Qualitative discrimination of yeast fermentation stages based on an olfactory visualization sensor system integrated with a pattern recognition algorithm

The volatile organic compounds produced in yeast fermentation are directly related to the degree of fermentation and product quality. This study innovatively proposes a method based on an olfactory visualization sensor system combined with a pattern recognition algorithm to ensure the correct discrimination of the yeast fermentation stages. First, the olfactory visualization sensor system was developed based on a colorimetric sensor array, which was composed of twelve chemical dyes including eleven porphyrins or metalloporphyrins and one pH indicator on a C2 reverse silica-gel flat plate. It was employed as an artificial olfactory sensor system to obtain odor information during the process of yeast fermentation. Then, principal component analysis (PCA) was used to reduce the dimension of the data, which were obtained from the olfactory visualization sensor system. Finally, three pattern recognition algorithms, i.e., support vector machine (SVM), extreme learning machine (ELM) and random forest (RF), were used to develop identification models for monitoring the yeast fermentation stages. The results showed that the optimum SVM model was superior to the ELM and RF models with a discrimination rate of 100% in the prediction process. The overall results sufficiently demonstrate that the olfactory visualization sensor system integrated with an appropriate pattern recognition algorithm has a promising potential for the in situ monitoring of yeast fermentation.