Data From Electronic Nose for Biopharmaceutical Powder Discrimination Comparative PCA–ANN Pattern Recognition

This dataset and code package represent the supplementary material for the study titled:"Multi-Chamber Electronic Nose for Biopharmaceutical Powder Discrimination: A Comparative PCA–ANN Pattern Recognition Approach" This repository contains the full preprocessed sensor response datasets, computational scripts, and experimental setup documentations used to discriminate and classify herbal/biopharmaceutical powder samples using a multi-chamber electronic nose (e-nose) system. Contents of this supplementary package:1. Appendix 1. Gas Sensor Responses after Data Preprocessing: - Complete tabular sensor response datasets (S1–S13) evaluated across six natural/biopharmaceutical powder samples: Black Pepper, Sambiloto (Andrographis paniculata), White Ginger, Lemongrass, Moringa Leaves, and Garlic. - Comparative data tables for both Absolute Data preprocessing and Normalized Absolute Data preprocessing methods. 2. Appendix 2. Principal Component Analysis (PCA) Script: - Complete MATLAB code for dimensionality reduction, variance explanation calculation (PC1 = 70.89%, PC2 = 17.12%), 2D score plot generation, and 3D PCA scatter visualization. 3. Appendix 3. Artificial Neural Network (ANN) Script: - MATLAB implementation code for feedforward backpropagation neural network training (`newff`, `trainlm`) including network initialization, hyperparameter configurations (hidden layer sizes h1=30, h2=40, learning rate lr=0.5, momentum mc=0.9), MSE performance monitoring, and simulation test scripts. 4. Appendix 4. Experimental Documentation: - Visualization outputs and hardware configuration details for the multi-chamber electronic nose setup. Appendix 4. Experimental & Hardware Documentation: Visualizations, hardware configurations, and chamber assembly details of the custom-built/in-house multi-chamber electronic nose system."