The mobility management skills, including the capacity to make scientific decisions can be improved as well as the reduction of personal injuries and property losses by timely and precise identification of traffic events. The detection effect is significantly impacted by the imbalance in the statistics on traffic incidents. Consequently, An technique for identifying traffic events inferred via factor analysis and adjusted random forests (FA-WRF) is developed. Analyze the flow of traffic parameters change rule to establish the first incident variable. The component analysis (FA) approach is used to minimize the original incident variables’ dimension. Using the Bootstrap improved technique, establishing the information collection criteria for the training set. Upon training, The decision tree's categorization impact is measured using the MCC coefficient value. The random forest (RF) procedure's overall ability to discriminate for imbalanced data can be improved by using the tree is used as a weight value to make sure that specimens with better classification skills have more influence during the voting phase. The area under the receiver operating characteristic curve, the classification rate, and the false alarm rate, and the recognition rate are typical metrics used to evaluate the effectiveness of detecting systems. The results of the experiment show how well the FA-WRF-based model performs in terms of categorization. It does just as well when classifying imbalanced data as the Support Vector Machine does.