Pronunciation Accuracy Calculator using Machine Learning
Priyanka Nandal, Yajash Kadian, Saurabh Upadhyay, Bhanu Pratap Mudgal · 2021
Pronouncing words in an error-free and correct way is one of the most relevant skills to acquire. Scientists all over the world are concentrating on identifying pronunciation mistakes with the help of machine learning. This paper describes calculating pronunciation accuracy of a person using machine learning. Pronunciation dictionary data mainly from CMU library is employed to convert the inputted voice data into digital format (ARPABET) and analyzing that data using machine learning models such as LSTM based RNN and then applying HMM. Hence, a neural network model was built to predict pronunciation of different words. Pronunciation accuracy is calculated that can be used by anyone for self assessment of pronunciation accuracy. The results demonstrate acceptable accuracy of the model.