Agriculture is essential to global food security and sustainable development. The use of machine learning (ML) and Internet of Things (IoT) has the potential to enhance crop production and improve resource management through enhanced precision agriculture. This article describes a smart crop recommender system developed to support both ML-based decision making and real-time environmental sensing. At the core of the system is a capacitive soil moisture sensor and a temperature-humidity sensor interfaced using an Arduino Nano. Using a nRF24L01 transceiver employing the 2.4 GHz Enhanced Shock Burst (ESB) protocol, the sensor information can be wirelessly transmitted. Data collected from the sensors are sent to an ESP32 module that posts the information to the Blynk web application through Wi-Fi, allowing real-time remote access. As the data source for the crop recommender model, the Blynk application provides the data needed to perform data analysis on the various ML algorithms (Random Forest, Bagging, Decision Tree, Gradient Boosting, and Enhanced Gaussian Naive Bayes EGNB) to determine the most suitable type of crop based on the current soil and environmental conditions. Based on the analysis, the EGNB model showed superiority with an accuracy of 99.55% with respect to precision, recall, and F1-score. This system provides an economically viable and expandable approach to smart agricultural practices, which integrates cutting-edge machine-learning modelling with digital tools for obtaining data in real-time (IoT). These capabilities allow users to receive and utilize actionable, data-driven guidance in real time when making decisions about crop production.
Sawant et al. (Fri,) studied this question.