Online shopper-vendor interactions: Identifying interestingness relations using clustering and association rules mining

Online shopping as a form of electronic commerce worldwide, has come to stay. While many studies have been authored on the motivation that drives the daily interactive transactions, the interplay between shoppers and vendors continues to increase in complexity, especially with the regular introduction of many advanced online services and features. Among the issues that have also increased the complexity are how different demographic profiles and geographic backgrounds impact online sales and customer retention. The interest in shopping online is motivated by a reduction in key barriers and the trust cultivated in an e-shop as compared to other alternatives. Against this background, this paper uses the a-priori Algorithm in Association rule mining to discover strong rules between variables of interest in a large dataset of e-shoppers in the Czech Republic. With a higher support and lift, the results reveal a high level of confidence between the reasons that propel the customer to start buying via the Internet, criteria for buying items on the Internet and the information gathered before the initiation of an online transaction. The paper further conducts a comparative analysis of shoppers' attitudes to foreign e-shops and local ones. In addition, the paper uses a clustering algorithm to segment demographic variables of respondents against the inclination of local Czech e-shoppers attitudes to foreign e-shops and their local shops. The results indicate that millennial Czech e-shoppers tend to patronize other foreign e-shops, especially female respondents, while adult e-shoppers are conservative when it comes to e-shopping. Finally, the paper proposes managerial recommendations for e-vendors in the Czech Republic on how best they can incorporate e-shoppers into their fold, given the notion that online shopping is here to stay. A total of 1,602 shoppers in the Czech Republic responded to the questionnaire, generating the data forthe analysis