A Risk Prediction Model Based on Crash History Data for Railway Trams

Risk prediction is an important task to ensuring the driving safety of railway trams. Although data-driven intelligent methods are proved be effective for risk prediction, accuracy still a top concern challenges data quality which mainly represent as unbalanced datasets. This study focuses on applying feature extraction and augmentation achieve trams, proposes approach based self-adaptive K-means clustering algorithm least squares deep convolution generative adversarial network (LS-DCGAN). The preprocessing proposed, include cluster locations trams extreme gradient boosting recursive elimination selection retain key features. LS-DCGAN model designed sparse sample expansion, aiming address category distribution imbalance problem. experiments implemented with public real datasets show that proposed can reach high 90.69%, greatly enhances tram safety.