Application Of Bayesian Statistical Method To Income And Food Security

This paper applied Bayesian network,Bayesian parameter estimation and naive bayes methods to the problem of income imbalance and food security.We obtained a Bayesian network composed of income-related factors,income distribution functions for each province in the northwest region,and the basic situation of diets in China.In recent years,although the total amount of GDP in China is rising,the problem of income imbalance still exists.Solving the problem of income imbalance can help enhance the stability of society and promote common prosperity.Firstly,this paper studied the factors related to individual income based on Bayesian network.It could be seen that gender,education level,interviewed location and job position had an effect on income from the established Bayesian network.We used the established Bayesian network to classify the income in the test set.The results showed that 80.46%of the total samples were correctly classified,indicating that the classification was effective.Then we used the generalized Gamma income distribution model based on Bayesian estimation to calculate the income inequality indexes,which include Gini coefficient,Theil index and Pietra index.The results showed that income inequality was more severe in Gansu Province and Inner Mongolia Autonomous Region relative to other provinces.In addition to income imbalance,food safety is also a great concern.We evaluated the health risk of mercury based on the results of the Fifth Total Diet Study(TDS)in China.HQ is the ratio between the exposure and the reference value of mercury.The HQs of total mercury in Shanghai,Fujian and Zhejiang were greater than 1,indicating that the exposure to total mercury in these three provinces was hazardous to the health of residents.Then we used naive bayes to classify the HQ of total mercury in each province.The results showed that 90%of the total samples were correctly classified.The accuracy of the classification was high and effective.