Data Set Analysis Using Rapid Miner to Predict Cost Insurance Forecast with Data Mining Methods

Abstract

The insurance protection program cannot be separated from everyday human life because there will always be risks in every human activity. Most people have entered into insurance agreements with state-owned and national private-owned insurance companies. The information system is one of the resources to increase competitive advantage. Information systems can be used to obtain, process, and disseminate information to support day-to-day operations and support strategic decision-making activities. The rapid growth of data accumulation has created data-rich but insufficient information conditions. Data mining is the mining or discovery of new information by looking for specific patterns or rules from large amounts of data expected to overcome these conditions. It is hoped that customer data can accurately produce information about insurance cost predictions. In this analysis, the authors use the RapidMiner Studio version 9.1 software. With the RapidMiner Studio app, authors can analyze the insurance data. A scientific novelty of this research is investigating data set cost insurance with data mining techniques consisting of classification, association, and clustering. Research goals for data mining techniques with classification, association, and clustering case studies implemented are to find all associative rules with high confidence, organize objects into groups whose members are similar, and collect objects between them. The following methods can be used: decision tree for data modeling, FP-Growth for determining which dataset occurs most frequently, and K-Means to classify the data attributes to facilitate the analysis.