A Novel Intelligence System for Hybrid Crop Suitable Landform Prediction Using Machine Learning Techniques and IoT

In a bid to increase the agricultural need of yielding a healthy and profitable production, crop yield forecasting employing machine learning and the Internet of Things combines powerful data analytics with sensor technology. IoT devices play an interactive role in the field of crop production as they periodically check the growth and intimate the conditions through signals. Integrating machine learning with the devices improves the performance and gives improved results with the prediction model output. The proposed system deploys an N (nitrogen), P (phosphorus), and K (potassium) NPK sensor that senses the nutrients in the soil and uses algorithms to process the inputted metrics and predict the output based on the sources. The combined technology of predicting the soil quality and the yield rate provides an improved and effective system for the process. Machine learning algorithms used in the model are logistic regression, K-nearest neighbour (KNN), and XGBoost. The Novel innovation development of the machine learning model, the dataset of the soil with nutrients is collected with which the quality of the soil, and the crop yield prediction rate can be determined. The data from the sensor is stored in the cloud module which feeds the necessary data whenever needed. The proposed work compares the three algorithms and results in the algorithm with higher accuracy. The accuracy of the algorithms used are 95%, 97%, and 99%, respectively, and by comparing the accuracy results the XGBoost algorithm is more efficient than the other machine learning algorithms employed.