Search for: [Abstrakt = "In the article an attempt to construct a model of insurance risk is made using a logistic regression. It is assumed that due to the information asymmetry on the insurance market the risk on the insurance policy is not additive with respect to partial risks. Owing to the specificity of insurance risk, particularly low probability of damage, the model was constructed on the balanced sample. Different methods of the variable coding and their influence on the model's content were taken into consideration. Optimal models were chosen on the basis of goodness of fit measures and power predictive measures. They were calculated for development and validation samples. The proposed model allows the classification of insured to homogeneous risk's groups with respect to insurance policy characteristics."]
