Application of Near-infrared Spectroscopy Combined with iPLS_SPA Band Screening in the Prediction Model of Yellow Water Alcohol Content
Article
Figures
Metrics
Preview PDF
Reference
Related
Cited by
Materials
Abstract:
In order to realize the rapid detection of the alcohol content of yellow water in the process of liquor fermentation, Fourier near-infrared spectroscopy (FT-NIR) technique was used to collect the spectrum of yellow water, and the partial least squares regression (PLSR) method was used to establish the alcohol content predictionmodel. In order to reduce the data redundancy of the full spectrum, reduce the complexity, and improve the modeling accuracy, the continuous projection algorithm (SPA) and the interval partial least squares (iPLS) method were used to screen the characteristic bands of the entire spectral region, and the coefficient of determination R2 and root mean squared error of prediction (RMSEP) were used to evaluate the predictive models. The results showed that compared with the original data set, the prediction model after abnormal sample removal, preprocessing and characteristic spectrum screening, the prediction set R2 also changed from initial 0.702 to 0.952 (an increase of 35.61%); the predicted RMSEP changed from 3.812 to 1.367 (a decrease of 64.14%); the number of variables also gradually decreased from 2,203 to 99 (a decrease of 95.51%). It showed that while reducing irrelevant information and noise, the complexity of the model has also been greatly improved, and the stability and accuracy of the model have been effectively increased, and finally, the rapid and non-destructive detection of the alcohol content in yellow water can be realized, in order to provide a new possibility in the liquor fermentation field, and provide a theoretical basis for the detectionof near-infrared in liquor fermentation by-products.
