Analysis of freeway accident frequencies: Negative binomial regression versus artificial neural network [An article from: Safety Science]

This digital document is a journal article from Safety Science, published by Elsevier in . The article is delivered in HTML format and is available in your Amazon.com Media Library immediately after purchase. You can view it with any web browser.

Description:
The Poisson or negative binomial regression model has been employed to analyze vehicle accident frequency for many years. However, these models have the pre-defined underlying relationship between dependent and independent variables. If this assumption is violated, the model could lead to erroneous estimation of accident likelihood. In contrast, the artificial neural network (ANN), which does not require any pre-defined underlying relationship between dependent and independent variables, has been shown to be a powerful tool in dealing with prediction and classification problems. Thus, this study employs a negative binomial regression model and an ANN model to analyze 1997-1998 accident data for the National Freeway 1 in Taiwan. By comparing the prediction performance between the negative binomial regression model and ANN model, this study demonstrates that ANN is a consistent alternative method for analyzing freeway accident frequency.