IMPLEMENTASI MAXIMUM MARGINAL RELEVANCE DAN MATRIKS COSINE SIMILARITY PADA APLIKASI PERINGKASAN DOKUMEN
Automatic Text Summarization are sorting paragraphs into shorter forms using a computer operated applications. This automatic summarization technique works in a computer to summarize the text inputted by user. A document entered into the computer application, then processed and summarized to produce a summary from original text. In research document extraction method used Maximum Marginal Relevance algorithm. MMR is a summary document extraction method is used to summarize a single document or multiple documents. MMR summarizing document by calculating the similarity between the sentences text in a paragraph. In this summarization contents for document segmentation process is carried out using a combination of gender-based matrix cosine similarity. Cosine similarity is used to calculate the relevance of the query approach on a document. The determination of relevance in a query against document is considered as measurement of similarity between queries with vectors of documents. The results from the automatic summary application is a sequential list words according to the results obtained algorithms. The closeness results with query assessed from the lambda limit values applied in the cosine similarity is 0 to 1.
Dibuat oleh : Jasman Pardede, Jordy Sinatria
E-mail : jasman@itenas.ac.id
Kata kunci :summarization, MMR, queries, similarity, cosine similarity.
Keterangan : Karya Ilmiah ini dimuat pada Prosiding Konferensi Nasional Sistem Informasi KNSI 2015, 26 – 28 Februari 2015