| Autor: |
Rana Hashmy, Mudasir Mohd, Iram Khurshid Bhat |
| Rok vydání: |
2017 |
| Předmět: |
|
| Zdroj: |
Advances in Intelligent Systems and Computing ISBN: 9789811056864 |
| DOI: |
10.1007/978-981-10-5687-1_56 |
| Popis: |
Summarization task helps us to represent significant portion of the original text in concise manner, while preserving its information content and overall meaning. Summarization approach can either be abstractive or be extractive. Our system is concerned with the hybrid of both the approaches. Our approach uses semantic and statistical features to generate the extractive summary. We have used emotion described by text as semantic feature. Emotions play an important part in describing the emotional affinity of the user and sentences that have implicit emotional content in them are thus important to the writer and thus should be part of the summary. The generated extractive summary is then fed to the Novel language generator which is a combination of WordNet, Lesk algorithm and part-of-speech tagger to transform extractive summary into abstractive summary, resulting in a hybrid summarizer. We evaluated our summarizer using DUC 2007 data set and achieved significant results compared to the MS Word. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
