Enhanced Product Recommendation System for e-Commerce Using Machine Learning

EasyChair Preprint 9176

Abstract

In the modern information technology age, finding users favorite product in large application databases becomes a serious issue for Recommendation developers. Recommendation means providing relevant suggestions to the user as per his/her interest and their needs. In this proposed Model we combined the corrwith() method which computes the Pearson correlation coefficients with our proposed collaborative filtering method which uses SVD++ that will help us in improving accuracy and targeting all sorts of users and recommends products supported ratings, previous purchase history, products on sale and recently viewed products etc. The experiments are performed on Amazon products dataset which consists the meta data of product reviews, feature, ratings, sales rank and similar products etc. The results of recommendations with this approach not only provide recommendations for specific products, but also provide recommendations on categories like groceries and gourmet foods. When a user selects a product, the model suggests related products that other users have also purchased with that product according to ratings, reviews and user’s purchase history. Thus, the recommendations are going to be more varied and in line with user interest. This model paper also discusses about various shortcomings in the current recommendation techniques and suggests potential solutions that that could enhance the existing recommendation systems used by e-commerce websites.

Keyphrases: Recommender System, collaborative filtering, e-commerce, user ratings, user reviews

@booklet{EasyChair:9176,
  author    = {R.M Rani and Sourav Singh and K Abinav and Suresh Krishna},
  title     = {Enhanced Product Recommendation System for e-Commerce Using Machine Learning},
  howpublished = {EasyChair Preprint 9176},
  year      = {EasyChair, 2022}}