Trainable Automatic Text Summarization Using Segmentation of Sentence

In this paper, we propose an automatic summarization method combining conventional sentence extraction and trainable classifier based on Support Vector Machine. To make extraction unit smaller than the original sentence extraction, we also introduce sentence segmentation process in our method. The evaluation results show that our system achieves the best result among all the other systems with regard to contents, and closer to the human constructed summaries (upper bound) at 20% summary rate. On the other hand, the system needs to improve readability of its summary output.