A Recommendation Engine by Using Association Rules

This study represents a recommendation engine which was developed to personalize an e-commerce website. Here, the personalization approach is collaborative filtering and the technique is association rule mining. The software was developed by the programming language C# and association rules were generated by Apriori algorithm. The recommendation engine had been tested by existing data before it was deployed to an e-commerce website. Testing phase was evaluated by accuracy and coverage while the deployment phase was evaluated by basket ratio, which is the ratio of the number of products added to the shopping cart to the number of keywords searched by users. The application has taken three weeks. Results show that the recommendation engine increases the basket ratio.