@article{1019,
  author = {Sher Afgun Khan, Muhammad Tanvir Afzal, Muhammad Abdul Qadir},
  title = {Metadata Based Classification of Scientific Documents},
  journal = {International Journal of Information Studies },
  year = {2012},
  volume = {4},
  number = {4},
  doi = {},
  url = {http://www.istudies.net/journal/sites/default/files/v4%20n4%20a3_0.pdf},
  abstract = {Finding a relevant paper has become a problem in exponentially growing web repositories of scientific documents, where scalability, semi structured information and evolving knowledge diffusion terms are escalating it. As a result, a user may receive millions of irrelevant hits through web interfaces like Google Scholar, CiteULike, CiteSeer etc. Proper Classification of these documents can address the problem with better results subject to the provision of up to date scientific document classification system and automated capability of correctly classifying the evolving terms like wikification, tweets, etc. Classification systems or taxonomies are being used in different domains like ACM classification in computer science. However, it lacks automatic updates and incapable of classifying modern scientific terms accordingly. Several approaches are being used for creating automatic classification i.e. co-occurrence analysis, keywords analysis, and using domain models ,etc. However, the current approaches have their limitations. In this paper, we proposed and implemented a new approach of finding the evolving classification, termed as concept classifier for the scientific community by finding the correlation between title terms and keywords of the paper. Initially the experiment has been conducted over the dataset of journal i.e. Journal of Universal Computer Science (J.UCS) .The resultant taxonomy has been verified and compared with the Growbag dataset (DBLP dataset of co-occurrence terms). The developed concept classifier is useful for classification of scientific documents.},
}