The Joint Framework for Dynamic Topic Semantic Link Network Prediction

To explore the maximum potential of textual data, a well-organized dynamic semantic structure of the topics is in fact of great importance for effectively supporting the advanced intelligent application. The proposed framework joints the Gaussian mixture model and the Bayesian network to conduct inference and prediction of topic relationships of a dynamic topic semantic link network. The approach is to identify the relationships between the topics and to infer the condition-dependent topic relationships for predicting the topic semantic link network structure, which not only describes the relationships between the topics under changing-dependent conditions but also provides a broader understanding of the relationships between the topics in dynamic evolution processes. The results of the evaluation and experimental analysis indicate that the proposed approach is effective, feasible, and well-suited to predict the dynamic and multi-dimensional relationship structure of topics.