Automatic Document Summarization System Based on Natural Language Processing and Artificial Intelligent Techniques

摘要

Extract summary optimization is the process of creating a small version from the original text Satisfy user requirements. Extraction approach is one of way of extracting the most important sentences in document, t his approach is used to select sentences after calculating the score for each sentence, and based on user defined summary ratio the top n sentences are selected as summary. The selection of the informative sentence is a challenge for extraction based autom atic text summarization researchers. This research applied extraction based automatic single document text summarization method using the particle swarm optimization algorithm to find the best feature weight score to differentiate between important and non important feature. The Recall Oriented Understanding for Gusting Evaluation (F measure) toolkit was used for measuring performance. DUC 2007 data sets provided by the Document Understanding Conference 2007 were used in the evaluation process. The summary that generated by Particle Swarm Optimization algorithm was compared with other algorithms namely Latent Semantic Analysis, Gong&lui, and Vector Space Model, and used Particle Swarm Optimization algorithm as benchmark. Experimental results showed that the summaries produced by the Particle Swarm Optimization algorithm are better than another algorithm.