Abstract
The purpose of this study is to clarify the influential factors in frontal collision using Japanese accident data. The injury prediction method and influential factors are proposed using ordinal logistic regression model. In this analysis, the significant factors are delta-V, seat belt use, vehicle damage grade and occupant's age. The prediction model using these factors can correspond to real accident data. On the other hand, the accident conditions which do not correspond to this model have several characteristics. These accident cases are frontal pole collision, short statured elderly female, etc. There is a possibility that these characteristics are the influential factors for injury level.
