Predictive Modeling of Agricultural Land Performance Using Digital Twin and Artificial Intelligence Techniques (Published)
Agricultural land performance is a key factor in ensuring productivity, resource efficiency, and sustainable investment in modern farming. Traditional methods fail to account for complex interactions between soil properties and climatic variability, leading to suboptimal decisions. In order to facilitate data-driven decision-making in agriculture, the goal is to provide a precise forecasting model for agricultural land performance. This research develops a model for agricultural land performance by integrating Digital Twin technology with advanced Artificial Intelligence (AI) techniques. A digital twin was created using 10,000 data points to represent physical agricultural land, combining sensor data and remote sensing inputs that included soil moisture, weather parameters, and crop conditions. Normalization was used to maintain data uniformity. To improve efficiency and accuracy, Principal Component Analysis (PCA) was used for feature extraction to reduce dimensionality, remove redundant information, and identify the most important aspects driving land performance. The hybrid Magnetotactic Bacteria Optimized K-Nearest Neighbour-Tuned Tree Algorithm (MBO-KNN-TA) accurately predicts agricultural land performance using crop yield. MBOA optimizes parameters by exploring solution space; KNN identifies patterns in multi-dimensional feature data; and XGBoost predicts the agricultural land performance. Experimental results conducted in Python 3.10 indicate that the proposed model outperforms existing approaches, achieving 98.2% accuracy. The digital twin enables scenario simulation, risk-informed decision-making, and optimized resource allocation, making it a robust tool for precision agriculture. These findings demonstrate that integrating artificial intelligence (AI) with a digital twin framework can significantly enhance predictive accuracy, sustainability, and data-driven management of agricultural land.
Keywords: Decision Support System, Digital twin, agricultural land performance, machine learning, precision agriculture, predictive modeling