Poverty improvement policies and household income: Evidence from China

The purposes of this research is to analyzed the impact of various policies on household incomes and to obtain the influential factors according to their importance. Utilizing a survey data of poor households supported by poverty improvement policies, for individual income, this research implements the traditional random forest method. To analyze the influence of strategies on the overall household earnings, we propose a novel splitting method centered on copula entropy and construct a novel choice tree. Finally, we conclude that the policies are generally effective but that levels of effectiveness are varied for people with different levels of income. We also identified the policy arrangement that has the greatest impact on household income.