Most of the document summary are arranged extractive by taking important sentences from the document. Extractive based summarization often not consider the connection sentence. A good sentence ordering should aware about rhetorical relations such as cause-effect relation, topical relevancy and chronological sequence which exist between the sentences. Based on this problem, we propose a new method for sentence ordering in multi document summarization using cluster correlation and probability for English documents. Sentences of multi-documents are grouped based on similarity into clusters. Sentence extracted from each cluster to be a summary that will be listed based on cluster correlation and probability. User evaluation showed that the summary result of proposed method easier to understanding than the previous method. The result of ROUGE method also shows increase on sentence arrangement compared to previous method. Keywords: Document Summarization, Cluster Ordering, Cluster Correlation, Probability AbstrakSebagian besar ringkasan dokumen dihasilkan dari metode extractive, yaitu mengambil kalimat-kalimat penting dari dokumen. Ringkasan dengan metode extractive sering tidak mempertimbangkan hubungan antar kalimat. Pengurutan kalimat yang bagus menunjukkan hubungan rhetorical, seperti hubungan sebab akibat, topic yang relevan, dan urutan yang kronologis diantara kalimat. Berdasarkan permasalahan ini, diusulkan sebuah metode baru untuk pengurutan kalimat pada peringkasan dari beberapa documen menggunakan cluster correlation dan probabilityuntuk dokumen berbahasa inggris. Kalimat dari beberapa dokumen dikelompokkan berdasarkan kemiripannya ke dalam cluster-cluster. Kalimat diekstrak dari setiap cluster untuk menjadi ringkasan, ringkasan akan diurutkan berdasarkan cluster correlation dan probability. Hasil evaluasi pengguna menunjukkan hasil ringkasan dari metode usulan lebih mudah dipahami dari pada metode sebelumnya. Hasil ROUGE juga menunjukkan peningkatan susunan kalimat dari metode sebelumnya.Kata Kunci: Peringkasan Dokumen, Pengurutan Cluster, Cluster Correlation, Probability
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.