Self Attentive Product Recommender – A Hybrid Approach with Machine Learning and Neural Network
Devanshu J. Dudhia, Sonal R Dave, Shweta Yagnik · 2020 International Conference for Emerging Technology (INCET) · 2020
People are choosing products from online ratings and comments. Experience from Amazon, Flipkart and other leading online shopping portals in India, the buyer `s product choice is mostly based on the other buyer's review. We have found an interesting case study of Netflix video recommendation based on so many criteria. Product recommendation is one of the demanding area of recent time where efficiency of prediction of which product a buyer can choose over other hundreds of product is a challenging task. Artificial intelligence has helped researchers in developing algorithm that has self-aware method for machine with machine learning, deep learning and natural language processing. In this research, we are introducing a hybrid approach of recommendation for products. There are many ways to find out the people who have similar choice and combining their choices can lead us to suggestions for other products. Collaborative filtering method has been being used for Recommender frameworks for quite a while now. They have been fruitful in comprehending numerous issues of the systems which are being used in the market. User- behavior analysis, sentiment score, product reviews, popularity score can be a decisive factor along with neural network with classification method can lead to more efficient results. In this research, we present some of more potential areas of working on Collaborative Filtering technique with machine learning and deep learning techniques. Self attentive product recommendation is one such technique which focuses on automated form for recommendation which is independent of a dataset and its data type. We have examined other approaches for joining different calculations for predicting client evaluations and furthermore examine a few outcomes from the investigation of different procedures utilized by earlier analysts and discover answers for them.