Angle Prediction between Document Vector and Ontology Vector, Using Multiple Linear Regressions

Considering the growing development of information at World Wide Web, the users find it difficult to have access to the documents s/he requires. The purpose of this paper is to present a method for making the search by user more systematic and limited using some statistical techniques. For this purpose, we will present a formula by multiple linear regression models in order to model the relation between lexical objects and ontology. Then for stating ideas on a sample document, we count view values in that document, which are conforming to lexical objects in ontology, and next we will form the document vector. With having optimized document value in the formula out of multiple linear regressions, we can predict the degree of angle between the document vector and ontology vector. The closer the angle to zero, the more relation the document has ontology. Experimental Results show the recommended method would be able to distinguish 100% accuracy of this angle.