Customer Behavior Predictive Modeling for Online Shopping using Tuned Decision Tree Method

Author Name(s): *D.Kalaivani, P. Sumathi, T.Arunkumar
Author Email: kalaithirunavukkarasu@gmail.com

Abstract

The crucial challenge of online shopping is to study and analyze the behavior of the buyers. The success of business is to know the requirement of the customer and providing the good offers in the right time. The development of the Internet usage is enormous. Therefore changes in the people mind leads to traditional buying behavior to online purchase. The younger generation focusses more on online shopping. It increases the online sales drastically every year. The focus of this study is to propose the novel Tuned Decision Tree Predictive model to analyze the customer behavior in online shopping. The proposed model Tuned Decision Tree Predictive model gives the accurate result and therefore prediction quality is certainly improved. The performance of proposed model is compared and validated with several online shopping data sets. In this work novel tuned decision tree predictive model is proposed and evaluated to analyze and rightly predict the customer behavior and therefore to optimize the Business outcome

Keywords

Customer Behaviour, Online Shopping, Data Mining, Decision Tree, Predictive Model, K-means algorithm, Apriori Algorithm, Cluster Analysis

Introduction

In early days, the online shopping is used to buy the limited items by the customers. The life-style of the people is not allowing them to spend time to buy the required items through traditional shopping. Owing to their busy schedule, most of them like online shopping to purchase the items like electronic items, furniture, books and dress are purchased using online shopping mode. The online shopping becomes very competitive today because of the involvement of multiple vendors in the business. Many vendors involved in the online shopping and they are spending lot of money to improve their online business to understand the customer online shopping behaviour. The item is available in multiple sites with different cost. To find the best offered cost of the item, people use to visit multiple vendor sites, and they will decide to buy the item with low-cost and good service. To improve online shopping, the online owners shown interest to find potential buyers based upon their transaction history and the number of time the web page is visited by the customer. The customer purchase behaviour is analysed based upon many factors such as the number of web page clicks, duration of the session and number of clicks per session.

Conclusion

Every company needs to stay competitive by using analytics for understanding their customers.  They need to adjust the content, design and advertisement to better serve their target personas. The capacity to identify the wealthy customers in their purchasing is deeply affected by the perceptions acquired the moment before the buying decision is made. Hence to escalate a predictive decision making solution close to a specific product or services would be greatly effective to rise ROI, conversion and customer happiness. In this work novel tuned decision tree predictive model is proposed to analyse the customer behaviour to improve the business. The performance of the proposed model is compared with existing predictive model and validated with the real-time data set.

1,407 total views, no views today

Download PDF File

About the author: tej