Leveraging Data Exchange and Random Forest Regression Model for Precision Farming: Predicting Soil pH to Enhance Agricultural Efficiency

Abstract

Precision farming is revolutionizing agriculture by incorporating advanced technologies and data-driven methods to improve crop  production. However, a gap persists in the practical use of machine learning (ML) models and real-time data exchange for optimizing soil  conditions to enhance crop yields, reduce resource waste, and limit environmental impact. This study aims to bridge this gap by  leveraging ML, particularly focusing on soil pH prediction to ensure optimal nutrient absorption for plant health. Data was collected from  Kigali Independent University ULK in Gasabo District, Kigali City, Rwanda, using seven soil sensors measuring soil moisture, temperature,  humidity, NPK levels, and pH. The study applied a Random Forest regression model to predict soil pH, achieving an impressive accuracy  of 99.9%, surpassing several contemporary models. The results highlight the effectiveness of ML in offering valuable insights to farmers,  promoting sustainable and profitable agricultural practices. The research emphasizes the importance of continuous technological  innovation and collaboration to push the boundaries of modern agriculture.