中国是世界马铃薯生产和消费大国,2015年初,国家农业部因势而谋推进马铃薯主粮化战略,对提升马铃薯重要地位和推进马铃薯加工进步具有积极作用。系统研究马铃薯原料性状与制品品质间关系,对促进马铃薯产业发展和升级具有积极的借鉴意义。试验采用最大-最小归一化处理方法将马铃薯油炸薯片、油炸薯条及雪花全粉各个品质指标转化为一维的综合评价指标,分别与44个马铃薯原料品种的特征指标进行拟合并建立回归模型,建立的马铃薯油炸薯片综合品质评价模型决定系数R2=0.802,调整后决定系数R2=0.776,随机误差估计值σ=0.101;马铃薯油炸薯条综合品质评价模型决定系数R2=0.731,调整后R2=0.704,随机误差估计值σ=0.141;马铃薯雪花全粉综合品质评价模型决定系数R2=0.705,调整后R2=0.682,随机误差估计值σ=0.170;利用K-means聚类算法将44个品种按加工用途划分为最适宜、较适宜和不适宜3类,得出最适宜加工油炸薯片的15个品种,最适宜加工油炸薯条和雪花全粉各10个品种。结果表明,3个综合品质评价模型拟合度较高,误差较小,模型效果可靠,可用于实际马铃薯加工制品品质评价;K-means聚类结果与实际应用情况相符,可为筛选加工专用品种提供参考和借鉴。
China is one of the world's largest countries in potato production and consumption, and the quality of processing products was significantly affected by characteristics of raw materials. In order to research the relationship of raw materials and their processed products(potato chips, French fries and potato flakes), 44 common potato cultivars grown in domestic were collected and tested in China National Food Machinery Quality Supervision Inspection Center and Chinese Academy of Agricultural Sciences from Sep. 2012 to Dec. 2014. Fourteen indicators including eye numbers, shape index, mean weight, reducing sugar, total sugar, starch, moisture, dry matter, ash, total soluble solids, vitamin C, protein, fat and crude fiber of raw potatoes were analyzed according to China national or industrial standards. Analysis of variance revealed significant differences and great variation among most of indicators. Three kinds of products including potato chips, French fries and potato flakes were processed using 44 cultivars collected respectively. The quality indicators of processed products were also tested in the laboratory. For potato chips, four indicators including sensory evaluation, protein, crispness and Hunter whiteness were tested. For French fries, four indicators including sensory evaluation, protein, hardness and Hunter whiteness were tested. For potato flakes, four indicators including sensory evaluation, protein, viscosity and free starch ratio were tested. Analysis of variance also revealed great variation among indicators of three kinds of processed products. Using the maximum- minimum normalized method, multiple quality indicators of potato chips, French fries and potato flakes were calculated into a one-dimensional evaluation indicator in order to establish a representative model. Each comprehensive indicator had a significant correlation with four original indicators of processed products. With stepwise regression method, significant indicators of raw materials were selected to establish the model of p