Application of GIS Spatial Interpolation Methods inAuto Insurance Risk Territory Segmentation and Rating

Abstract

Evolution in the field of Geographic Information Systems (GIS) hasgiven rise to sophisticated scientific techniques for collection, analysisand visualization of location based data. These GIS analysis processesare used to reveal some critical patterns of occurrences. Due toinaccurate analysis and covering of insurance risks in Kenya, severalcompanies have closed down prompting the Insurance RegulatoryAuthority (IRA) and Association of Kenyan Insurers (AKI) set upmaximum and minimum premium rates on insurance risks. The setpremiums discounts are given to the insured based on records of theirannual claims. The main problem here is that the rates cover the entirenation without considering the distribution of risk in various regions.The objective of the paper is to show that GIS can be used to analyseand generate auto insurance risk territories for insurance companiesfrom which an insurance rating model can be developed. We used GISanalysis methods such as inverse distance weighting (IDW)interpolation, data smoothing and clustering techniques and data onauto insurance accidents and crime, geo-coded police stations, roads,socio-economic, aerial and satellite imagery for Nairobi County. Arisk territory map showing the distribution of auto insurance risk andother related maps were generated. A prescriptive insurance ratingmodel was then developed that uses generated risk territories tocalculate varying rates for auto insurance premiums rates for the  respective regions. This research shows that GIS techniques can beused for better visualization of risk at a given location for accuraterisk analysis and uptake.Keywords: Auto Insurance Risk Territory, Spatial Interpolation, Inverse DistanceWeighting (IDW), Prescriptive Auto Insurance Rating Model