Resource Recommendation Algorithm Based on Text Semantics and Sentiment Analysis

Traditional recommendation systems rarely take the contextual semantics of the application scenarios into account when implementing the resources recommendation, which results in those algorithms having serious deficiencies in real-time, robustness, and quality in the actual learning circumstance. On the other hand, sentimental factors and individual preference also have great impacts on users' demands. The objective of this study was to determine a resource recommendation scheme based on the semantic similarity and sentiment analysis of review text. Extracting the semantic and sentiment information of the resources, filling user rating matrix, and calculating users' similarity with adjusted cosine measures, will obtain the personalized recommendation results. Experiment results demonstrate that the proposed algorithm can better characterize user preference by obtaining information-in-depth, and outperforms the state-of-the-art methods.