IoT-Based Crop Yield Prediction System in Indian Sub-continent Using Machine Learning Techniques

The agricultural sector holds significant importance in the socio-economic structure of India. The inability of farmers to make informed decisions regarding the optimal crop selection for their land, based on non-scientific and traditional methods, poses a significant challenge for a nation where farming constitutes the livelihood of nearly 58% of the population. At times, farmers have experienced challenges in selecting appropriate crops that align with the soil conditions, sowing season, and geographical location. By collecting data from sensors in the fields, ranchers can easily ascertain the extent of the crops, which helps preserve scarce resources and lessens the negative effects of weeds on the environment. The majority of Indian farmers choose conventional farming methods when deciding how their crops should be grown. Regardless, ranchers do not jeopardize agricultural yields, which are sensitive to abiotic factors such as soil type. In this proposed method, two types of datasets have been collected from the southern and northern parts of the country through IoT-based sensors. The crop recommendation system has been processed with various machine learning algorithms like logistic regression, KNN, decision tree, and support vector machine through the collected dataset. The accuracy of 97% and 96% has been achieved for processed datasets in the southern and northern parts of India. © 2023, The Author(s), under exclusive licence to Springer Nature Switzerland AG.