Abstract

Significant proportion of crops are damaged in farms by bacterial attack, erosion and lack of knowledge to follow right agricultural practices. This has negative effects on yields to be gotten from farms. Farmers prefer using previous farming experience to estimate numbers of yields, which is not a good practice. Future solutions to feeding massively expanding population involve digitalizing and automating agriculture. The use of big data and machine learning is a crucial instrument for the digitalization of the agriculture industry and other sectors. This study shows how machine learning contribute to digital agriculture in terms of crop yield prediction. Dataset were gotten from three (3) geopolitical zones of the country, modelled for the prediction using three (3) machine learning techniques using decision tree classifier, random forest and support vector so as to choose the best in terms accuracy after evaluating all the models using Root Mean Square Error as evaluation metric. Cross Industry Standard Processing for Data Mining (CRISP-DM) and Agile method were used and result shows North West yield better and accurate during dry season with accuracy of 98.257%