基于投影寻踪的西藏野生波伐早熟禾坪用价值评价

针对以往草坪质量评价方法存在人为权重赋值误差,以及计算繁琐等方面的问题,提出了对草坪坪用质量指标综合评价的投影寻踪等级评价方法。该方法依据标准样本自身的数据特性寻求最佳投影方向,通过最佳投影方向与评价指标的线性投影得到投影指标值,依据投影指标值对样本进行统一评价和分类,并利用最佳投影方向判断各评价指标对综合评价目标的贡献大小。应用该方法较好地对西藏林芝地区野生波伐早熟禾(Poa poophagorum Bor)的坪用等级进行了评价,并对单一坪用指标对坪用质量综合评价的贡献进行了分析,波伐早熟禾草坪等级为4级;此评价方法精确度较高,分析结果平均绝对误差为0.15%,平均相对误差为6.21%。

The Application of Projection Pursuit Grade Evaluation Model in Quality Assessment of Poa poophagorum Bor Turf

Projection pursuit grade evaluation model based on genetic algorithm was used to avoid jamming of weight matrix and to simplify the mathematic calculation in other comprehensive evaluation system.The projection index value was given by multiplying projection vector and grading standardization value of evaluation indices,which can synthesize the characteristics of turf samples.The method had better applicability for evaluating the turf quality grade criterion of Poa poophagorum Bor in Tibet.The turf quality was classified as 4 grades.The error analysis results showed that the absolute error was 0.15% and the relative error was 6.21%.