IMPROVING CROP YIELD FORECASTING WITH AGRICULTURAL ENVIRONMENT FEATURES: FEATURE SELECTION AND CLASSIFIER-BASED APPROACHES

Abstract

The study of agriculture as an academic discipline is very recent. Soil and climatic variables, such as rainfall, temperature, and humidity, significantly impact crop yields, making crop yield prediction a crucial part of agriculture. Farmers used to have greater leeway in terms of when and what crops they could grow when I was a kid. The rate of global change is making it impossible for farmers to maintain their traditional methods. Since machine learning approaches can now estimate, this study used many sorts of these algorithms to predict the productivity of farms. If you want to be sure that a specific ML model is doing its job, you need to employ excellent feature selection methods to transform raw data into an ML dataset. It is critical to add only data points that are relevant to the model's output in order to maintain the accuracy of the machine learning model and prevent redundancy. Careful feature selection is required to ensure that the model contains just the most crucial attributes. An overly complex model would result from adding all raw data traits without first determining if they were beneficial for developing the model. The accuracy of the output would also be affected by adding qualities that simplify the ML model in terms of space and time. According to the research, the current classification system is not as effective as an ensemble strategy when it comes to producing predictions.