Litcius/Paper detail

A comparative study of machine learning models in predicting crop yield

Bobo Mafrebo Lionel, Richard Musabe, Omar Gatera, Célestin Twizere

2025Discover Agriculture10 citationsDOIOpen Access PDF

Abstract

This paper explores different machine learning techniques applied in the prediction of crop yield based on meteorological parameters. Recurrent Neural Networks (RCN), Convolutional Neural Networks, Random Forest, Decision Trees, Gradient Boosting machine (GBM) was explored. To rate the performance of machine learning models, evaluation metrics like Mean absolute error, Root mean squared error and coefficient of determination were used for decision making. The results shows that Random Forest shows high accuracy in predicting crop yield with R 2 of 0.875 for Irish potatoes and 0.817 for maize, however, for the prediction of cotton Extreme gradient boost had limited error of 0.07. Convolution neural network (CNN) was compared to traditional machine learning in the case of grading tomato and come up with a conclusion that combination of convolution neural network (CNN) with support vector machine (SVM) performed better with an accuracy of 97.54%

Topics & Concepts

Machine learningGradient boostingArtificial neural networkArtificial intelligenceRandom forestSupport vector machineMean squared errorBoosting (machine learning)MathematicsConvolutional neural networkComputer scienceCrop yieldDecision treeAlternating decision treeConvolution (computer science)Correlation coefficientYield (engineering)Mean squared prediction errorExtreme learning machineStatisticsMean absolute errorComputational learning theoryPattern recognition (psychology)Word error rateAlgorithmRegressionSmart Agriculture and AILeaf Properties and Growth MeasurementAgricultural Economics and Practices