Research on Tournament Influencing Factors Based on Logistic Regression Modeling and AHP

Logistic regression algorithm is a supervised learning algorithm widely used in classification problems. To expand its application, this paper applies it to the modelling examination of ‘momentum’, a special factor that is difficult to measure and has no obvious change characteristics, in tennis matches. In this paper, a quantitative method of ‘momentum’ based on hierarchical analysis method (AHP) is firstly proposed to evaluate ‘momentum’ quantitatively. Then, a prediction method of ‘momentum’ change based on regression model is proposed. The linear regression model is used to predict the change of ‘momentum’ and the most relevant indicators during the race. The experimental results show that the quantitative method proposed in this paper indicates that the player's physical strength and his own level are the most important indicators affecting ‘momentum’, with weights of 35.095% and 28.38%, respectively. The accuracy of the prediction method proposed in this paper is 91%. The study provides a scientific method and theoretical basis for the measurement of such factors that are difficult to measure the characteristics of insignificant changes and provides a reference for the modelling of other similar scenarios.