Rainfall Prediction and Cropping Pattern Recommendation Using Artificial Neural Network

Abstract

Ethiopia’s economy is primarily agricultural, with agriculture employing more than 85% of the country’s population. In a country like Ethiopia, where agriculture is the main source of income, reliable rainfall data is critical for water resource management, disaster avoidance, and agricultural productivity. In a circumstance where the amount of rainfall varies from time to time, cropping pattern recommendation is also highly important. In this paper we perform report, and discuss results of rainfall prediction and cropping pattern recommendation specifically for Amhara region using different combination of metrological parameters. Our Radial Basis Function Neural Network (RBFNN) prediction demonstrates better performance than the techniques used by Ethiopian National Metrological service agency (ENMSA) and other statistical techniques. For the recommendation system we used Model based collaborative filtering technique; that is K-means algorithm. We used Mean Squared Error (MSE), Mean Absolute Error (MAE), Root Mean Squared Error (RMSE) and Sum of Squared Errors (SSE) evaluation metrics to evaluate our prediction results. Generally, we can say that this is the first work which combines rainfall prediction and cropping pattern recommendation.