PREDICTIVE ANALYTICS IN AGRICULTURE: MACHINE LEARNING MODELS FOR CROP YIELD

Machine learning applications are transforming the way data is processed and decisions are made, which is having a big effect on the global economy. Agriculture is one of the industries that is significantly impacted by the global food supply issue. This research aims to find out what modern farmers may gain by employing machine learning techniques. Optimising agricultural productivity while reducing waste is primary goal of these algorithms, which use intelligent judgements for planting, watering, and harvesting. Study delves into present situation of machine learning in farming, drawing attention to important obstacles and possibilities. It provides experimental evidence that changes to labels affect the precision of data processing algorithms. By analysing the massive amounts of data acquired from farms and utilising online, real-time data from IoT sensors, results suggest that farmers may make better decisions on variables influencing crop growth. Eventually, incorporating these Technology that increases agricultural yields while decreasing waste has the potential to change modern agriculture. Among the fifteen different algorithms, one newly developed algorithm that improves feature combining schemes is provided in order to help determine which method is most appropriate for usage in agriculture. According to the results, 99.46% classification accuracy can be achieved with the Hoeffding Tree and Naïve Bayes Classifier methods, and 99.59% classification accuracy can be achieved with the Bayes Net methodology. These results will demonstrate higher output rates. Keywords: Random Forest, Logistic Regression, Naïve Bayes, ANN, Smart farming, Crop analysis and prediction