A Trigraph Based Centrality Approach Towards Text Summarization
| Autor: | M. Geetha, Devika Raj |
|---|---|
| Rok vydání: | 2018 |
| Předmět: |
Context model
Computer science business.industry Graph based 02 engineering and technology computer.software_genre Graph model Automatic summarization 020204 information systems Trigraph 0202 electrical engineering electronic engineering information engineering Graph (abstract data type) 020201 artificial intelligence & image processing Trigram Artificial intelligence Centrality business computer Natural language processing |
| Zdroj: | 2018 International Conference on Communication and Signal Processing (ICCSP). |
| DOI: | 10.1109/iccsp.2018.8524528 |
| Popis: | As the electronic documents are increasing due to the revolution of information there is an urgent need for summarizing the text documents. From the previous works we observed that there is no generalized graph model for text summarization and low order ngrams could not preserve the contextual meaning. This paper focuses on an extractive based graphical approach for text summarization, based on trigrams and graph based centrality measure. Trigraph is generated and the centrality of the connected trigraph is taken to extract the important trigrams. A mapping is done between the original words and the trigrams to regain the link between the words. And after comparing the centrality from the graph, the summary is extracted. The ROUGE-SU4 F-measure obtained for the proposed approach is 0.036 which is significantly better than the previous approaches. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|