Explainable AI for Soil Fertility

Soil fertility is crucial to permaculture and depends on numerous chemical, physical an d biological factors that affect plant growth. As the demand for agricultural products continues to increase and arable land decreases, new solutions are essential to increase agricultural productivity without affecting environmental justice. To this end, this paper proposes a new method that uses artificial intelligence (XAI) to predict soil fertility with unprecedented accuracy and transparency. Our model uses random forest workers t o predict the relative soil fertility of a sample bas ed on its physicochemical properties. More importantly, the model provides participants with informed consent by clarifying the logic behind each pregnancy prediction from the user’s representative images. Our results showed an accuracy of up to 97.02%, exceeding traditional machine learning models. In addition to predictive power, our XAI model illuminates the interaction between soil parameters, revealing the underlying mechanisms of soil fertility control. The transition to a transparent model is consistent with the United Nations Sustainable Development Goals, particularly SDG 2 (Zero Hunger), SDG 13 (Climate Act ion) and SDG 15 (Life on Earth). Our approach not only makes agriculture more profitable by in creasing soil fertility, but also reduces climate change, promotes environmental protection and contributes to world food security. In addition, the versatility of our model extends to global application and offers practical solutions to ameliorate and longterm soil fertility degradation. Stakehol ders can use the predictive power of our XAI model to develop evidence based permaculture strategies through collaboration