Satyam Dwivedi (IIT BHU)
Abstract ID: 411
Topic: Language Documentation
General Session Papers
Negation is a pretty complex phenomenon to handle when it comes to NLP applications. Especially in the case of morphologically rich languages, the flexible position and respectively the scope of negation poses a substantial NLP problem for scholars dealing with sentiment analysis and similar NLP applications, where the granularity of results depend on the insights from each and every linguistic device.
The idea and the products namely NEGFramator and NEGFrame repository are the original contributions of this work. For this paper, the scope of the work is delimited to handle negation in Indian languages. Initially, the repository is being created for Hindi and sister languages of Indo-Aryan descent and the framator is being used to cover the negation for sentiment analysis for testing purpose. Later on, this can be utilised for various NLP applications such as MT. Based on the ease of use and scalability, these resources can later be scaled for granular analysis in other languages as well.
This paper presents the idea and the architecture of NEGFramator and NEGFrame repository, various problems that came in the process of development, and their possible solution.
Keywords: NEGFramator, NEGFrame repository, NLP