Using Pattern Recognition To Establish A Screening Method About Effect Of Soybean Quality To Plaster Tofu

The tofu processed by different soybean varieties differs greatly in texture,yield and even composition.However,for companies and breeders,they do not know how to select varieties that are more suitable for processing soybean products.In particular,soybean breeding scientists often face thousands of varieties.The strains cannot be screened for the dominant varieties that are suitable for processing soy products.Therefore,it is necessary to study the screening methods for high-throughput tofu processing suitability for soybean strains or varieties.In this experiment,68different soybeans in different regions collected by the Processing Research Laboratory of the National Soybean Industrial Technology R&D Center was used as sample,and determination of theirs protein,fat,water content,phytic acid,7S and 11S content of protein subunits,calcium,iron,phosphorus,magnesium,lipoxygenase activity,trypsin inhibitors.And calculate the ratio of protein and fat and the ratio of 7S,11S protein subunit.Then,the soybean milk was prepared,the soybean protein was determined and the protein recovery rate was calculated.The tofu was prepared and studied for texture.Principal component analysis,cluster analysis,and discriminant analysis were performed by pattern recognition method.Finally,a method for screening soybeans suitable for processing tofu was established.It provides a convenient and theoretical basis for future soybean breeding,a good and method for the selection of varieties in tofu processing,and a theoretical basis for industrialization and mechanized production of tofu.The main findings of this paper are as follows:?1?The contents of protein,fat,water content,phytic acid,protein subunits,and the contents of calcium,iron,phosphorus,magnesium,lipoxygenase,and trypsin inhibitors in 68different soybean samples used in the present study.Both are significantly different.?2?Regression analysis leads to the following conclusions The yield of tofu Y₁=-14.234X₄+33.246X₆-1.312X₇+363.073,hardnessY₂=-19.499X₂+294.143X₆-16.774X₁₁+257.143,glutinosityY₃=-4.706X₁₁-1.379X₁₂+0.507X₁₆+238.135,ChewinessY₄=-4.175X₁₁-1.463X₁₂+0.537X₁₆+227.753,resiliencyY₅=-0.025X₃-0.110X₄-0.003X₁₂+1.680,CohesionY₆=-0.015X₃-0.078X₄-0.002X₁₂+0.826,ResponsivenessY₇=0.0001X₁₄+0.0001X₁₅+0.052The texture index of tofu is mainly influenced by fat content,ratio of egg to fat,soymilk protein,phytic acid,trypsin inhibitor,calcium,iron and magnesium content.?3?Pattern recognition studies:Through the study of the physical and chemical indicators of 68 different soybean samples and the texture characteristics of soybean milk and tofu,principal component analysis,regression analysis,cluster analysis and discriminant analysis were used to reveal the intrinsic relationship between them.6 main components were extracted from the total complex relationship of 17 physical and chemical indicators of soybean and soybean milk.Clustering 68 kinds of soybeans into the most suitable for tofu production,suitable for the production of tofu but lower yield,can produce tofu but general class of texture index and not suitable for tofu production 4 subclasses,and got 4 typical discriminant equations,The total discriminanant accuracy was 95.6%...