Gbm Based Policy Influence Analysis Of Agent Simulation Parameters

In this paper, GBM(generalized boosting model) based policy influence analysis is presented for house market agent simulation. In order to execute the agent simulation, various simulation parameters must be set. Among the parameters to be set, there are policy parameters that affect the policy goal, which is the main result of the simulation. Therefore, if simulation is executed by setting policy parameters variously, the results of simulation may be different. Simulation is usually performed by combining policy parameters based on policy objectives and then the result is analyzed whether the simulation result meets policy goals or not. However, it is difficult to analyze what kind of policy parameters affect policy objectives among policy parameters. In this paper, in order to analyze how the policy parameters affect policy goals, we set policy parameters in various combinations, and then execute social phenomenon agent simulations, and then analyze the impact of policy parameters using GBM algorithm.