Automatic generation method to derive for the design variable spaces for interactive Genetic Algorithms

In the growing E-commerce market, many online shopping sites have adopted product recommendation systems to expand their business opportunities. We have focused on iGA (interactive Genetic Algorithm) as a solution to product recommendation algorithms. Although iGA is an optimization technique for user's preference through the interaction between systems and users, iGA requires extraction of design variables to apply the product recommendation. Extracting design variables from existing products by hand is unrealistic because there are a wide variety of products on shopping sites that are updated rapidly. To address this problem, we have proposed an automatic generation method of a design variable space based on the collective preference data on the web. In this paper, we introduce several design variable spaces generated by books on Amazon, which maintain recommendation relations among products. The distributions of books generated by the proposed method are influenced by their authors. Then, subjective experiments confirmed that test subjects searched solutions by their unique preference.