Life Insurance-Based Recommendation System for Effective Information Computing

Life Insurance-Based Recommendation System for Effective Information Computing

Asha Rani (GGN Khalsa College, India), Kavita Taneja (Panjab University, India) and Harmunish Taneja (DAV College, India)
Copyright: © 2021 |Pages: 14
DOI: 10.4018/IJIRR.2021040101
OnDemand PDF Download:
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Due to the rapid advancements in information and communication technologies, the digital data is exponentially growing on the internet. The insurance industry with tough competition has emerged as information rich domain based on health, assets, and life insurance for public. Customers expect to receive personalized services that match their needs, preferences, and lifestyles. But a large portion of population is still unfriendly to the insurance selection. Major reasons could be the time and complexities involved in selection of suitable policies. This paper presents the state of the art of the research done in insurance recommendation systems at national and international levels. Multi-criteria decision-making methods are compared with collaborative filtering and data mining techniques. Their suitability to the field of life insurance recommendation is analyzed. The paper identifies the lack of public dataset of customers and life insurance policies and highlights the need for a personalized, neutral, and unified model for effective information computing for life insurance recommendations.
Article Preview
Top

Introduction

World Wide Web is a repository of heterogeneous data both in the structured and unstructured format. The information derived from such type of data has resulted in vigorous information exchange over the web (S. Dhuria, et. al., 2016; Sinha, S., et. al., 2017). Information computing for life insurance recommendation needs to draw insights from data about individual preferences, behaviors, attitudes, lifestyle details, and hobbies to create personalized offers and policy recommendations. This huge volume of data necessitates the use of efficient, effective, and contemporary techniques to retrieve process and manage this data from multiple dimensions. On one side, an ample amount of information is desired to derive good decisions and on the other side, too much information induces the problem of information overload. Information retrieval (IR) works for the user by retrieving the relevant information and ranking it as per its relevance and attractiveness for the specific application areas. An essential requirement for today’s IR is to deal with the heterogeneity of data and adapting to aggressively dynamic conditions (Suri, Pushpa et. al., 2010). There is large information flow and hence filtering is required. The recommender system has come up towards the rescue of this problem (H.T. Cheng, 2016). Recommender the system is a subclass of information filtering system which gives suggestions on the basis of matching of 'preferences', 'ratings' or 'attributes'. Recommender systems are working in almost every field like music, movies, restaurants, online shopping, online games, books, research articles, a wide range of services, and so on. (F.O. Isinkaye, et al.,2015). Commonly used recommendation systems have the ability to predict items as per user choice, which are calculated based on explicit (rating score) and implicit data (Logs) available for users. ‘Figure 1’ shows the general outline of recommender system. The implicit and explicit preferences and user ratings are used to create user profile which is matched with the item features. This matching and filtering provides the desired recommendations.

Figure 1.

General Outline Structure of Recommender system

IJIRR.2021040101.f01

Recommendation techniques can be categorized in Collaborative filtering based (recommends items by identifying other users with similar taste and uses their opinion to recommend items to the active user), Content based filtering (recommends items that are similar to items previously preferred or rated positively by user) (Balabanovic M. et al.,1997; Binder, S. et. al., 2017) Knowledge-based filtering (Knowledge representation, extraction, and system design techniques) (Sae-Ueng, et al., 2008), Context awareness-based filtering (identified attributes which are known a priori and illustrate the context) (Gediminas A., et al., 2011) and hybrid filtering techniques (the combination of two or more techniques) (Balabanovic M., et al., 1997).

Financial liberalization have shown a rapid growth of insurance demands in our country which has resulted in the emergence of a considerable number of insurance companies and their insurance products. The global insurance industry is experiencing phenomenal growth with the growth in insurance companies and their products. Also, the price comparison websites, digital revolution and other technological developments have proved to be tectonic shifts (Binder S., et al., 2017). Indian Government has also increased the foreign investment limit from 49%to 100%on insurance intermediaries, concentrating on the insurance sector (Indian Union budget, 2019). Insurance-based recommender systems facilitate the companies, customers, and agents to identify the client risks, mark the up-selling and cross-selling opportunities and offering a simple and transparent platform for insurance products, insurance literature and marketing facts and figures to boost the sales cycle. The Recommender system in insurance domain rescues the customers from confusion, agent's biased information and commission and wrong selection of products and provides suggestions based on personalized requirements. Insurance recommender system suggest adequate solutions on the basis of customer's demographics, interaction history, stated preferences, similarity in clients or the similarity between the content of policies (Shishehchi, et. al., 2012) The traditional recommender systems are developed majorly for e-commerce and could not be directly employed in the insurance sector because of domain size, item complexity, customer expertise, demographic features, interaction frequency, and user constraints.

Complete Article List

Search this Journal:
Reset
Volume 12: 4 Issues (2022): 1 Released, 3 Forthcoming
Volume 11: 4 Issues (2021)
Volume 10: 4 Issues (2020)
Volume 9: 4 Issues (2019)
Volume 8: 4 Issues (2018)
Volume 7: 4 Issues (2017)
Volume 6: 4 Issues (2016)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing