اسم الباحث     :    Mossa ghurab Hiba Mohammed Al-Marwi
سنة النشر     :    2016
ملخص البحث     :   

Abstract:- with a rapid expansion of new available information presented to us on the internet ,information retrieval gaining importance. Several approach introduce in this filed which record success ,but it take long time. in this paper we discuses a model based on ontology which we built using naives bayes algorithm, and apply k-means clustering algorithm to find semantically similarly terms on the Ontology ,Also we present document using improved concept vector model   ICVS  which is improved model for traditional CVS also we apply fuzzy classification to limited number of retrieval document to particularly threshold.

Index Terms—concept Relevance, Concept vector space, Query Expansion ,K-means clustering, Naivse Bayes .

I.     INTRODUCTION

As number of document on the internet increase in everyday life. Traditional approach on information retrieval proven to be less efficient in providing relevant information to user query .Several methodology appears to generate more accuracy result. To achieve this goals new upcoming semantic approach try to establish  a semantic relationship among the document . In semantic web information is stored in conceptual hierarchy referred as ontology to build some domain specified ontology. we use Naïve Bayes algorithm . Many researches are developed to enhance Naïve Bayes [1] [2] [3].We use naives bayes using map reduce to enhance its performance[5].

Text document are represent using vector space model[4]. Vector space model consist of concepts extracted from ontology and concept relevance which is calculate using frequency of concept occurrence . We use improved vector space model  ICVS [4]. which take into account both concept frequency and concept important which computed using page rank algorithm as one of the most well known of markov based algorithms[4].

Fuzzy logic very useful in information retrieval .IR system find difficulty to make decision in providing accurate information .Each element of fuzzy set has a membership value between 0 to 1.This membership function play important role in defining degree of membership of element in fuzzy set.

Rest of the paper is organized as section 2 provide related work; section 3 presented discuses and section 4 conclude paper and finally section 5 presented my opinion.

II.       RELATED WORKS

There is a lot of research in IR field such as[1][2] [5] which extract concept from internet using Naïve bayes algorithm which used as preprocessing step to build specific domain ontology.

Another research focus on enhance models which used to represented document as helper technique to enhance information retrieval[4 ].

In [6] research developed approach to get more relative document to user query by using k-means clustering algorithm fuzzy classification ,query expansion technique .