A framework for an intelligent agro-climate decision support system for small-scale farmers in Swayimane

Abstract

Machine Learning (ML) is fast becoming a technology of choice with enormous potential to transform small-scale agriculture, particularly in helping farmers make informed decisions about crop choices. On the other hand, there is evidence that small-scale farmers face several challenges, such as lack of access to market information, poor soil quality, and inadequate farming techniques. ML technology can be used to provide real-time information on weather patterns, soil quality, and other factors that affect crop growth and yields. By providing this information, ML can help farmers choose the right crops to plant and optimize their yields. In this paper, the authors report on the use of AI to select the appropriate crop to plant, which resulted in crop choices that are 99% accurate. This was achieved by collecting climatic and edaphic data, and using different multi-class classification algorithms to train the dataset. The results of the different algorithms were compared and contrasted using different metrics to determine the best fit for the development a framework for prediction of crop suitability at pre-planting stage. The resulting framework utilizes ML as well as observed Indigenous Knowledge (IK) to synthesize edaphic and climatic factors to support decision management for small-scale farmers.