MODELING INSURANCE PRICING POLICIES USING CLUSTER AND REGRESSION ANALYSIS

The study aims to develop a methodology for determining optimal insurance service pricing based on cluster analysis and regression modeling techniques. The relevance of the research is driven by the growing need to improve pricing accuracy in insurance markets characterized by increasing competition and uncertainty. The percentage of settled claims and the average claim payment amount were selected as the primary determinants of insurance policy cost. The methodological framework combines multivariate statistical analysis, hierarchical clustering, and econometric modeling. Insurance companies were classified using standardized data, the furthest-neighbor clustering method, and the block distance metric. Linear and nonlinear regression models were subsequently developed for each cluster to describe the relationship between policy cost and key insurance performance indicators. The empirical analysis identified three stable clusters of insurance companies with distinct claim settlement and payment characteristics. The results indicate that the relationship between policy cost and explanatory variables is heterogeneous across clusters. Linear specifications provide the best fit for certain groups, whereas nonlinear models demonstrate superior predictive performance for others. The developed models exhibit strong explanatory power and forecasting accuracy. The practical significance of the study lies in the applicability of the proposed cluster–regression framework for insurance pricing, policy cost forecasting, and portfolio management optimization. The methodology is universal in nature and can be adapted to various insurance segments provided that relevant statistical data are available.