Support Vector Machines and Intelligent Irrigation Systems: A Scoping Review

Recent global climate change, global warming, and global water scarcity have turned out to be critical world issues. To tackle these vital challenges, effective irrigation systems management is very important to ensure proper water usage without risking crop yield and improving production. This scoping review investigates the applications for Support Vector Machines (SVM) to enhance intelligent irrigation systems for sustainable agriculture. A systematic search was conducted using four major databases such as Web of Science, Google Scholar, Scopus, and ScienceDirect; after that, 48 relevant research studies were selected out of 491 research studies pointed out from 2020 to 2024. It is demonstrated by our findings that the SVM model performs best (19 research articles) as compared to other state-of-art machine learning models (such as DT, RF, and ANNs) while classifying irrigation requirements and predicting different soil moisture levels based on IoT devices and sensor data such as humidity, temperature, and soil moisture content. Furthermore, SVM determines strong performance but its computational cost and time needs increase as the dataset expands. Moreover, its performance is vulnerable to the selection of parameters and requires substantial tuning through methods such as grid search or cross-validation. Integrating nature-inspired or bio-inspired algorithms can be further enhanced. These relevant enhancements would lead to more accurate and efficient irrigation predictions to support the development of advanced and sustainable irrigation practices.