A Machine Learning-Driven Crop Recommendation System with IoT Integration

Bangladesh's economy depends heavily on agriculture, yet farmers confront a significant obstacle in choosing the best crop which affects production and regrettably causes financial hardships, migration, and suicides. Because climate and ingredients of soil are changing continuously. This paper proposes a system that utilizes several types of soil and environmental characteristics to determine the ideal crop for a particular land. Via Internet of Things (IoT) devices, environmental characteristics that include temperature and humidity, as well as soil parameter that is pH will be immediately retrieved from the land, enabling instantaneous data gathering. Using a variety of algorithms, the suggested approach Gaussian Naive Bayes which we got 99.55% validation accuracy, determines which crop would be best for cultivation. Following the integration of the model into an intuitive interface, farmers are provided with a useful tool to improve decision-making and eventually support Bangladeshi agriculture's sustainability. The most efficient and accurate method is selected to build a machine-learning model after undergoing extensive testing on several different algorithms. In order to provide farmers with exact recommendations for choosing the most suited crop for cultivation which is integrated into an easy-to-use interface.