Food is the moral right of all human beings but we are living in a world where people are facing acute food insecurity. To grow more, we need to address the issues and challenges faced by our farmers. In India, to address the queries of farmers, the government has established Kisan Call Center (KCC). We consider the queries of farmers from Odisha which is an eastern state of India. From our analysis, it is found that most of the farmers grow paddy, which is consumed in most part of India and abroad. We developed a machine learning model that will classify the query based on whether the query is related to paddy farming or not. If the query is related to paddy crop then we try to answer the queries based on the past records in the dataset. We applied Long Short Term Memory (LSTM) and Support Vector Machine (SVM) with some pre-trained embedding model like Glove and sentence transformer. We achieved classification model accuracy 98.66% and 98.47% with SVM and LSTM respectively.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Machine Learning Approach to Classify and Answer Paddy-Related Queries Using KCC Dataset

  • Mahaprasad Mohanty,
  • Deepak Sahoo

摘要

Food is the moral right of all human beings but we are living in a world where people are facing acute food insecurity. To grow more, we need to address the issues and challenges faced by our farmers. In India, to address the queries of farmers, the government has established Kisan Call Center (KCC). We consider the queries of farmers from Odisha which is an eastern state of India. From our analysis, it is found that most of the farmers grow paddy, which is consumed in most part of India and abroad. We developed a machine learning model that will classify the query based on whether the query is related to paddy farming or not. If the query is related to paddy crop then we try to answer the queries based on the past records in the dataset. We applied Long Short Term Memory (LSTM) and Support Vector Machine (SVM) with some pre-trained embedding model like Glove and sentence transformer. We achieved classification model accuracy 98.66% and 98.47% with SVM and LSTM respectively.