Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/321283
Title: | Word Sense Disambiguation For Punjabi Language Using Intelligent Techniques |
Researcher: | Walia, Himdweep |
Guide(s): | Rana, Ajay and Kansal, Vineet |
Keywords: | Computer Science Computer Science Artificial Intelligence Engineering and Technology |
University: | Amity University, Noida |
Completed Date: | 2020 |
Abstract: | Natural Language Processing is one of the major sub-domains under Artificial Intelligence. It forms the basis of the technique that allows a machine to communicate in a manner similar to humans. This implies that the machine is capable of understanding the context in which a discussion is going on and is able to give an intelligent response to it. The different algorithms under Machine Learning have been instrumental in defining a framework that helps in this process.The algorithms for word sense disambiguation can be divided into supervised, unsupervised, semi-supervised and knowledge-based methods. Supervised systems need to be trained with sense-tagged corpus, learn the relationship between the specific sense and the context, and get a classifier for each word. Unsupervised approach utilizes clustering technique to cluster words based on their context to distinguish senses. Semi-supervised systems adopt bootstrapping methods which learn knowledge from a small sense-tagged corpus and extend their knowledge from a small sense-tagged corpus and extend their knowledge with the existing knowledge. Knowledge based methods mainly utilize external knowledge base, such as dictionary and ontology to choose the most appropriate sense. The supervised approach has shown good results in deciphering the context of the ambiguous word.On experimentation, we found that the results were moderate as compared to supervised classifier which indicated that the results indicated a positive swing towards determining the right context to be looked for the given ambiguous word.The experimental outcomes have showcased that the results are above moderate and can be improved further by increasing the stored cases in the case repository. The work has been done in Punjabi language which is one of the regional languages of India.Thus the hybrid methodology of combing the un-supervised technique with case-based reasoning would prove to be beneficial for deciphering the many contexts of the Punjabi ambiguous word. |
Pagination: | |
URI: | http://hdl.handle.net/10603/321283 |
Appears in Departments: | Amity Institute of Information Technology |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
01_title.pdf | Attached File | 92.89 kB | Adobe PDF | View/Open |
02_certificate.pdf | 215.96 kB | Adobe PDF | View/Open | |
03_preliminary pages.pdf | 83.97 kB | Adobe PDF | View/Open | |
04_chapter 1.pdf | 367.51 kB | Adobe PDF | View/Open | |
05_chapter 2.pdf | 309.11 kB | Adobe PDF | View/Open | |
06_chapter 3.pdf | 601.99 kB | Adobe PDF | View/Open | |
07_chapter 4.pdf | 636.5 kB | Adobe PDF | View/Open | |
08_chapter 5.pdf | 456.3 kB | Adobe PDF | View/Open | |
09_chapter 6.pdf | 634.18 kB | Adobe PDF | View/Open | |
10_chapter 7.pdf | 169.83 kB | Adobe PDF | View/Open | |
11_content.pdf | 201.74 kB | Adobe PDF | View/Open | |
12_bibliography.pdf | 270.21 kB | Adobe PDF | View/Open | |
80_recommendation.pdf | 258.68 kB | Adobe PDF | View/Open |
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: