COMPARISON OF VARIOUS RECOMMENDATION TECHNIQUES EMPLOYING LOCATION BASED SERVICES FOR E-COMMERCE WEBSITE

Abstract

As data has grown tremendously over the last few decades, the management of data also follows its footsteps and created the ocean of opportunities for researchers to come up with best data management techniques. One aspect of data management that is widely used now-a-days is in the E-Commerce industry to showcase the relevant items to the users and predict which items the user will be most interested in. This hurdle Race allows the data scientist/Machine learning Engineers/Data Engineers and even Geo Data Scientist to give the user the best experience once he visits the website for online shopping. In other words, every website tries to showcase the limited products that the user might be most interested in rather than displaying all its trillion items and destroying the customer experience as he will get irated in searching for his product of interest. So the website should carefully design its recommended products palette as it can destroy as well as build the customer’s experience. Another aspect is that some portals display the product that are far from the customer but these match the user’s profile the most. It will also be of no use as the delivery cost as well as time will also be a deciding factor whether the customer will buy that product or not. So there is a need to build location based Recommendation techniques. In this thesis work, it has been attempted to incorporate location based services into the traditional techniques i.e. Collaborative filtering and content based filtering and name them as “LOCOL” and “LOCONT” and the displayed results are in very good terms with the customer.