Research On Path And Decision Analysis Model Application Based On Regression Analysis

Smog pollution seriously affects air quality,and at the same time it also has a major hazard to human health.The concentration of PM_2.5in the air can reflect the severity of haze pollution.There are many factors that affect PM_2.5concentration.It is necessary to scientifically determine the importance of PM_2.5influencing factors.Reasonably controlling these factors to alleviate haze pollution and optimizing air quality is extremely urgent.Similarly,air pollution,changes in annual precipitation,and untimely water demand and irrigation have caused important challenges to dryland wheat breeding.It has become very important to study the relationship between the various factors that affect wheat yield and determine the importance of these factors in order to optimize breeding to increase yield and promote long-term stable increase in dryland wheat.This paper uses regression-based path analysis and decision analysis methods to study the importance of PM_2.5influencing factors and wheat yield in arid areas.The core of this method is the decomposition of the total correlation coefficient and the total determination coefficient between the dependent variables and the independent variables.The interaction between influencing factors is quantified,and the calculation of decision coefficients comprehensively reflects the order of importance of influencing factors.The specific results are as follows:Take the winter haze in Taiyuan City from 2017 to 2019 as an example.Select PM_2.5as the dependent variable,air quality indicators and weather indicators,a total of 10 influencing factors as independent variables to carry out empirical application of"one-to-multiple"path analysis and decision analysis.Path analysis shows that there are six factors that have a greater effect on the total PM_2.5,but there are obvious differences in the way of action.PM₁₀,CO and humidity are mainly controlled directly,while NO₂,SO₂and visibility are mainly controlled by indirect regulation related to PM₁₀,CO and humidity.The result of decision analysis showed that PM₁₀,CO,humidity and O₃are all significant decision factors.Comprehensive analysis shows that PM₁₀,CO and humidity are the three most important factors affecting PM_2.5concentration.That is,in the process of controlling the concentration of PM_2.5,we must first adjust these three factors.When exploring the breeding law of dryland wheat,the number of spikes per plant,the number of grains per spike,and the weight of 1000 grains were selected as dependent variables,and the biomass per plant,grass weight per stalk,and economic coefficient were used as independent variables for research.First,the three-repetition"one-to-multiple"path decision analysis model is used to analyze each dependent variable separately.However,the contradictory decision-making measures of different independent variables versus dependent variables cause breeders to be confused.Furthermore,the"multiple-to-multiple"path decision-making model is used for analysis,and three independent variables are considered as the correlation between the dependent variables caused by the common cause,and a total of five paths are obtained.The generalized determination coefficient is divided into five paths,which clearly quantifies the control relationship between variables.The calculation results of the generalized decision coefficients show that the economic coefficient should be increased while reducing the biomass of a single plant in the breeding process.