IoT-Driven Enhanced Crop Recommendations with MachineLearning Algorithms

Abstract

The Indian economy is heavily reliant on the agricultural sector. The farming industry has the potential to proliferate in the coming decades. 'The globe has to produce 70% more food in 2050 than 2006 in order to feed the expanding population of the Earth,' states the UN Food and Agriculture Organisation. Food consumption is increasing at a high pace parallel to the rise of the global population. Changes to farming practices are required to bring supply and demand into balance. Identification of suitable crops based on soil properties is a major problem that farmers face. Irregular and traditional soil testing methods further contribute to huge losses in the agriculture sector. Internet of Things (IoT) along with machine learning methodologies, will become one of the technological advances to solve this problem. This article proposes a new method “Enhanced Crop Recommendation System (ECRS)” for effective crop recommendation to the farmers based on soil properties using the Internet of Things and Machine learning technical advancements. In the first stage, The Internet of Things (IoT) with a variety of sensors and actuators has the potential to revolutionize contemporary farming. It provides insights and detailed data that help farmers figure out the optimal conditions for crops to grow. This way the farmers won’t waste soil and fertilizers, increase the quality and quantity of the crop, water conservation, remote monitoring, and contribute to the overall economic growth. In the second stage,different machine learning algorithms such as Random Forest, XGBoost, Gradient Boosted, Trees, Logistic Regression, Light GBM, Decision Tree, K Nearest Neighbors, and SVM are applied to give crop recommendations based on a dataset from the Kaggle database. This system helps the farmers choose ideal and appropriate crop varieties according to soil characteristics, and weather conditions with the help ofthe Enhanced Crop Recommendation System using LightGBM algorithm.  Additionally, it also provides suitable crop recommendations via a user-friendly Mobile Application ‘Crop Recommender’ in the regional language (Telugu).Sensor data will be validated with soil testing laboratory values to know the accuracy of the IoT system.