Research on Diabetes Prediction Based on the Randomized Search CV Method

Early detection of diabetes is crucial for improving patient treatment healthcare costs. In diabetes Random Forest model is widely used due to its excellent classification d dimensional data. However, its performance heavily depends on hyperparameter settings, leaving room for improvement. To enhance predictive accuracy, this study employs the RandomizedSearchCV method to optimize the hyperparameters of the Random Forest model and conducts comparative experiments using the Pima Indian Diabetes (PID) dataset before and after optimization. The results indicate that the optimized model improves accuracy from $81 \%$ to $95 \%$, while the AUC value increases from 0.81 to 0.95. This demonstrates that hyperparameter optimization based on RandomizedSearchCV plays a significant role in improving model performance and provides valuable technical support for the early diagnosis and prevention of diabetes.