Proactive caching is crucial for optimizing cellular network performance by predicting and prefetching user-requested content. This study investigates cache priority prediction using three classifiers: SVM, KNN, and Random Forest. The dataset comprises 2621 samples categorized into three classes: Low, Medium, and High cache priority. The models were evaluated using accuracy, precision, recall, and F1-score. The SVM classifier achieved the highest accuracy of 97%, along with precision (95%), recall (89%), and F1-score (92%), demonstrating its superior predictive capability. The KNN classifier followed with an accuracy of 94%, precision (91%), recall (92%), and F1-score (91%). The Random Forest classifier achieved an accuracy of 88%, with precision, recall, and F1-score values of 85%. The experimental results indicate that SVM outperformed the other classifiers across all evaluation metrics. Future research could focus on exploring alternative classifiers and optimizing current models to enhance the accuracy and efficiency of cache priority prediction.