An Overview of Relevant Literature on Different Approaches to Word Sense Disambiguation
C. Pavithra, Supriya Mandal · 2021 Fourth International Conference on Electrical, Computer and Communication Technologies (ICECCT) · 2021
WSD (Word Sense Disambiguation) is a common issue in Natural Language Processing (NLP) and Machine Learning technology. In NLP, word sense disambiguation is described as the capacity to detect which meaning of a word is activated by its use in a specific context. WSD is a solution to the uncertainty that occurs when words have different meanings in different contexts. Contextual word meaning plays an important role in various applications such as sentiment analysis, search engine, information extraction, machine translation etc. It is a challenge for these systems to detect and overcome the uncertainty that emerges from the lexical ambiguity. Many studies have been conducted over the decades to propose various approaches to the WSD problem. In this manuscript, a comparative study of three approaches namely LESK algorithm, embedding techniques, and Neural Network techniques based on the text collected from children's story books is performed. We explored an approach that combines Bi-LSTM neural network with Knowledge Graph to predict contextual word meaning. Our study shows that the combined approach accuracy is 80.34approaches