An Investigation of Summary Based Classifiers

The work presents a comparative investigation of the performance of some of the well-known machine learning classifiers in the task of summarization. The objective of this paper is to evaluate the capability of classification algorithms in predicting summary sentences and compare its prediction performance against ten well-known machine learning models in the context of the DUC 2002 dataset. Classical classification algorithms based on Trees, Rules, Functions, Bayes and Lazy learners are used in the study. We have used 350 text documents from DUC2002 (Document Understanding Conference 2002) for prediction. This paper evaluates the capability of classification algorithms in predicting candidate sentences for summary generation. The results indicate that the prediction performance of machine learning classifiers in the task of text summarization is better than the baseline performance.