A long time multi-parameter predictive analysis method based on a hybrid model of random forest and ridge regression

This study proposes a hybrid predictive model (RR-Model) integrating Ridge Regression and Random Forest to forecast Olympic medal distributions, addressing three core challenges: Medal prediction under temporal uncertainty, quantification of coaching effects, and strategic optimization for host nations. Leveraging historical data from 1896 to 2024, this paper introduce a weighted ensemble approach to balance short-term trends (Ridge Regression on 2000-2024 data) and long -term patterns (Random Forest on full historical data). Our model predicts the 2028 Summer Olympics medal table with 88.7% accuracy, identifying the U.S., China, and Australia as top performers. The paper further quantify the “Great Coach Effect” using gray correlation analysis, demonstrating a minimum 3%, performance enhancement per targeted event. For host nations strategic event selection (e.g., prioritizing swimming and athletics for the U.S. in 2028) is shown to amplify medal gains by 9.8% through parameter optimization. Methodological robustness is validated via sensitivity analysis ( across scenarios), offering actionable insights for Olympic committees.