Mining Consumer Behavior Patterns in E-Commerce Using Apriori Algorithm and Sequential Pattern Analysis

The consumer behaviour in e-commerce is the main focus of this research and data has been collected using Apriori Algorithm association rule mining method followed by the use of sequence analysis. The work applies real large transactional databases to generate the frequent items and the next sequences of the consumers’ buying habits. Looking at association and sequence mining in Python using MLxtend and Seq2Pat libraries named those consumer patterns that are relevant, for instance, smartphones often go hand in hand with the screen protectors and chargers are advised that they have support of 10%, and a confidence level of 85%. The subsequent further sequence analysis revealed that 4 percent customers who bought smartphone also bought chargers thereafter headphones reveal an appropriate sequence. The proposed two paradigms static (association rules) and dynamic (sequential patterns) were tested for validation and the results for Prec = 82% and Recall = 85% signifies that the proposed model will prove efficient and accurate and hence can be used for predicting patterns. Together, these findings will enable the e-commerce platform providers to enhance and/or optimise channel explicit Eric & Baumeister (2011) approach, delivery mechanism and/or recommendation system of their customers as well as tailor their stocks according to the flow of customer movement. This work highlighted the need for the combined approach of association and sequence mining analysis to enhance knowledge regarding the consumer behavior and offers a feasible source for ecommerce platforms to enhance customer experience. This approach can be enriched in future work and adjusted further to obtain better reactions on current changes in behavior of consumers by applying real-time data streaming. The above research framework is helpful for both business and consumers in digital marketplace as it provides right time to market products and services using data which is consistent with consumers’ actual behavior.