Target Movement Research Based on Spearman's Rank Correlation Analysis and Ridge Regression: Key Identification, Patterns, and Prediction

Abstract

This paper proposes a research methodology for the long jump that integrates kinematic analysis with statistical modeling. It focuses on determining the timing of takeoff and landing, analyzing the movement patterns during the flight phase, and investigating the relationship between influencing factors and performance outcomes, including predictive capabilities. First, using key foot joints as the core, precise timing of takeoff and landing is determined by calculating average vertical foot velocity and position. This is combined with three-tier dynamic indicators—displacement, velocity, and acceleration—along with quantifying fluctuation thresholds across phases to obtain critical timing parameters for athletes. Second, treating athletes as point masses, a projectile motion model is constructed. The flight phase is divided into ascent, apex, and descent stages. The movement patterns of parameters such as center-of-mass trajectory and joint angles are analyzed, comparing differences in movement characteristics among athletes. Finally, a Spearman rank correlation analysis model identifies 33 features significantly correlated with performance. Ridge regression then addresses multicollinearity among features and quantifies each factor's contribution to performance, enabling the construction and validation of a long jump prediction model. This model accurately captures critical movement nodes, clarifies movement patterns, effectively screens core influencing factors, and reliably predicts long jump performance while resolving multicollinearity issues, yielding robust analytical results.