Crop Decision Using Various Machine Learning Classification Algorithms

Abstract

Agriculture is the backbone of a country's economic development. Crop decision is the most fundamental decision for every farmer. Digital transformation in agriculture has enabled many farmers in the country to make the right decisions of crops according to their location conditions. Modern techniques like machine learning can be used for this purpose. There are many algorithms involved in this technique. A comparison of various classification algorithms based on accuracy parameters is presented in this paper to determine an appropriate crop depending on field conditions. Machine learning algorithms like Naive Bayes, decision tree, logistic regression, K-nearest neighbors (K-NN), support vector machine, and random forest are used for this process. The analysis is performed using the WEKA software. The open-source dataset consisted of location parameters such as nitrogen, phosphorus, potassium, temperature, humidity, Ph, and rainfall along with labels of 22 crops. By this comparison, it has been concluded that both random forest and Naive Bayes are good algorithms for crop decisions based on accuracy parameters. Many parameters such as root mean squared error, precision, recall, TP rate and FP rate, and F-measures along with accuracy. This analysis can be used for proper agri-inputs for farmers and can be also used in many agricultural applications comprising location parameter measurement for weather prediction, etc.