Product Recommendation System Design Using Cosine Similarity and Content-based Filtering Methods

Abstract

The wide variety of products offered by a company, combined with the consistent demands of specific products from customers, create a certain problem for the organization when they want to market a new product. Organization need information that could help them promote the most suitable product based on their customer’s characteristics. The organization also need to suggest alternative products for customer if the requested product is unavailable. In this research, we design a Recommender System that could suggest either new or alternatif products to customer based on their characteristic and transaction history. This proposed system adopts Cosine Similarity method to calculate product similarity score and Content-based Filtering to calculate customer recommendation score and used as a model for the proposed system. Subsequently, these models are used to classify customers as well as products according to their transaction behavior and consequently recommends new products more likely to be purchased by them. Based on the testing results of the proposed system, it can be concluded that the chosen methods can be utilized to recommend products and costumer of products. It is shown that Precision and Recall of product similarity scores and customer recommendation for product scores are 100% and 93.47%.