Content based apparel recommendation system for fashion industry

Abstract

Present apparel e-commerce system that encourage online shopping,has major issues to deal with catalog based online shopping. As there is a lack of customized services, the users may face difficulties to find discrimination over different types of retailers available on electronic product catalogs, they may also be confused with complex navigations that redirect to other pages based on their selection. This drawback can be overwhelmed by following suggestions on categories that they have chosen or from the products that they have already viewed. Multiple number of online marketing companies around world-wide,has been practicing the naive method for apparel marketing website. This paper aims to simulate this recommendation system on real world data set taken from the marketing giant, Amazon’s Product Advertising API, in a policy compliant manner by following the procedure in three steps:Analyzing the data to select the pivot for the recommendation system, Data preprocessing to remove invalid sections and to implement and find appropriate choices among the techniques like Bag of Words(BoW) and TF-IDF for better recommendations.