DigiFarm – A Machine Learning based holistic Crop Prediction Platform for Farmers

Abstract

The world population is increasing everyday and to keep up with it an extensive food production system has to be designed. Farming is one of the most important occupations in India. Traditional farming is used by most of the farmers these days, however it has its own disadvantages like less yield leading to less income. It makes use of pesticides which degrades the quality of the soil and the crops. To overcome these disadvantages farmers nowadays are shifting to digital farming which yields more crops thus increasing their income. This research work intends to delineate the details of the newly built platform “DigiFarm” which aids the farmers in making informed decisions based on their soil pattern and weather condition. The central aim of DigiFarm is to help the farmers in predicting the crop which is most suitable for their land to gain maximum yield. To get accurate crop prediction, in this research work two Machine Learning algorithms have been compared. They are Random Forest(RF) and Gradient Boosting (GB) algorithms. The RF algorithm outperformed with better accuracy compared to the GB algorithm. Hence, RF algorithm has been used in the platform for crop prediction which helps the farmers in predicting the best crop for their land based on various factors like soil composition, weather, etc., The platform also provides an Artificial Intelligence based chatbot named ‘AgriBot’ which helps the farmers in guiding them through the platform and also helps them in predicting the crops as well as answering their other queries.