Document Classification: A Novel Approach Based on SVM

摘要

Science and Engineering fields are very large and important areas of research in which the scientists are conducting studies and experiments, as a result a voluminous datasets are generated from such research. An example is the biological databases like the one we use in this paper, the PubMed database, it contains millions of articles. In this paper we discuss the main challenge for these types of databases which is the problem of retrieving information that are relevant to a specific topic from a huge collection of articles. In other words are classifying those articles. In this paper we present a generic algorithm that can classify biomedical articles efficiently related to Minimotifs. Our algorithms are based on algorithm that has been used for a related problem called SVM. We are comparing our algorithm with the Gene selection algorithm that uses the SVM in other ways. The Gene selection is the problem of identifying a minimum set of genes that are responsible for certain events (for example the presence of cancer). Our proposed algorithm will take as input a set of articles (that characterize the information of interest) and will develop a learner model that will identify a small subset of the keywords that are capable of classifying papers into two types. The first type is the articles that have information of interest the second type are the articles that don't have information of interest. Experiments show that our new algorithm give a higher classification accuracy using a smaller number of selected keywords when compared to one of the best algorithms reported in the literature (Gene Selection).