INTELLIGENT COACHING AGENT FOR ETHIOPIAN AGRICULTURE PRODUCTIVITY WITH MACHINE LEARNING

Abstract

Ethiopia is one of the countries where agriculture is a predominant occupation. The country's economy heavily relies on agriculture, particularly crop production. Technological advancements and big data progress have led to the development of more connected, accurate, and efficient precision farming instruments. Mechanization has gradually replaced manual labor in the agricultural sector, resulting in increased land productivity and economies of scale. This transition has enabled farmers to manage larger fields and farms more effectively. In Ethiopian agriculture, various factors such as land area, rainfall, temperature, humidity, fertilizer usage, sunshine, rainfall patterns, and soil type significantly influence agricultural outcomes. However, accurately estimating crop production remains a major challenge. The existing system for Ethiopian farming faces difficulties in detecting crops, identifying crop types, and predicting crop production. The primary objective of this study is to predict crop productivity by forecasting crop types. Additionally, the research involves the analysis and prediction of crop production. The dataset used for the study was compiled from diverse sources, including crop data from the agricultural office and meteorological data from Ethiopia's national meteorology agency. Data collection techniques, such as interviews and document analysis, were employed. The proposed work utilizes machine learning algorithms, specifically Artificial Neural Networks (ANN) and Support Vector Machines (SVM). Performance metrics, such as accuracy, are employed for crop type prediction and addressing crop production concerns. Based on experimental results conducted on agricultural data, the following outcomes were obtained: The SVM model achieved a crop prediction accuracy of 96.8%, while the ANN model achieved an accuracy of 90.69%. Consequently, the SVM model was determined to be the most suitable for crop type prediction and was utilized in developing an intelligent coaching agent system. In conclusion, the proposed system employs SVM for the development of an intelligent coaching agent that predicts crop types and offers guidance in Ethiopian agriculture.