Optimizing Sports Performance Through Data-Driven Food Classification With C4.5 Algorithm

As the economy grows and national fitness initiatives gain traction, fitness sports have become more popular. Effective physical activity depends on both physical condition and diet. Optimizing the link between food nutrition and exercise is essential for nutrient absorption and fitness. This paper explores the impact of food preparation on sports performance and suggests methods to combine nutrition with exercise. An improved C4.5 algorithm, using sample selection and cosine similarity, is proposed to enhance classification accuracy and reduce training time for large food datasets. The optimal sample size is determined through a statistical strategy, followed by iterative adjustments to improve performance. Cosine similarity merges highly similar attribute pairs, refining the training set. The C4.5 algorithm selects the best splitting attribute and builds a decision tree, improving efficiency and accuracy. Experimental results confirm the model's effectiveness in overcoming limitations of existing classification algorithms for food data.