Real Time Crop Prediction based on Soil Analysis using Internet of Things and Machine Learning

Research in the realm of the agriculture sector is expanding. More than half of India's population depends on agriculture for livelihood, and it is a major contributor to the country's economic growth. Soil quality is changing drastically, affecting the agricultural crop yield. Machine learning and deep learning algorithms are effectively helping to predict the crop based on the soil quality of the land. Data on temperature, humidity, rainfall, soil moisture, and pH are needed to train the machine-learning models. This work has been carried out using the following machine learning models: Decision Tree classifier, K-Neighbor classifier, and Random Forest classifier models. The accuracy of the Random Forest classifier is 93.11 percent, which is higher than the accuracy of the Decision Tree classifier (90.96 percent) and the accuracy of the K-Neighbors classifier (87.63 percent). Along with accuracy, the following performance metrics, such as precision, F1 score, recall, mean absolute error, and log loss, are taken into account. Web-based software has been developed to forecast the crop prediction of farmland based on soil conditions. The real-time data on the soil quality is gathered using the IoT devices from the farm, and the data is saved in the cloud. The data is fed to the machine learning model to predict the crop that would be most suited for cultivation on the farm. Since this is a real-time strategy, farmers can predict the crop with greater accuracy, resulting in higher yields.