Soil nutrient status prediction using machine learning algorithms

Soil nutrients play a pivotal role in facilitating optimal plant development and enhancing agricultural yield. The precise assessment of soil nutrient levels is paramount for making informed agricultural choices, encompassing crop selection, land preparation, and fertilizer application. This study incorporates diverse supervised machine learning approaches, including K-Nearest Neighbour, Decision Tree, Random Forest, Support Vector Machine and Naive Bayes, to analyse soil nutrient profiles. A total of 12 soil parameters were employed to categorize soil nutrients into low, medium, and high classifications. Post pre-processing, the dataset underwent a division into training and testing datasets. Algorithms were applied to the training collection and then assessed with the test dataset, using Python for coding. The random forest model achieved the highest accuracy, reaching 99%, thus surpassing alternative methodologies. The research highlights that the application of machine learning strategies, notably the random forest method, can greatly advance the accuracy of soil nutrient forecasts, allowing farmers to make wiser decisions that increase productivity and optimize land management.