A hybrid model for recommender system in e-commerce
Teng-Chun Yu, Shie-Jue Lee · 2023
With the popularity of the Internet, e-commerce has grown rapidly. Many online applications use recommender system to predict user preferences and make predictions on items based on user preference information to provide a better service experience and increase sales of items. Since online applications often add new items into the system, it is difficult to recommend new items to users without any feedback from them. Similarly, when a new user registers to the system, it becomes very difficult to recommend items to the new user because there is no previous purchase history of the new user. This condition is known as cold-start problem. There are two types of cold-start problem: user cold-start problem and item cold-start problem. In order to solve the cold-start problem, we propose a hybrid approach that combines traditional machine learning methods and neural network techniques. For the user cold-start problem, we extract additional information between users and items and convert the information into latent features as a basis. Experiments will be done with real-world datasets to verify the effectiveness of this method on the user cold-start problem.