Purchase Recommendation and Product Inventory Management using Content Based Filtering with Sequential Pattern Mining Approach

Abstract

Today, the product sales at XYZ Bookstore are increase in accordance to the trend in society. In that case, high sales must be supported by good supply and on target. Product sold based on needs of consumers will make possibility to achieve high sales. Using the Sequential Pattern Mining approach, we can specify sales patterns of products in relation to another products. SPADE (Sequential Pattern Discovery using Equivalence classes) is an algorithm that can be used to find sequential patterns in a large database. This algorithm finds frequent sequences of the sales transaction data using database vertical and join process of the sequence. The results of SPADE algorithm is frequent sequences which are used to form the rules. Those can be used as predictors of other items that will be purchased by consumers in the future. The result of this study is a lot of unique sequence appears that can provide the best advice for Merchandiser Officer, for example, there are 1.468 sequences that prove the customer who bought the product in Children’s Book category will always bought the same thing in the others day. This research produce some recommendation, one of the recommendation is Children's Book category has a very high chance of being a Best Seller for a long time so that the purchasing officer on XYZ bookstore should ensure that the product's supply of the category is always safe throughout the year. It means SPADE is successfully used to provide the advice and Merchandiser Officer must ensure the stock of that product is always available to avoid Lost Sales.