ABSTRACT
Soil is an important component to consider when forecasting crop yield. Soil analysis can benefit farmers and landowners. Analysts believe that the arrangements are allowing for higher agricultural output. This study uses decision tree (DT) regression and classification, random forest (RF) regression and classification to forecast soil fertility. The experimental study used soil data collected from various district soil samples office in Maharashtra state. For evaluating the decision classification and RF techniques, such as the area under the receiver-operating characteristic curve (AUC) accuracy, Out of Bag (OOB) error rate (ER), recall, F-score, ER, and precision, and regression techniques, such as the mean squared error, the R2 score, and the ER, only a few dimensions have been evaluated to determine the efficiency of each studied technique. The most well-known study was on precise soil prediction algorithms that were trained using a machine learning approach, and the results were published in peer-reviewed journals consequently. Testing and regression techniques, as well as classification procedures, are presented. The results of the experiments showed that the RF tree and the DT are both more effective decision-making methods. RFs and DTs are the most accurate methods for predicting soil fertility.
