Improvement of Classification for Agriculture Bibliographic Data

Abstract

In general, Classifying bibliographic data using lexicon-based might not enough in terms of classification efficiency. In this paper, we propose the agriculture bibliographic classification model by improving lexicon set and by using feature selection techniques. The techniques of Information Gain (IG), Chi Squared (CHI) and Gain Ratio (GR) are used in order to select the distinguish properties for feature selection process. Then three algorithms Decision Tree (DT), Naïve Bayes (NB) and Support Vector Machine (SVM) are applied to classify those features. The experiments were done by using 2,580 papers from agriculture publication database. The results show that the proposed method gave better performance than using only lexicon-based in terms of precision/recall and F-measure, respectively 1.3%.