Hybrid Approaches for Customer Segmentation and Product Recommendation

The trend of influencer marketing is increasing massively in the digital age. Despite this, the majority of marketing strategies are still not optimally using this trend. Marketing strategies can be implemented by applying customer segmentation and product recommendations through the implementation of artificial intelligence technology. Customer segmentation can be implemented using clustering techniques, and product recommendations can be applied using association rules. The clustering methods K-Means and K-Medoids will be compared to find the best clusters, and then each clustering method will be combined with the Apriori association rules method. The objective of this research is to identify the combination of clustering methods and association rules that produces the best performance. The elbow method is used to determine the optimal number of clusters. Based on three used evaluation metrics, it is found that between the two clustering methods, K-Means has a better evaluation metric. After the associations are combined, the result shows that K-Medoids and Apriori have a slightly higher confidence score than K-Means and Apriori with 39.265s for runtime and 244.175Mb for memory. This research will help companies identify various groups of customer characteristics and products that have strong ties to these groups. Therefore, this will help companies gain the necessary business insights.