Enhancing Precision Agriculture with Machine Learning & IoT: SVR and DTR Comparative Analysis

In many nations, including Bangladesh, the importance of agricultural revenue and employment has dwindled over time. A sizable section of the population works directly in the agricultural industry, typically depending on past crop cultivating expertise. Because of this lack of accurate agricultural information, crop selection may be poor. Furthermore, agricultural institutions rely heavily on manual data, which is insufficient for accurate crop selection and output forecasts. Precision farming remains a distant dream. Using multiple regression methods, this research analyzes the most efficient way for agricultural production prediction. There has been little study into using machine learning methods integrating with IoT to estimate agricultural production. This research also incorporates multiple regression techniques on the obtained dataset using machine learning algorithms. The fundamental goal of this study is to improve the predictability of optimum crop selections, mitigate food issues, and encourage intelligent information prediction analysis in agriculture. This strategy is intended to help farmers make better informed and profitable agricultural choices. Data from Bangladesh Bureau of Statistics & Bangladesh Meteorological Department are used, which spans 43 years. This study takes into account one important crop—Aman rice. Furthermore, this research takes into account seven environmental characteristics, including monthly averages of Bangladesh’s hottest and lowest temperatures, yearly rainfall, cloud cover, wind speed, relative humidity, and strong sunlight and proposed an IoT model for data collection. The characteristics of optimum control for Bangladesh’s primary crop agroecosystems are used. The Decision Tree Regression technique is proposed in this study for agricultural crop selection & yield prediction.