Extracting and Representing Higher Order Predicate Relations between Concepts
| Autor: | Parnab Kumar Chanda, Bhavi Bhagwan Jagwani, Vibhor Mittal, Nitish Varshney, S. Chatterji |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
060201 languages & linguistics
Evaluation strategy Theoretical computer science Content retrieval Knowledge representation and reasoning Binary relation business.industry Computer science 06 humanities and the arts 02 engineering and technology Predicate (mathematical logic) computer.software_genre Semantic network 0602 languages and literature 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Artificial intelligence business computer Natural language processing |
| Zdroj: | Natural Language Processing and Information Systems ISBN: 9783319417530 NLDB |
| DOI: | 10.1007/978-3-319-41754-7_15 |
| Popis: | In a text, two concepts can hold either direct or higher order relationship where function of some concepts is considered as another concept. Essentially, we require a mechanism to capture complex associations between concepts. Keeping this in view, we propose a knowledge representation scheme which is flexible enough to capture any order of associations between concepts in factual as well as non-factual sentences. We utilize a five-tuple representation scheme to capture associations between concepts and based on our evaluation strategy we found that by this we are able to represent 90.7 % of the concept associations correctly. This is superior to existing pattern based methods. A use case in the domain of content retrieval has also been evaluated which has shown to retrieve more accurate content using our knowledge representation scheme thereby proving the effectiveness of our approach. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|