• A Data Mining Approach: Application to the Extraction of the Characteristics of IARD Products in the Insurance Sector

    Nora MOSBAH LOUNICI1, Khadidja SADI2 and Nadjib LOUNICI3

    1. High National School of Statistics and Applied Economics ENSSEA, Universities of Kolea, Algeria. Laboratory of applied statistics

    2. High National School of Statistics and Applied Economics ENSSEA, Universities of Kolea, Algeria. Laboratory of applied statistics

    3. School of Commercial High Studies.

    Abstract: In this study, we were interested to multirisk home products (MH) and professional multirisk (MP) of the branch IARD insurance, which covers fire, accidents and various risks of an Algerian insurance company (SAA). We want to know the variables that best characterize each of the two products. To solve this problem, we have combined three methods borrowed from Data Mining and decisional statistics. The use of data mining tools to realize a classification and to bring out the most informative variables is usual tasks of this discipline. The aim of this work is to extract knowledge from an actuarial database by adopting an approach to data mining and to compare the results obtained by querying three techniques, implemented in data mining software: TANAGRA.

    Keywords: Data mining, Insurance IARD, Regression, Step Disk, V-test.

    Pages: 311 – 321 | Full PDF Paper