Text abstraction based on user intent and deep analysis

In recent years, text analysis is mostly inaccurate and incomplete. To solve the above problem, a text abstraction method based on user intent and deep analysis is proposed. This method analyzes user intent, constructs user intent subtree by ontology theory, fully understands the various forms of user search keywords, mines and integrates features that conform to reality, uses deep learning model to train, and outputs text information that meets the requirement. Experiment shows that compared with keyword method and user intent method, by using the text abstraction method based on user intent and deep analysis, the number of returned result of abstraction text on related topics is higher, which indicates that the accuracy rate has been improved to a certain extent.