Sindhi Named Entity Recognition (SNER)
Main Article Content
Abstract
Natural Language processing is one of the least advanced field of Artificial Intelligence (AI) due to the variety of languages and differences amongst the language rules. Named Entity Recognition (NER) is also a least advanced application of natural language processing and branch of information retrieval in which text is understood, classified, labeled and retrieved. A difference in language impose new challenges as some of the entity recognition systems for English perform at the level of human response. This paper introduces Sindhi Named Entity Recognition (SNER) as of its very first attempt for NER for this language. Sindhi is one of the seven languages of the history of the mankind. SNER is intelligently extract and classify entities from Sindhi text with pre-defined categories. Challenges and issues are analyzed and presented the solutions of these identified problems in Sindhi Named Recognition System. The system works on more than 200,000 words including Sindhi names, surnames, numbers, names of cities and other entities. The proposed system performs NER tasks successfully and presents 97% accuracy.Ambiguity handling is the next step along with the grammar understanding, which will lead to a Sindhi chat bot.
Article Details
COPYRIGHT
The Government-Research Journal of Political Science follows creative commons copyright policies, authors are instructed to send a copyright form attached with article while submission. Click here to download the copyright form.