Performance Analysis Rice Yield Model based on Historical Weather Dataset in Bangladesh

The study’s aim is to evaluate the efficacy of the Rice Production Model in Bangladesh by applying the Historical Weather Dataset and a sustainable Machine Learning (ML) model that is compatible with Industry 4.0. In order to extract high-level information from enormous meteorological datasets, machine learning models are increasingly widely used for rice yield production projections. Throughout the course of this research, a number of different machine learning models, such as Radial basis function, Multiple Linear Regression, Support Vector Regression and Multilayer Perceptron were constructed in order to make predictions about the production of rice. Maximum and Minimum temperature, rainfall, and humidity are the climatic factors considered to develop and evaluate the frameworks. The model was developed using climatic and rice yield data from Bangladesh’s fifteen regions between 2006 and 2016. The result indicate that the Support Vector Machine Regression (SVR) exceeds other existing frameworks for reliably predicting future rice yields in Bangladesh leveraging the developed framework, as shown by the findings.