关键词: 水稻；主成分分析；隶属函数；品质综合评价；聚类分析

Abstract: In order to explore the differences in the quality indexes of different rice varieties, this study used principal component analysis (PCA) and membership function analysis (MFA) for comprehensive evaluation of 13 quality indexes of 46 high-quality japonica rice varieties in southern regions of Northeast China, and classified the rice varieties by cluster analysis (CA). The results showed that there were significant differences in all quality indexes among different rice varieties. Among the quality indexes, the coefficient of variation for grain chalkiness rate was the largest (66.32%), while that for amylose content was the smallest (3.28%). Correlation analysis showed that there were different degrees of correlation among the 13 quality indexes. Through PCA, four principal components (PCs) were determined to comprehensively evaluate the quality of different rice varieties. PC1 was synthesized from appearance, hardness, viscosity, balance degree and taste score, PC2 from the percentage of brown rice, milled rice and whole milled rice, PC3 from grain chalkiness rate and chalkiness degree, and PC4 from length-to-width ratio, amylose content and protein content. The quality of the 46 rice varieties was evaluated and ranked by MFA. The top 10 varieties were Liaogeng 1925, Yangeng 935, Tiegengxiang 3, Shennong 625, Liaogeng 1499, Liaogengxiang 1, Yangeng 241, Liaogeng 419, Tiegeng 11 and Yangeng 313. The 46 varieties were classified into five categories by CA. Class I included one variety (2.17%), whose quality was the best. Class II included 18 varieties (39.13%), and their rice quality was excellent. Class III included 18 varieties (39.13%), and their rice quality was moderate. Class IV included 7 varieties, and their rice quality was poor. Class V included two varieties (4.35%), and their rice quality was the worst. This conclusion can provide an important reference for screening high quality rice varieties.

Key words: rice; principal component analysis; membership function; comprehensive quality evaluation; cluster analysis