基于回归分析的元胞自动机模拟精度改进——以蔚县剪纸艺人分布研究为例

The Improvement of the Simulation Accuracy of the Cellular Automata(CA) Based on the Regression Analysis: the Case Study of the Distribution of Yuxian Paper-Cut Artists

利用回归分析方法对蔚县剪纸知名艺人数据进行筛选,并将筛选结果应用于元胞自动机模拟,该方法使元胞自动机模拟精度明显提高。以蔚县知名剪纸艺人分布村庄时间序列数据为研究对象做对比试验,首先在不考虑离群数据以及增长率对模拟效果影响的情况下,抽取两期数据进行模拟分析并加以验证,正确率为60.12%,表明CA_Markov模型可以应用于对文化景观的模拟。然后应用回归分析的方法,考虑离群数据以及增长率对模拟效果的影响筛选试验数据进行,并应用模拟规则进行模拟,其模拟正确率达到74.46%,预测精度显著提高。

By using regression analysis method,the data of the famous paper-cut artists in Yuxian were screened.And the result obtained was applied to the cellular automata simulation.This method could improve the simulation accuracy of the cellular automata significantly.Taking continuous data simulation of the distribution of Yuxian’s villages which had well-known paper-cut artists from 1834 to 1966 as an example,without considering the effect of outliers and growth rate,datas in these two groups were extracted for simulation analysis and verification,and the correct rate turned out to be 60.12%.The experiment showed that the CA-Markov model could be applied to cultural landscape simulation.Besides,when the impact on simulation results due to outliers and growth rate was considered and regression analysis was used to select experimental data,the correct rate of the simulation results could be over 74.46%,which meaned the accuracy of prediction of this method had been improved remarkably according to hypothesis testing.