ANALISA PERBANDINGAN ALGORITMA MAXIMUM MARGINAL RELEVANCE DENGAN TEXTRANK DAN LEXRANK PADA KASUS PERINGKASAN TEKS OTOMATIS
Daftar Isi:
- Automatic text summarizing is done with the aim of summarizing shorter texts without losing their meaning and content. In this study, a comparative analysis of the results of text summarizing is done using the Maximum Marginal Relevance algorithm with Textrank and Lexrank. The processes carried out in this research are Pre-Processing, TF-IDF Weighting, Application of MMR Algorithm with Textrank and Lexrank, and calculating the results of concise accuracy. In the pre-processing, input processing is case folding, tokenizing, and stemming. Then summarize using the MMR algorithm with Textrank and Lexrank, then evaluate the level of accuracy with manual testing using the parameters recall, precision, and f-measure. The data tested in the form of Indonesian text documents amounted to 10 texts. The results showed the MMR method had a recall accuracy level of 19%, precision 37%, and f-measure 25%; and the Textrank and Lexrank methods have 51% recall accuracy, 69% precision, and 59% f-measure.