Abstract

The number of casualties and fatalities from road accidents presents a significant global challenge. This research advocates a proactive approach by pre-positioning ambulances, reducing response times, and improving medical assistance efficiency. Using deep learning techniques, particularly deep-embedded clustering, the study introduces a novel method to predict optimal ambulance locations by recognizing patterns in accident occurrences and geographical factors. The integration of Cat2Vec enhances model construction, improving accuracy in location prediction. A comparative analysis with K-means, GMM, and Agglomerative clustering highlights the superiority of the proposed approach. Additionally, a novel scoring function assesses real-time response time and distance calculation, further optimizing emergency medical services.