Identyfikacja potencjalnych nabywców polis ubezpieczeniowych w warunkach mocno niezbilansowanej próby uczącej
Identification of potential purchasers of the insurance policies under hard unbalanced training set

Summary/Abstract: Having given the data set with executed transactions and customer demographic features one can use marketing scoring to support sales campaign. The discrimination methods used in the scoring often face the problem of imbalance classes and irrelevant variables. In this paper, we analyze the insurance market, where the scoring is performed with a use of the weighted k nearest neighbors and multivariate filters. The feature selection significantly contributed to increasing the number of correctly identified potential purchasers of the insurance policy.

• Details
• Contents