AHDAN-MLP: An Adaptive Hybrid Dynamic Adjusting Neural Classifier for Two Stage Soil Climate Based Crop Recommendation

Accurate crop recommendation requires effective modeling of complex interactions between soil nutrient composition and climatic variability. Conventional single-stage classification approaches often fail to adequately capture hierarchical Agro-environmental dependencies, leading to reduced robustness under diverse soil conditions. This study proposes a two-stage hierarchical soil–climate framework for crop recommendation. In the first stage, soil samples are clustered into five agronomically meaningful soil types using nitrogen (N), phosphorus (P), potassium (K), and pH attributes, enabling abstraction of soil suitability. In the second stage, climatic variables (temperature, humidity, and rainfall) are incorporated to perform crop recommendation using supervised classification models. Baseline methods including Logistic Regression (LR), Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), and Random Forest (RF) are evaluated and compared with a proposed Adaptive Hybrid Dynamic-Adjusting Neural MLP (AHDAN) classifier. Experimental results demonstrate that the proposed approach significantly outperforms baseline models, achieving 99.8% overall Top-3 recommendation success. Soil-wise analysis further confirms the robustness and generalization capability of the hierarchical framework. The results highlight the effectiveness of integrating structured soil abstraction with adaptive neural classification for reliable and scalable crop recommendation systems.