Skip navigation
st. Mary's University Institutional Repository St. Mary's University Institutional Repository

Please use this identifier to cite or link to this item: http://hdl.handle.net/123456789/6243
Full metadata record
DC FieldValueLanguage
dc.contributor.authorHaile, Mebeki-
dc.date.accessioned2021-09-24T07:49:16Z-
dc.date.accessioned2021-09-24T07:49:17Z-
dc.date.available2021-09-24T07:49:16Z-
dc.date.available2021-09-24T07:49:17Z-
dc.date.issued2021-01-
dc.identifier.uri.-
dc.identifier.urihttp://hdl.handle.net/123456789/6243-
dc.description.abstractThe main objective of study was to classify medical insurance customers with high claim ratio in order to take an appropriate measures during underwriting process to save profit making customers under medical insurance class of business. Globally insurance companies are spending high amount of claim costs due to medical insurance. It is a concern for companies to have a system that could differentiate whether the customers are profit making or loss incurring from upcoming claims. In the insurance industry the claim costs are needed to be minimized as much as possible. The main cause which result in high claim costs knowing profit making and loss incurring customers without the knowledge of claim experience in the company. To tackle the problem of high claim cost in medical insurance class of business, predictive data mining techniques has been employed using Support Vector Machine, Naïve Bayes and Logistic Regression predictive models. The dataset used for the experiment in this study was collected from Awash Insurance Company specifically from underwriting and claim data tables of medical insurance class of business. After cleaning irregularities and incomplete data in the dataset, a total of 41,151 records have been used to train the models in the ratio of 80:20. To meet the aforementioned objective of the study, the CRISP-DM methodology, which involves six steps was adopted to undertake data mining process and to address the business problem systematically and iteratively. A six steps process model is used to guide the entire knowledge discovery process. Support Vector Machine, Logical Regression and Naïve Bayes classification algorithms are used to build predictive model. Experiments are conducted and the resulting models show that the Support Vector Machine (SVM) is found to work well in classifying medical insurance customers with 99.39% classification accuracy. A prototype is developed based on the predictive model. Finally recommendations and future research directions are forwarded based on the results achieved.en_US
dc.language.isoenen_US
dc.publisherST. MARY’S UNIVERSITYen_US
dc.subjectPredictive data mining, CRISP-DM, medical insurance class of business, SVMen_US
dc.titleApplication of Data Mining to Classify Medical Insurance Customers Based on Claim Experience: The Case of Awash Insurance Company S.Cen_US
dc.typeThesisen_US
Appears in Collections:Master of computer science
Master of computer science

Files in This Item:
File Description SizeFormat 
Mebeki Haile(Final thesis).pdf1.48 MBAdobe PDFView/Open
Show simple item record


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.