Loading...
Please wait, while we are loading the content...
Similar Documents
Combing Paragraph Embedding and Density Peak Sentence Clustering based Multi-Document Summarization
| Content Provider | Semantic Scholar |
|---|---|
| Author | Wang, Baoyan Zou, Yuexian Zhang, Jia-Yu Yang Liu, Yan-Li |
| Copyright Year | 2017 |
| Abstract | We present a novel unsupervised extractive multi-document summarization (MDS) method by combing paragraph embedding and density peak-based sentence-level clustering. Word embedding is a widely used text representation method due to its remarkable performance. However, we are aware that paragraph embedding is relatively few used in MDS. Besides, both relevance and diversity should be properly considered when generating summary. Whereas most existing MDS methods tend to quantify the degree of relevance between sentences and the other firstly, while the diversity of summary is ensured through a post-processing module. Based on these observations, three contributions are proposed in this paper. First, we compare different text representation methods for MDS thoroughly, including three classical bag-of-word methods, two word embedding methods and two paragraph embedding methods. Next, we employ density peak clustering to cluster sentences and the integrated sentence scoring method to rank them, which take relevance, diversity and length constraint into account concurrently. Third, we evaluate our method on the benchmark datasets and compare it with the other state-of-the-art methods. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://web.pkusz.edu.cn/adsp/files/2015/10/Combing-Paragraph-Embedding-and-Density-Peak-Sentence-Clustering-based-Multi-Document-Summarization-2.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
