| In the fierce market competition, the quality of soybean paste and the development of new products is the key factor affecting its market share. Accurate and objective evaluation of the quality of soybean paste is a top priority in product improvement and development. In our country at present the study the assessment of the quality of soybean paste, is according to the related standards of state or health of physical and chemical indicators, mainly is conducted on the samples of single index to compare or simple summary analysis. The statistical pattern recognition method of soybean paste quality control and improvement of research rarely reported. This study 20 kinds of commercial soybean paste as the research object, firstly examined seven physical and chemical indicators of 20 kinds of soybean paste, soybean paste are analyzed based on the test results of physical and chemical properties and its corresponding quality analysis. Then based on the physical and chemical properties of soybean paste for pattern recognition research, the 20 kinds of soybean paste is divided into the types of different quality; Finally analyzes the soybean paste that the relative content of different kinds of volatile flavor substances and verified based on the physical and chemical index of the pattern recognition research the rationality of soybean paste quality, for the quality control and improvement of soybean paste and soybean paste the establishment of the new product quality indicators provide certain theoretical basis.(1) The selection of 25 kinds of physical and chemical properties of the soybean paste were sold in accordance with the related provisions of GB/T 24399-2009, the soybean paste samples collected during the ideal quality, each product is not in conformity with the provisions; Different soybean paste products on the physical and chemical properties of the difference was significant(P < 0.01);Seven measures of physical and chemical properties in the same grade of different brand samples differ significantly(P < 0.01).(2) Pattern recognition research: based on the 25 kinds of soybean paste samples of physical and chemical indicators, using principal component analysis, clustering analysis and discriminant analysis, samples reveal the inner link between the physical and chemical indicators. Results will be 20 kinds of soybean paste clustering for two subclasses, seven physical and chemical indexes based on two main component model is established, and the discriminant equation of soybean paste typical D1=-3.505X1-8.403X2+0.335X3+0.790X4+ 0.393X5-1.802X6 + 0.305X7-12.026. Total discriminant accuracy was 100%.(3) Using headspace solid-phase micro extraction and gas chromatography- mass spectrometry technology(HS- SPME/GC/MS) analyzed the volatile flavor components in soy sauce samples, 20 kinds of commercial soy sauce samples to detect the average contains 44 kinds of volatile flavor components, quality good analogy to other categories with more kinds of volatile flavor substances, richer esters and phenols content, etc., so its flavor than other categories also good quality, which can be separated from other kinds soy sauce. Of soy sauce samples volatile flavor components analysis is verified based on the physical and chemical index of the pattern recognition research sauce and the rationality of the quality of soy sauce and classification for the quality control and improvement of soy sauce and soy sauce the establishment of the new product quality indicators provide certain theoretical basis. |
PDF Full Text Request