An Innovative mode: An Integrated CRITIC-Multivariate Linear Regression Framework with TOPSIS Multi-Criteria Evaluation

Abstract

 As the world's most influential multi-sport event, the Olympic Games have long faced challenges in constructing medal prediction models due to integrating multi-source data and quantifying uncertainties. Based on official data from the International Olympic Committee (IOC), this study proposes a prediction framework that combines multi-dimensional feature quantification with uncertainty analysis. First, a National Comprehensive Sports Strength Index is constructed using the CRITIC (Criteria Importance Through Intercriteria Correlation) method. This method calculates each factor's weight based on its variance and its correlation with other criteria. The index integrates factors such as event participation coverage, athletes' competitive levels, and historical medal data to reveal key influences beyond the host country effect. Next, a linear regression model optimized by gradient descent is applied. The coefficient of determination () of the test set improves by 13.7% compared to the baseline model, thus confirming the model's generalization ability. An innovative Project Competition Coefficient is introduced to address the challenge of predicting medal counts for countries that have not yet won medals. Combined with the TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) multi-criteria evaluation model, this coefficient quantifies the potential for winning medals. The analysis identifies countries like Samoa, which have the potential to succeed in less competitive events. Finally, a Monte Carlo simulation is introduced to apply random perturbations to the input data, generating 10,000 samples. A 95% confidence interval is constructed, with a coverage rate of 91.2%, significantly reducing prediction uncertainty. These findings give event organizers a basis for dynamic resource allocation and offer a quantitative tool to help non-traditional sports powerhouses develop differentiated strategies. This, in turn, contributes to the balanced evolution of the Olympic competitive landscape.