Hybrid Approach for Automated Answer Scoring Using Semantic Analysis in Long Hindi Text

Abstract

In the case of the Hindi language, the technology that underpins automated scoring is still in its infancy in terms of its development. These systems have shown much better accuracy and reliability in their operations Nowadays, several studies are being carried out in addition to saving individuals money and time. with the intention of providing nuanced feedback on grammatical as well as semantic problems. This paper's main objective is to develop a hybrid methodology for automated answer scoring using semantic analysis for long Hindi text. Deep Learning and Recurrent Neural Network method have been taken into consideration throughout this research study. The ability of recurrent neural networks, to learn the temporal dependency of sequential input gives them an edge over feed forward neural networks, when it comes to the scoring of musical responses. Research work has integrated PSO and Roberta to improve accuracy. Based on the research findings, the recommended approach has been shown to outperform the currently recognized revolutionary techniques. It shows that the Hybrid PSO-Roberta based deep learning strategy performs better than the old system in terms of precision, recall, and f1 score. It reduce the amount of paperwork they need to do, teachers won't have to be concerned about any evaluation issues going away either.