<![CDATA[Increasing the amount of digital data quickly each year, one of which is text data of documents that can be text news articles can make it difficult for the reader to understand all the information so as to affect the acquisition of accurate information and need a longer time to extract an information on a document. Therefore, it is necessary to have an automatic extract text extracting system in Indonesian language health articles in order to help the reader or user to facilitate the process of extraction of information in the document with a fast, concise and clear time. This study uses the latent semantic analysis (LSA) method which is a method that extracts semantic structure or hidden meaning in a sentence and produces a general or broad meaningful summary. The LSA method uses the linear value decomposition (SVD) linear algebra approach by forming a representation matrix of term associations which are words in the closely related document of the TF-IDF calculation process. The LSA cross method is used to construct a summary sequence in the summary extraction stage. Tests of this research resulted that the result of text summary with LSA method obtained the accuracy value of precision, recall and f-measure in consecutive order at compression rate 50% with value 0.668, 0.743, 0.700 and 0.690 while at compression rate 40% equal to 0.696, 0.605 , 0.642 and 0.663.]]>
Copyrights © 2018
