A Study on Development of Time Series Warning Module in Warranty Claims Database

Abstract

Warranty for a product purposes to improve product quality and reliability as well as consumers' satisfaction. Therefore, a producer analyzes field reliability data. There exist many sources to collect reliability-related data. However, most of them need high cost, or an experiment cannot be continued until a product life-cycle ends. Thus, from the perspective of a producer, the warranty decision-making system which uses age-based warranty claims data needs to be developed and operated since it is easy to analyze most cases statistically. Nevertheless, we have problems that the statistical hypothesis is too stubborn in age-specific analysis that uses such failure data and that there is a high possibility of statistical errors in realistic terms. Therefore, the study suggests to using a decision-making algorithm that determines an early warning module about collected warranty claims data with a simple time series method. The suggested algorithm can be useful as a pre-process stage for an age- based claims analysis, an aggregated warranty claims analysis and a warranty claims analysis using covariates.