Building a Recommendation System of Consumers' Preferable Choice Using a Collaborative-Filtering
Nishanthinee Velloo, Noor Farizah Ibrahim, Zuraini Zainol · 2023 17th International Conference on Ubiquitous Information Management and Communication (IMCOM) · 2023
Nowadays, one of many challenges in e-Commerce is to target the right customer by recommending relevant products or services that suit their preference. To overcome this challenge, various sources have been utilised including customer comments and reviews from social media to analyse their preference based on their interests and sharing. Numerous large e-Commerce web sites are now investing in recommender systems to ease their customers into recommending the relevant products that suit their preferences. However, there are still limitations within the existing recommender systems that are lacking in discovering the real sentiment values of how customers perceived the products or services offered. This study aims to develop a product recommender system that incorporates the sentiment values from customer post on social media and predict the top five recommended products using the Collaborative-Based filtering technique. The tweets were collected from top global and local retailers, Amazon and Shopee to understand the customer response from social media followed by the development of a recommender system for the customer. The findings of this study highlighted a total of 8 clusters such as electronics, entertainment, and books from the clustering experiment. The top 5 best seller product predictions for each category also showed a score of 0.59, which is higher than the threshold score in the evaluation stage.