GA-LR参数优化与城市用地动态模拟

In order to better understand the process of urban expansion, how to improve the accuracy of the urban model based on geographic cellular automata has been the direction of academic efforts. Aiming at the deficiency of the conventional logistic regression model, the GA-logistic regression model is constructed, and the dynamic change of the urban land is simulated as an example of the Chengdu Chongqing Economic Zone. In other words, the genetic algorithm is used to fit and obtain the best regression coefficient of logistic regression model to optimize the dynamic model of urban land use. The results show that the accuracy of construction land in urban and rural areas with logistic regression simulation model based on genetic algorithm (GA) is increased from 78% to 85.3%, and the overall kappa index is increased from 74.5% to 81.2%, which is respectively improved by 7.3% and 6.7%.Research shows that the simulation result of parameter optimization of genetic algorithm to optimize variables in logistic regression method than the conventional logistic regression method has better, and more suitable for a large range of city land in dynamic simulation of uneven distribution of data, data features more etc.. Research shows that the effect of simulation of parameter selection method based on genetic algorithm to optimize variables in logistic regression is better than conventional logistic regression method, and genetic algorithm to optimize variables in logistic regression is more suitable for a large range of region with uneven distribution of data, and more variable data etc(10 Points, Times New Roman)