关键词: 车胎安全；机器学习；随机森林；特征工程；预警系统

Abstract: This study employs machine learning techniques to develop an efficient tire safety detection and early warning system designed for early identification of potential hazards and risk alerts. Through data collection, problem analysis, and methodological discussion, the research utilizes an optimized random forest approach to examine factors such as brand, age, mileage, and weather conditions affecting tire performance, thereby establishing a classification model for predicting the severity of tire issues. Experimental results demonstrate that the optimized model effectively distinguishes between four categories-"minor" "moderate" "severe" and "critical"-with the random forest model achieving over 95% accuracy on the test set, confirming that tire age and mileage are critical determinants of safety performance. The presented data and analytical methods provide valuable insights for supporting autonomous driving applications in new energy vehicles, holding significant research and practical importance.

Key words: tire safety; machine learning; random forest; feature engineering; early warning system