Research Of Papers Automatically Generated System Based On Ontology

The semantics of current resources on the web can not be recognized precisely by machine. To find some proper methods for handling them, this paper takes the application in the field of computer teaching as example to design architecture for ontology based on papers automatically generated system and implement the generation papers classifier. The methods that we propose play an important role on fully sharing and smartly utilizing information in computer teaching area. The method used in this paper can also be referred in other areas.Firstly, we study the background of ontology based on papers automatically generated system and overview the current application in computer teaching area in this paper. Through deep research in ontology technology, we decide the suitable ontology description language and ontology constructing method of questions ontology samples and so on.Secondly, we describe the architecture for ontology based on papers automatically generated system. We present how to build Ontology in computer teaching area and generate partial ontology samples of papers in computer teaching area by using OWL DL. And we propose a new entropy-based closest-neighborhood algorithm of attributes weighting fittest for generation papers by using questions ontology repository.Finally, according to the entropy-based algorithm of attributes weighting proposed in this paper, we code it in java and conduct an experiment. We take the computer introduction and computer english as examples to run the experiment in Eclipse. Then we compare the result of the experiment with the result of IBK, naive Bayes, Bayes network, J48 decision tree and AdaBoostM1 methods in weka on the same dataset. The experiment demonstrates that the accuracy of ontology based on papers automatically generated system in this paper is finer than that of other system, which elucidate the significance of this research.