Source Code for RBF Neural Network-Based Wine Evaluation Model

During the 2012 "Higher Education Cup" Mathematical Contest in Modeling, our team investigated wine quality evaluation methodologies. We developed an assessment model based on Radial Basis Function (RBF) neural networks and successfully implemented the corresponding source code. The implementation utilizes MATLAB's neural network toolbox, featuring automatic determination of hidden layer neuron centers through clustering algorithms and optimized weight calculation using least squares methods. We collected extensive datasets containing wine physicochemical properties and sensory evaluations, alongside implementing classical data preprocessing techniques including normalization, outlier detection using z-score analysis, and feature scaling to ensure model accuracy in assessing wine quality and taste characteristics. This research holds significant practical implications for the wine production and sales industry, as it enables better understanding of consumer preferences and facilitates production of higher quality products through data-driven quality assessment.