Incident Duration Prediction Based on Latent Gaussian Naive Bayesian classifier Dawei Li , Lin Chen , Jiangshan Ma

oleh: Dawei Li, Lin Cheng, Jiangshan Ma

Format: Article
Diterbitkan: Springer 2011-05-01

Deskripsi

The probability distribution of duration is a critical input for predicting the potential impact of traffic incidents. Most of the previous duration prediction models are discrete, which divide duration into several intervals. However, sometimes the continuous probability distribution is needed. Therefore a continuous model based on latent Gaussian naive Bayesian (LGNB) classifier is developed in this paper, assuming duration fits a lognormal distribution. The model is calibrated and tested by incident records from the Georgia Department of Transportation. The results show that LGNB can describe the continuous probability distribution of duration well. According to the evidence sensitivity analysis of LGNB, the four classes of incidents classified by LGNB can be interpreted by the level of severity and complexity.