PERINGKASAN DOKUMEN BAHASA INDONESIA PADA CERPEN MENGGUNAKAN METODE MAXIMUM MARGINAL RELEVANCE
Daftar Isi:
- Reading activity is an activity that reqiures a relatively long time, a long time is very influential on a person’s productivity. Therefore, summarizing text document automatically is needed to help get the short story core. The data used in this study are documents that have been formatted in the form of TXT. Data is obtained directly form the website www.cerpenmu.com. The amount of data collected is 20 short stories. In this study there are four stages, namely first, text preprocessing such as sentence solving folding case, tokenizing, stopword, and stemming. Second calculate the weight values in summary sentences using various features. The features contained are words that resemble titles, frequency of occurrence of words in sentences, and also the lenght of sentences. Third, determine the results of summary sentences obtained form sentence extraction using the maximum marginal relevance method. Fourth, sequencing the results of summary sentences so as to produce a summary that is easy to understand for the readers. The level of accuracy obtained by testing the summary results of the shoftware compared to the summary results form expert experts so as to produce average precision 76.7%, recall 77.5%, f-measure 77%.