Enhancing Road Safety: A Novel Fuzzy-RBF based Approach for Car Accident Prediction

The danger posed by traffic accidents is significant, as they have become a primary cause of death in many regions of the world. Lessening the frequency of such events is crucial for improving city and community resilience and saving lives. Machine learning and data analysis techniques decipher the root causes of accidents and propose corrective actions. The volume and speed with which data is being generated about road incidents necessitates the adoption of big data approaches. This system analyzes patterns in road car accident data and proposes a predictive model, with a focus on key data variables like the number of victims, accident severity, and the number of vehicles involved. The suggested methodology incorporates preprocessing, feature selection, and model training. The proposed approach employ Tokenization, Stop-word elimination, Stemming, and Lemmatization in preprocessing data. For classification models, feature selection is a crucial post-processing step because many input space variables are likely to be highly correlated with one another or have negigible effects on crash occurrences. The models are then trained using RBF-Fuzzy after the attributes have been gathered. The proposed method significantly outperforms the RBF and Fuzzy methods. This plan is nearly guaranteed to succeed at a rate of 98.23%.