摘要
试验以三种测度结合的分组主成分法对7个番茄品系的果实商品性状进行了综合评价分析,构建了综合评价函数,结果表明:应用因子分析法提取了2个公共因子,累计方差贡献率达86.59%,变量划分为两组,营养品质和耐贮品质变量为一组,外观品质变量为一组,分别对两个组的变量进行主成分分析,变量的权系数偏重于营养品质和耐贮品质,为0.66,外观品质为0.34,合乎评价的要求,将7个品系的商品性分为四个层次,第一层次是YH02-6,第二层次是YH02-9,第三层次是YH02-7、YH02-1和YH02-11,第四层次YH02-5和YH02-8。此方法把因子分析法和主成分法结合起来,易于变量的分组和降维,考虑了变量测度的转换,使综合评价变量易于正向化,对番茄果实商品性的综合评价有较好的适用性。
The method of the grouped principal component analysis consisting of 3 kinds of measuring methods was applied to evaluate the trade traits of the fruits of tomato lines, comprehensively. The comprehensive evaluation function was built up. The results showed that with the factor analysis, 2 factors were abstracted, the cumulative rate of variance was 86.59%, the variables were classified into 2 groups, the variables of the nourishment quality and the preservative quality were in group 1, and the variables of external appearance quality were in group 2. Then the variables of every group were analyzed with the component analysis, respectively. The coefficient of group 1 was 0.66, and the one of group 2 was 0.34. This was met up to the demand of the evaluation job. The variation coefficients of the nourishment quality and the preservative quality were higher, which was met up to the demand of tile comprehensive evaluation. According to the comprehensive score of the trade traits, YH02-6 was in the first place, YH02-9 was in the second place, YH02-7, YH02-1and YH02-11 were in the third place, and YH02-5and YH02-8 were in the lowest place. This approved that this analytic method was a practical and efficient way to analyze the trade traits of the fruits of tomato comprehensively. It integrated job of grouping variable and lessening the factor analysis and the principal component analysis to make the the dimension easily, and took it to consideration that transforming the variables with 3 kinds of measuring methods.
出处
《中国农学通报》
CSCD
2008年第1期335-339,共5页
Chinese Agricultural Science Bulletin
基金
云南省自然科学基金项目"云南花卉
水稻等植物空间诱变育种技术基础研究"(2001-C0035M)
关键词
因子分析
主成分分析
测度转换
综合评价
Factor analysis
Principal component analysis
The transforming of variable
The comprehensive evaluation
