaches to increasing models' interpretability. Among them, the SHapley Additive exPlainnation (SHAP) is widely used [29]. SHAP is based on local approximation and cooperative game theory to interpret models [30], and it has been used in many industries with excellent results [31,32].…”