土壤中微生物含量影响因素的统计方法分析

发布时间：2018-05-30 15:13

本文选题：方差分析 + 多元线性回归分析　；参考：《微生物学通报》2016年12期

【摘要】：【目的】采用多元方差分析、多元线性回归分析、通径分析等统计方法,并利用多元统计方法对土壤中微生物含量进行分析。【方法】通过研究这3种常用统计方法的原理、应用条件、适用于解决的问题等,用不同的方法分析影响土壤中微生物含量的各种因素,利用SPSS软件求解。【结果】得出不同降雨量下土壤中的微生物含量存在差异性;铵态氮含量和海拔高度与变形菌门含量呈显著的线性关系,土壤中变形菌门含量随着铵态氮含量增多而增多,且铵态氮含量对变形菌门含量所起的直接作用最大;土壤中铵态氮含量、总氮量和海拔高度与放线菌门含量呈显著的线性关系,且均为负相关;其中总氮量对土壤中放线菌门含量的变化所起的负面作用最大,但铵态氮含量对放线菌门含量所起的直接作用最大。对比描述性统计方法与多元方差分析方法可知,应用描述性统计方法只能得出降雨量是影响土壤中不同细菌群落相对含量的因素之一,而不知降雨量对微生物含量变化的影响程度,6个梯度降雨量引起地显著差异究竟是由哪种微生物引起的,方差分析方法可以解决以上的问题。对比多元线性回归分析方法与通径分析方法可知,多元线性回归分析得出的相关系数只能表示铵态氮含量、总氮量和海拔与变形菌门含量(放线菌门含量)之间关系的密切程度,但无法解释和分析这种关系的构成和来源,通径分析方法可以解决这个问题。【结论】通径分析方法在分析相关问题时要优于多元线性回归分析方法,通径分析比多元线性回归分析方法更能体现因素对变形菌门含量(放线菌门含量)的直接影响和间接影响,分析结论更直观、更能说明问题。
[Abstract]:[objective] to study the principles of the three commonly used statistical methods, such as multivariate variance analysis, multivariate linear regression analysis, path analysis and so on. Different methods were used to analyze the factors affecting the microbial content in soil, and SPSS software was used to solve the problem. [results] the difference of soil microbial content under different rainfall was obtained. There was a significant linear relationship between the content of ammonium nitrogen and altitude, and the content of Proteus in soil increased with the increase of ammonium nitrogen content, and the content of ammonium nitrogen had the most direct effect on the content of Proteus. The contents of ammonium nitrogen, total nitrogen and altitude in soil were linearly correlated with actinomycetes content, and all of them were negatively correlated with the content of actinomycetes, and the total nitrogen content had the most negative effect on the change of actinomycetes content in soil. However, the direct effect of ammonium nitrogen content on actinomycetes was the greatest. By comparing descriptive statistical method with multivariate variance analysis, we can only conclude that rainfall is one of the factors affecting the relative content of different bacterial communities in soil. However, we do not know the degree of influence of rainfall on the change of microbial content, the significant difference caused by the six gradient rainfall is caused by which kind of microorganism, the variance analysis method can solve the above problems. By comparing the multiple linear regression analysis and path analysis, we can see that the correlation coefficient obtained by multiple linear regression analysis can only indicate the ammonium nitrogen content. The degree of closeness between total nitrogen and elevation and the content of Proteus (actinomycetes), but it is impossible to explain and analyze the composition and source of this relationship. Path analysis method can solve this problem. [conclusion] path analysis method is superior to multivariate linear regression method in analyzing related problems. Path analysis can reflect the direct and indirect influence of factors on the content of Proteus (actinomycetes), and the conclusion of the analysis is more intuitionistic and more illustrative than the method of multivariate linear regression.
【作者单位】：中国科学院大学数学科学学院;中国科学院大学生命科学学院;
【基金】：国家自然科学基金项目(No.71271204,11331012,11571014) 科教融合专项基金项目(No.KJRH2015)~~
【分类号】：S154.3

【相似文献】