Using Feature Selection to Reduce the Complexity in Analyzing the Injury Severity of Traffic Accidents

When analyzing the traffic accidents in terms of predicting injury severity, past studies often use too many variables and thus lead to over fitting and complicate the interpretation of the analysis. By adopting feature selection technique, irrelevant and redundant features from a dataset will be filtered out such that high discrimination power and informative features will be provided. This paper selects twenty eight factors by adopting feature selection to analyze the injury severity of traffic accidents in Taiwan. The method facilitates to reduce the complexity of analyzing the injury severity of traffic accidents. The findings show that nineteen factors are classified into important, one is categorized as marginal, and five are grouped into unimportant.