A hybrid recommendation algorithm–based intelligent business recommendation system

Abstract

In order to solve the current issues of e–commerce recommendation systems: low accuracy, inflexible, lack personalized, etc. a solution based on hybrid recommendation algorithm is proposed, aiming at building a personalized recommendation system for e–commerce. To make up for the lack of a single recommendation algorithm, we implement the hybrid recommendation algorithm including three algorithms: content based recommendation algorithm, item based collaborative filtering recommendation algorithm and demography based recommendation algorithm. To expand recommendation dimensions, we adopt the algorithms in classification and clustering to mine the historical data of items and users. Then we do the performance evaluation and the result shows that the improved scheme has a better effect of recommendation for E–commerce.

Keywords:

• Recommendation system
• Hybrid recommendation
• Intelligent recommendation
• Collaborative filtering
• Data mining