Linear Regression Crash Prediction Models: Issues and Proposed Solutions

摘要

The paper develops a linear regression model approach that can be applied to crash data to predict vehicle crashes. The proposed approach involves novice data aggregation to satisify linear regression assumptions; namely error structure normality and homoscedasticity. The proposed approach is tested and validated using data from 186 access road sections in the state of Virginia. The approach is demonstrated to produce crash predictions consistent with traditional negative binomial and zero inflated negative binomial general linear models. It should be noted however that further testing of the approach on other crash datasets is required to further validate the approach.