Classification of Full Text Biomedical Documents: Sections Importance Assessment

The exponential growth of documents in the web makes it very hard for researchers to be aware relevant work being done within scientific community. task efficiently retrieving information has therefore become an important research topic. objective this study is test how efficiency text classification changes if different weights are previously assigned sections that compose documents. proposal takes into account place (section) where terms located document, and each section a weight can modified depending on corpus. To carry out study, extended version OHSUMED corpus with full have been created. Through use WEKA, we compared abstracts only texts, as well weighing combinations assess their significance article process using SMO (Sequential Minimal Optimization), WEKA Support Vector Machine (SVM) algorithm implementation. experimental results show proposed preprocessing techniques feature selection achieve promising document classification. We also evidence conclude enriched datasets from certain better than titles abstracts.