基于不同评价方法的番茄生长综合评价研究

Research on Comprehensive Evaluation of Tomato Growth Based on Different Evaluation Methods

为了探明不同评价方法对水分-沸石耦合调控策略下交替膜下滴灌番茄生长综合评价结果的影响。选取主成分分析法、熵值法和优化层次分析法对不同处理下番茄株高Kh、茎粗Kt、根体积Rv、净光合速率Pn、叶面积指数Lai、维生素C含量VC、可溶性固形物含量SS、产量Yield和水分利用效率WUE共九项指标进行了综合评价,计算得出各处理的综合评价得分。运用主成分分析法提取出生长品质因子和水分效率因子两个主成分,方差贡献率分别为74.420%和17.023%。基于主成分分析法的各处理综合评价得分排序为:Z_(6)W_(100)>Z_(9)W_(100)>Z_(0)W_(100)>Z_(3)W_(100)>Z_(6)W_(75)>Z_(9)W_(75)>Z_(3)W_(75)>Z_(0)W_(75)>Z_(6)W_(50)>Z_(9)W_(50)>Z_(3)W_(50)>Z_(0)W_(50)。运用熵值法得到各指标权重最大的三个依次为WUE、Kh和Yield。基于熵值法的各处理综合评价得分排序为:Z_(6)W_(100)>Z_(6)W_(75)>Z_(3)W_(100)>Z_(9)W_(75)>Z_(9)W_(100)>Z_(6)W_(50)>Z_(3)W_(75)>Z_(9)W_(50)>Z_(0)W_(100)>Z_(3)W_(50)>Z_(0)W_(75)>Z_(0)W_(50)。运用优化层次分析法得到各指标权重最大的3个依次为Yield、WUE和Rv。基于优化层次分析法的各处理综合评价得分由大到小依次为:Z_(6)W_(100)>Z_(6)W_(75)>Z_(3)W_(100)>Z_(9)W_(100)>Z_(9)W_(75)>Z_(6)W_(50)>Z_(3)W_(75)>Z_(0)W_(100)>Z_(9)W_(50)>Z_(3)W_(50)>Z_(0)W_(75)>Z_(0)W_(50)。研究中主成分分析法能反映出每个指标对研究目标的整体影响,但其在数据降维时会有信息丢失的问题。熵值法能较全面的反映每个指标对研究目标的影响和作用,但其在权重赋值过程中会使一些重要程度较高但离散程度较小的指标被赋予较低权重。而优化层次分析法能够兼顾评价指标的重要性和差异性,克服了其他两种方法的不足,为最佳综合评价方法。基于不同评价方法的综合排序得分结果各异,但均表现Z_(6)W_(100)处理为交替膜下滴灌番茄优质高产最优调控策略,Z_(0)W_(50)处理为最劣的调控策略。

In order to explore the effects of comprehensive evaluation results of different evaluation methods for tomato growth under alternate film drip irrigation under the coupled regulation strategy of water-zeolite,in this paper,principal component analysis,entropy method and optimized analytic hierarchy process were selected to comprehensively evaluate the nine indicators of tomatoes under different treatments,these indicators include plant height(Kh),stem diameter(Kt),root volume(Rv),net photosynthetic rate(Pn),leaf area index(Lai)vitamin C content(VC),soluble solid content(SS),yield(yield)and water use efficiency(WUE),and the comprehensive evaluation score of each treatment was calculated.Using the principal component analysis extracts two principal components,growth quality factor and water efficiency factor,and the variance contribution rate were 74.420% and 17.023%.The comprehensive evaluation scores of each treatment based on the principal components analysis are ranked as follows:Z_(6)W_(100)>Z_(9)W_(100)>Z_(0)W_(100)>Z_(3)W_(100)>Z_(6)W_(75)>Z_(9)W_(75)>Z_(3)W_(75)>Z_(0)W_(75)>Z_(6)W_(50)>Z_(9)W_(50)>Z_(3)W_(50)>Z_(0)W_(50).Using the entropy method,the three indicators with the largest weights are WUE,Kh and Yield.The comprehensive evaluation scores of each treatment based on the entropy method are ranked as follows:Z_(6)W_(100)>Z_(6)W_(75)>Z_(3)W_(100)>Z_(9)W_(75)>Z_(9)W_(100)>Z_(6)W_(50)>Z_(3)W_(75)>Z_(9)W_(50)>Z_(0)W_(100)>Z_(3)W_(50)>Z_(0)W_(75)>Z_(0)W_(50).Using the optimization analytic hierarchy process,the three indicators with the largest weights are Yield,WUE and Rv.The comprehensive evaluation scores of each treatment based on the optimized analytic hierarchy process are ranked as follows:Z_(6)W_(100)>Z_(6)W_(75)>Z_(3)W_(100)>Z_(9)W_(100)>Z_(9)W_(75)>Z_(6)W_(50)>Z_(3)W_(75)>Z_(0)W_(100)>Z_(9)W_(50)>Z_(3)W_(50)>Z_(0)W_(75)>Z_(0)W_(50).It can be concluded that the principal component analysis method can reflect the overall impact of each index on the research objective,but it has the problem of information loss when reducing the data dimension.The entropy value method can comprehensively reflect the influence and function of each index on the research target,but in the process of weight assignment,some indexes with a high degree of importance but a small degree of dispersion are given a low weight.The optimized analytic hierarchy process can take into account the importance and difference of evaluation indicators overcome the shortcomings of the other two methods,and is the best comprehensive evaluation method.The comprehensive ranking score results based on different evaluation methods are different.However,all of them show that Z_(6)W_(100)treatment is the optimal control strategy for high quality and high yield of tomato under alternating drip irrigation under mulch,and Z_(0)W_(50)treatment is the worst control strategy.