Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Locating online loan applicants for an insurance company

Locating online loan applicants for an insurance company Purpose – This study aims to investigate insurance policy loan applicant characteristics. Additionally, it reveals the behaviour patterns of heavy users who have applied for at least two loans. A policy loan prediction model is established which is designed to increase loan application rates. Design/methodology/approach – The proposed model is implemented using data‐mining techniques and comprises two mechanisms: a business rule generator and a recommendation mechanism. Two analytical approaches, the C.5 and Apriori algorithm, are employed to analyse the profile and browsing log DBs of insured individuals. The prediction model is verified by actual data from a Taiwanese insurance company. Findings – The data‐mining results reveal that five attributes are ultimately used to establish the prediction model, namely: gender, marketing channel, insurance type, area of policy owner, and assumed interest rate. Additionally, the analytical results also indicate that insured individuals apply for loans as a result of arbitrage inducement. The accuracy of loan applicant prediction can exceed 70 per cent. Finally, some interesting patterns emerge for heavy users, such as the finding that loan applicants are used to applying for loans continuously (loan application repetition is on average two to three times). Research limitations/implications – Some policy owners who are unfamiliar with the web interface prefer to contact insurance personnel directly to discuss their insurance needs, and thus no browsing records are available for such users. In such cases only the profile could be collected and analysed. Practical implications – The proposed model enables insurance firms to locate potential loan applicants according to the data‐mining results. As in the illustration scenario in the paper, insurance personnel can contact these potential loan applicants before they submit loan applications to the bank. Additionally, loan‐related information is provided for online insurance users based on their browsing logs. The loan application rate is thus expected to increase, along with interest revenue. Originality/value – As long as the policy proceeds, the interest income from the policy loan seems to be a good option for extending insurance company operational earnings. Understanding the characteristics of loan applicants will provide helpful information. Besides, the proposed mechanism will be more appropriate to online users, who are unwilling to deal with unwanted information. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Online Information Review Emerald Publishing

Locating online loan applicants for an insurance company

Online Information Review , Volume 32 (2): 15 – Apr 11, 2008

Loading next page...
 
/lp/emerald-publishing/locating-online-loan-applicants-for-an-insurance-company-y8s02nU0R2
Publisher
Emerald Publishing
Copyright
Copyright © 2008 Emerald Group Publishing Limited. All rights reserved.
ISSN
1468-4527
DOI
10.1108/14684520810879845
Publisher site
See Article on Publisher Site

Abstract

Purpose – This study aims to investigate insurance policy loan applicant characteristics. Additionally, it reveals the behaviour patterns of heavy users who have applied for at least two loans. A policy loan prediction model is established which is designed to increase loan application rates. Design/methodology/approach – The proposed model is implemented using data‐mining techniques and comprises two mechanisms: a business rule generator and a recommendation mechanism. Two analytical approaches, the C.5 and Apriori algorithm, are employed to analyse the profile and browsing log DBs of insured individuals. The prediction model is verified by actual data from a Taiwanese insurance company. Findings – The data‐mining results reveal that five attributes are ultimately used to establish the prediction model, namely: gender, marketing channel, insurance type, area of policy owner, and assumed interest rate. Additionally, the analytical results also indicate that insured individuals apply for loans as a result of arbitrage inducement. The accuracy of loan applicant prediction can exceed 70 per cent. Finally, some interesting patterns emerge for heavy users, such as the finding that loan applicants are used to applying for loans continuously (loan application repetition is on average two to three times). Research limitations/implications – Some policy owners who are unfamiliar with the web interface prefer to contact insurance personnel directly to discuss their insurance needs, and thus no browsing records are available for such users. In such cases only the profile could be collected and analysed. Practical implications – The proposed model enables insurance firms to locate potential loan applicants according to the data‐mining results. As in the illustration scenario in the paper, insurance personnel can contact these potential loan applicants before they submit loan applications to the bank. Additionally, loan‐related information is provided for online insurance users based on their browsing logs. The loan application rate is thus expected to increase, along with interest revenue. Originality/value – As long as the policy proceeds, the interest income from the policy loan seems to be a good option for extending insurance company operational earnings. Understanding the characteristics of loan applicants will provide helpful information. Besides, the proposed mechanism will be more appropriate to online users, who are unwilling to deal with unwanted information.

Journal

Online Information ReviewEmerald Publishing

Published: Apr 11, 2008

Keywords: Loans; Insurance companies; Data handling; Taiwan

References