SIMILARITY MEASUREMENTS OF CANDIDATE’S ANSWERS TO MARATHI QA SYSTEM
Question Answering is becoming into a significant topic of research for researchers as there is an increasingly high level of technological advancement. Data is available in large amounts on the internet. People have become interested in searching for their needs on Internet, as it is the simplest way of gaining information with a click. QA systems are essential in education. Which has been provided during the Covid-19 pandemic as online education becomes the generic way of teaching-learning. There are two types of question-answering systems: closed-domain and open-domain. The best approach to get answers to user questions which are asked in natural language as compared to a query is through question answering. Popular languages spoken in India include Hindi, Punjabi, Telugu, Bengali, Malayalam, and others. These languages are now being researched extensively and taken into consideration by researchers. Currently, these languages are taken into consideration by the researchers and a lot of work is being done on these languages. During our study, Comparing the Marathi language to other Indian languages, we found that relatively little study has been done in the Indian context. In this study, a method for evaluating the system performance is presented. It is intended to examine how closely the candidate's paraphrased answers to the model answer match up with our model answers.
Question, Answer, Question Answering System, NLP