AI-Driven Precision Farming: A Scalable Neural Approach for Crop Prediction

Abstract

Precision farming is an application that uses infor mation to maximize crop production and resource use. The paper states that we have JK soil analysis data and it has been trained on a machine learning model to give the prediction of the crops. Environmental factors (Rainfall, Temperature) were evaluated and soil properties (pH, N, P, K, Zn, Fe, Mn). In order to achieve a high data quality and reliability, sophisticated preprocessing methods were used. Z-score-based filtering was used to eliminate outliers and the Min-Max scaling was used to standardize the input space by normalizing numerical features. The Chi-Squared test was used to reduce the most significant 10 features to further do the prediction. The preprocessing pipeline was thorough and the noise and redundancy were reduced to provide a solid dataset to train the model. The neural network was created using two hidden layers (64 and 32 neurons) and attained 98.45 percent test accuracy. Other important contributions are strong outlier management, environmental data incorporation and scalability in newer datasets. The method provides a basis of real-time and scalable accurate farming frameworks, and possible utilizations of pest and fertilizer management.