Simultaneous Prediction by Hyphenated ANN Regression Model (GA-PLS-ANN) for Amino Acids in Dried Ground Tea Leaves

Application of Chemometrics to Absorbances in Near Infrared Region

Abstract

The technique of chemometrics which combined the artificial neural networks (ANN) with Genetic Algorithm (GA) and Partial Least Squares (PLS) was applied to the near infrared spectroscopic absorbances obtained from dried ground tea leaves. This hyphenated artificial neural networks regression model (GA-PLS-ANN) was constructed in order to predict the total content f 7 kinds of amino acids, and individual glutamine, arginine, and theanine, simultaneously.
Using the 4 latent variables as the input signal, the 4-4-4 hierarchy type ANN regression model as applied in the training set, and was evaluated in a prediction set. Correlation coefficients (r) between the calculated value, and the actual amount of amino acid in a prediction set were as follows, Glutamine: 0.952, Arginine: 0.874, Theanine: 0.965, Total amount of amino acids: 0.991
The estimation of the ANN regression model obtained was very precise. The GA-PLS-ANN regression model, when compared with the conventional multiple linear regression (MLR) model, greatly reduced the number of essential variables, and was more adaptable for the prediction of other unknown samples. In addition, the GA-PLS-ANN regression model was also suited for a nonlinear relation between variables and amino acids.