Soil Classification based on Machine Learning for Crop Suggestion
Keywords:Soil classification, Machine learning, Random Forest, Naïve Bayes, k-Nearest Neighbor (k-NN)
A system for classifying and arranging information about soil is known as soil classification. This category of soil was formed in response to a need for a simple, consistent, and easy-to-understand way to classify lands, which is especially important for plantation and agricultural decision-making. However, the current method of assessing soil type is time consuming and heavily relied on agricultural experts. The implementation of machine learning is expected for better soil classification to suggest the crop. The three algorithms are tested, which is Random Forest, Naïve Bayes, and k-Nearest Neighbor (k-NN). Classification techniques are being chosen as a data mining task to produce a classify model. Random Forest has the best accuracy (97.23 percent), Naïve Bayes has the second highest accuracy (96.82 percent), and k-Nearest Neighbor (k-NN) has the lowest accuracy (92.92 percent).