Classifying Beers With Memristor Neural Network Algorithm in a Portable Electronic Nose System

Quality control and counterfeit product detection have become exceedingly important due to the vertical market of beers in global economy. China is largest producer beer globally has a massive problem with alcoholic beverages. In this research, modular electronic nose system 4 MOS gas sensors was designed for collecting models from four different brands Chinese beers. A sample delivery subsystem fabricated inject clean samples. software-based data acquisition programmed record time-dependent chemical responses 28 models. back-propagation neural network based on memristor proposed classify quality Data collected were then used train, validate, test created model. Over 70 tests changes setup parameters, feature extraction methods, parameters performed analyze classification performance hardware network. Samples experiments showed deviation 9% mean value. The able beers, 88.3% accuracy. Because algorithm easy fabricate hardware, it reasonable design an instrument low cost high accuracy near future.