Multi-document Text Summarization: SimWithFirst Based Features and Sentence Co-selection Based Evaluation

Document summarization is an emerging technique for understanding the main purpose of any kind of documents. To visualize a large text document within a short duration and small visible area like PDA screen, summarization provides a greater flexibility and convenience. In this paper we study various text summarization techniques e.g. RANDOM, LEAD and MEAD. Then, we propose two techniques for both single and multi document text summarization. One is adding a new feature SimWithFirst (similarity with first sentence) with MEAD (combination of centroid, position, and length features) called CPSL and another is the combination of LEAD and CPSL called LESM. Finally we simulate and compare the results of new techniques with conventional ones called MEAD with respect to some evaluation techniques. Simulation results demonstrate that CPSL shows better performance for short summarization than MEAD and for remaining cases it is almost similar to MEAD. Furthermore, simulation results demonstrate that LESM also shows better performance for short summarization than MEAD but for remaining cases it does not show better performance than MEAD.