Machine Learning Approach for Text Summarization

With the abundance of interminable text documents, providing summaries can help in retrieval of relevant information very quickly. The technique is to extract those sentences from the document that contain important information. This paper presents the results of our research on extractive summarization with a method based on Support Vector Machines (SVMs). The SVMs are trained using DUC-2002 dataset and the importance of sentences is judged on the basis of salient features. To evaluate the performance of our system, comparisons are conducted with two existing methods. ROUGE scores are used to compare the system generated summaries with the human generated summaries, and the experimental results show that our system's performance achieved high metrics.