Recommender System for E-Commerce Application based on Deep Collaborative Conjunctive Model

Nowadays, recommender systems play a crucial role in human lives. The recommendation process is involved in many items and many users' decision is based on this process. Collaborative filtering technique is one of the widely applied techniques in various types of recommender systems that uses the reviews of products and services. Word2Vec is adopted to extract information from the users' comments made on the items they bought. To group the items into definite sets, the clustering algorithm is used. A deep collaborative conjunctive recommender (DCCR), which is a hybrid approach. It consists of the combination of an autoencoder which extract the user and item features along with a multilayered perceptron which represents the interaction them. By applying together, the modern technique of deep learning and the traditional technique of collaborative filtering, the proposed system extract some set of the features like the features from users, explicit and implicit ratings to items along with user information. The proposed system uses real-world datasets to achieve effectiveness.