学习云空间中基于情感分析的学习推荐研究

Emotion Analysis and Its Learning Recommendation Applications in the Learning Space Based on Cloud Environments

In the context of the “three links and two platforms”， learning space is important for improving efficient and social learning. Supported by cloud computing and interconnected private spaces， the cloud space wins the advantage of ubiquitous resource supply and convenient interactive support， so a learning cloud space can provide learners with a highly efficient and customizable learning environment. Like other online learning platforms， “emotional loss” has also been existing as a major bottleneck which has important impact on the learning efficiency in the learning cloud space. This paper focuses on the acquisition and applications of learner’s emotion in a cloud space， and presents a method of emotion analysis with a technique of big data based on spatial interactive text， and its learning recommendation mechanism. Firstly， the emotion analysis model of learner is constructed by using the LSTM neural network adaptive to deal with time-series data， and an execution algorithm for real-time emotion analysis is designed. Secondly， after the Bayesian network is applied to interpret the emotion attribution， the emotion-driven learning recommendation mechanism and strategy are established， and the learning recommendation function system， utilizing the above theoretical achievements， is realized in the built learning cloud space platform. Finally， the effectiveness of the system is verified. The experimental results show the proposed approach and corresponding system can better obtain learners’ emotions and recommend learning support services to meet the requirements of personalized knowledge building in customized cyberspaces.