Research And Implementation Of Information Filtering Based On Ontology

Today, the network implies a flood of information, face the sea of information, people often feel helpless, creating opportunities for more and more spam. However,conventional filtration systems, generally are based on keyword matching, there is no semantic extension, so many of our results need to be filtered out. How according to a need for the automatic classification information filtering, information technology has become a hot issue.This paper full use of ontology and semantic web-related theory, ontology semantic model applied to the field of information filtering, and finally achieved an ontology-based information filtering system. Contents of the study about the establishment of domain ontology, storage, reading. The text feature vector generation, vector matching filter algorithm, and several other areas. Ontology-based filtering, semantic relations into the system. So greatly improving the filtration effect. The system's algorithm has algorithm of dominance-class, algorithm of dominance-class instance, algorithm of implicit class, algorithm of hidden instance of a class. Most importantly, it introduces a vector based on keywords scattered fusion technology. Since a single message or a sentence that match the query can be related to various fields of information resource library for comparison. This consumes a lot of query time, but does not meet our requirements. Therefore, the dispersion of all kinds of information, integration is very important. In this paper, the strategy is to source the document with the domain ontology integration, while for the source document keyword feature extraction using distributed. A fundamental solution to the drawbacks of the traditional information filtering.In this paper, ontology-based information filtering. Includes five major modules: document pre-processing module, feature vector generation module, the document feature vector extraction module, matching module, factor estimating module. The system of the model proposed in the strategy and methods for modular and hierarchical, following the hierarchical structure of Semantic Web, for Semantic Web researchers, business groups can provide a reference solution.Finally, we make experiment on filter the key algorithm. In the case of f = 0.7, the accuracy was 97.9%, the recall rate was 96% through experimental data analysis of the results shows that our idea is correct and feasible.