由于马铃薯品种繁多，不同原料品质会影响马铃薯面条的品质，为研究不同马铃薯品种加工面条的适宜性，指导马铃薯产业选择加工种植品种。试验采用最大-最小归一化处理方法将马铃薯面团和马铃薯面条的各个品质指标转化为一维的综合评价指标，分别与40个马铃薯原料品种的品质指标（淀粉、维C、可溶性蛋白、还原糖、粗纤维、钾、灰分、干物质、游离氨基酸、硬度、弹性、内聚性、咀嚼性、胶黏性、回复性）进行拟合并建立回归模型，采用聚类方法，将40个马铃薯品种按照加工适宜性分为三类。结果表明，建立的马铃薯面团综合品质评价模型决定系数R2=0.911，调整后决定系数R2=0.904，随机误差估计值=123.113；马铃薯面条综合评价模型决定系数R2=0.973，调整后决定系数R2=0.971，随机误差估计值=432.226；利用K-means聚类算法将40个马铃薯品种按加工适宜性划分为最适宜、基本适宜和不适宜3类，其中最适宜加工的品种为05-44-1、克9、79（2）、C11、D17、C3、T3、庄3、会2、T4、L7、黑共12个品种，基本适宜品种为L0524-2、冀8、T2、F5、红、郑7、S3-28、中901、S4-32、中13、T5、冀12、甘农5、晋18、F6、T18、78共17个。该研究得到的综合品质评价模型拟合度较高，误差较小，模型效果可靠，可用于实际马铃薯面条的品质评价；K-means聚类结果与实际应用情况相符，可为新疆地区筛选加工专用品种提供参考和借鉴。

Due to the existence of many potato varieties, the quality of different raw materials will affect the quality of potato noodles. The purpose was to study the suitability of different potato varieties for processing noodles and guide the potato industry to select processing and planting varieties. The maximum-minimum normalization method was used to transform each quality index of potato dough and potato noodles into a one-dimensional comprehensive evaluation index, which was compared with 40 varieties of potato raw materials (starch, vitamin C, soluble protein, reducing sugar, crude fiber, potassium, ash, dry matter, free amino acid). Fitting and regression models were established for hardness, elasticity, cohesion, chewiness, stickiness and resilience. Forty potato varieties were classified into three categories according to processing suitability by clustering method. The results showed that the determinant coefficient R2 was 0.911, the adjusted determinant coefficient R2 was 0.904, the random error estimate was 123.113, the determinant coefficient R2 was 0.973, the adjusted determinant coefficient R2 was 0.971, and the random error estimate was 432.226; K-means clustering algorithm was used. According to processing suitability, 40 potato varieties were divided into three categories: the most suitable, the basically suitable and the unsuitable. Among them, 12 varieties were suitable for processing, including 05-44-1, Ke 9, 79(2), C11, D17, C3, T3, Zhuang 3, Hui 2, T4, L7 and Hei. The basically suitable varieties were L0524-2, Ji 8, T2, F5, Hong, Zheng 7, S3-28, Zhong901, S4-32, Zhong13, T5, Ji 12, Gannong 5, Jin 18, F6, T18 and 78. The comprehensive quality evaluation model obtained in this study had high fitting degree, small error and reliable effect. It can be used to evaluate the quality of potato noodles. The K-means clustering results are consistent with the actual application, which can provide reference for the selection of processing varieties in Xinjiang.

乌鲁木齐市科学技术计划项目（G161210004）；国家重点研发计划项目（2016YFD0401304-04）