Mining Agricultural Data to Predict Soil Fertility Using Ensemble Boosting Algorithm

Agriculture is the most important resource of livelihood and an emerging field the forms the backbone of India. Present challenges of the agriculture domain include uncertain climatic changes, poor irrigation facilities, weather uncertainty. Machine learning is one such technique that is employed to predict the fertility of the soil in agriculture. ensemble machine learning techniques aim to create meta-classifiers to produce better predictive performance. The primary focus of this paper is to analyze the soil data that is collected from the soil testing laboratory to predict fertility from a collected dataset by using multiple ensemble machine learning algorithms such as bagging, boosting, and stacking for better prediction, accuracy, and higher consistency. The soil fertility classes were evaluated using 10 selected attributes. Measurements of different soil parameters have been used for predicting soil fertility. The experimental result shows that the boosting method on the C5.0 algorithm achieved higher accuracy than other ensemble classifiers with 98.15%.