Hybrid Method for Named Entity Recognition in Kumauni Language Using Machine Learning

Named Entity Recognition (NER) identified entities in a text according to specified rules. Machine interpretation, question answering, and automated summarization are among NER's numerous NLP applications. Language barriers make identifying people harder. They recognized words having various meanings or uses in other formulations. Writing similar words might be tough. Finding unlabeled terms is difficult. To mitigate these challenges in this research, a novel fusion chain model for NER in Kumauni Language using Machine Learning (ML) is proposed. A fusion of Convolution Neural Network (CNN) with Bidirectional Long- and Short-Term Memory (LSTM) and Conditional Random Field (CRF) is utilized as the proposed model for training and testing in which Support Vector Machine (SVM) is used for tagging. Finally, for measuring the performance of the proposed model, the performance metrics are calculated. The results show that the precision, recall, and F-measure of the proposed model with skip gam embedding are 76, 75, and 75.5%, and with GloVe embedding the precision, recall, and F-measure of the proposed model are 78, 62, and 69.08%, in both the cases, the proposed model performs better than other conventional approaches.