Customer segmentation-based basket analysis with association rule mining combined with SVD and K-Means algorithms

In retail companies, there are many customers and products available. The basket analysis on retail datasets is important for retail companies to recommend new products to their customers in order to increase their sales and profits. In the literature, there are many studies applying basket analysis by Apriori or Frequent Pattern (FP)-Growth algorithms with other methods. There are two common problems in classical basket analysis: huge sparse data, profile-based segmentation. To eliminate huge sparse data, dimension reduction is required while clustering is necessary for customer segmentation. In our recent study, we bring forward a new basket analysis model for association rule mining by using Singular Value Decomposition (SVD) for dimension reduction and K-means clustering for customer segmentation before FP-Growth algorithm. We obtained more reliable purchase rules in our model than classical FP-Growth.