Abstract:
Automatic text summarization is a process of rewriting text into a shorter
compressed version from the original text. Extraction focuses on the
selection of particular pieces of text from a document where the sentences
and/or phrases with the highest score are considered as salient sentences and
are chosen to form the summary. The selection of the informative sentence is
a challenge for extraction based automatic text summarization researchers.
This research applied an extraction based automatic single document text
summarization method help differentiate using the genetic algorithm (GA) to
find out the best feature weight score to difference between important and
non-important features. The Recall-Oriented Understanding for Gusting
Evaluation (ROUGE) toolkit was used for measuring the performance. DUC
2002 data sets provided by the Document Understanding Conference 2002
were used in the evaluation process. The summary that generated by GA
algorithm were compared with other evolutionary algorithm (PSO,ACO) and
used DE algorithm as benchmark. Experimental results showed that the
summaries produced by the DE algorithm are better than other algorithms.
In the meantime, recently propped algorithms such as (ACO) could out
performance GA.
