Event-Aware Predictive Route Recommendation Using Random Forest and A* Search
Abstract: Urban traffic conditions can change considerably during peak hours and around public events. Conventional route selection generally reacts to the traffic condition observed at the time of travel, whereas a predictive approach can estimate likely travel duration before a journey begins. This paper presents a student-level prototype for event-aware route recommendation that combines supervised learning with graph-based path search. A Random Forest regression model is used to estimate travel duration from route origin, destination, distance, travel hour, and an event indicator. The predicted duration is then used as the cost of road segments in an A* search procedure. A small urban road graph representing selected locations is used for demonstration. The implementation consists of a Python-based machine-learning module, a Flask backend, and a React user interface. The project report records an R? value of 0.9687 for the main modeling experiment, while a separately generated event-focused dataset produced an R? value of 0.9291. These results indicate that the prototype can learn useful relationships in the prepared dataset and can translate predicted travel time into route-selection decisions. The work is intentionally presented as an academic prototype rather than a live city-scale navigation system; real-time traffic feeds, geographical coordinates, larger datasets, and extensive field validation remain future work.
Keywords: Traffic prediction, route recommendation, Random Forest regression, A* search, event-aware routing, smart mobility, machine learning
How to Cite?: Dr. Sandhiya S, "Event-Aware Predictive Route Recommendation Using Random Forest and A* Search", Volume 15 Issue 10, October 2026, International Journal of Science and Research (IJSR), Pages: 32-35, https://www.ijsr.net/getabstract.php?paperid=SR26930064033, DOI: https://dx.doi.org/10.21275/SR26930064033