Sentiment Extraction from Naturalistic Audio of Agricultural Expert Opinions

Abstract

Opinions of Agri Experts play a vital role in planning and managing crop productivity. We use a variety of text mining techniques to search for the strong sentiment keywords in social media, fiction, non-fiction, movies, novels, comics, stock exchanges, etc. Among the most widely used NLP techniques is sentiment extraction. Today, farmers are facing some difficulties maintaining crop productivity. They need to consult with agricultural experts for maintaining productivity. However, it is difficult for farmers to keep in touch with experts in agriculture industry. In the Proposed Study, experts offered different opinions based on their experience. Using decoded ASR (Automatic Speech Recognition) and sentiment model, audio data was processed by a machine learning algorithm. Then it was stored as a context for training. As a result of analyzing the stored text, the input by the user was analyzed and relevant answers were classified as output. Afterwards, it was translated into a national language like Urdu in Pakistan. This was done to make it more understandable to the farmer who lacks formal education, as opposed to textual data, which he may not understand. This solution became easy to understand and resulted in better outcomes before and after the harvesting of crops in different areas.