Security on electronic transactions in developing countries: A cluster and decision tree mining approach

Abstract

The dawn of information technology has impacted tremendously on customer information management and marketing activities of companies. The huge amount of data generated by companies, its management and evaluation has posed a great challenge as well as offered opportunity to businesses to gain competitive advantage. The opportunity created by the advent of the internet has been a blessing and a challenge. It is a blessing since businesses reach a wider coverage of customers and a challenge because the security of customers' data on Electronic Transaction (eTransactions) need to be protected. The increase in eTransaction activities especially in developed countries has been tremendous in recent years. The correlative increase in developing countries has however been very slow. Many researchers have attributed this slow growth of eTransactions to a couple of obstacles with security of transactions prominently featuring in the research outcomes. There is a general outcry of security and cyber fraud concerns on eTransactions in developing countries. Research on the category of consumers who are most concerned about security of transactions has however been limited. This paper has researched five sub-Saharan African Countries (Kenya, Ghana, South Africa, Nigeria and Zambia) to elicit their views on security of eTransactions and their readiness to shop online. The paper surveyed 600 people across business sectors, age group, income levels and educational qualification in the countries. The paper, using clustering and decision tree analysis of data mining, (1) discovers clusters mostly associated with security of eTransactions in Sub-Sahara Africa, (2) identifies the variable that is more congruent with security as per eTransactions of citizens in Sub-Sahara Africa, (3) details the category of users primarily and mostly concerned about security in eTransactions in the sub-region, and (4) presents a predictive model for predicting customers with likely eTransaction security (payment/data) concerns.