Genetic Algorithm-Artificial Neural Network Modeling of Capsaicin and Capsorubin Content of Chinese Chili Oil

Chili oil, which contains large amounts of capsaicin and capsorubin, is one of the most consumed seasonings in China. These compounds significantly affect the quality, antioxidant activity, pungency, and color of chili oil. This study aimed to investigate the effect of stewing temperature, stewing time, and amount of oil on the capsaicin and capsorubin contents of Chinese chili oil. The partial least squares (PLS) regression and genetic algorithm-artificial neural network models were established and used to predict capsaicin and capsorubin contents. The genetic algorithm was applied to optimize the parameters of the network. The developed genetic algorithm-artificial neural network, which included ten hidden neurons, predicted capsaicin and capsorubin contents with correlation coefficients of 0.995 and 0.986, respectively. The neural network exhibited more accurate prediction and practicability compared with the PLS regression model.