Multilingual Language Detection and Translation System with Multinomial Naive Bayes
Sowjanya Vuddanti, Dhathri Sri Lekha Ariveni, Niharika Sri Sandhya Kethepalli, Lokesh Manchimi, Jaswanth Yajjavarapu · 2024
Smooth communication becomes essential as the world unites different linguistic groupings. The proposed system for language recognition and translation addresses this difficulty. The proposed system uses a Google Translate-powered “bridge” with a machine learning “detective” to properly identify text language and translate between any supported combination. The Multinomial Naive Bayes classifier, a machine learning approach, is used by the language detection component to determine the language of input text. As the “detective,” the Multinomial Naive Bayes classifier correctly determines the language of the input text with a remarkable accuracy rate of $97 \%$. The model can accurately classify language in a range of languages by preprocessing the text data and generating a representation of words in a bag. To translate text between languages, the text translation component uses the Google Translate API via the Google library. With the flexibility to choose the source and target languages provided by this component, users can translate text from any supported language to the language of their choice. Easy to use, it enables people to promote intercultural understanding and overcome language barriers in social media, business, education, and other domains, making the world more inclusive and connected.