陕西省县域城镇化发展水平的BP神经网络测定研究

运用BP神经网络的理论和方法,构建了县域城镇化发展水平BP神经网络模型,并对2008年陕西省83个县域城镇化发展水平进行了综合评价。结果表明：陕西省县域城镇化发展水平区域差异显著,评价结果与专家的判断基本一致;根据评价结果,结合ARC-GIS9.2中的自然断点法（natural breaks）将BP测度得出的得分值分为5个等级,发现县域城镇化发展水平较高的县域主要集中在陕北地区,关中地区次之,陕南地区最低。BP神经网络用于评价县域城镇化发展水平简单、实用,且避免了人工确定指标权重的主观性,具有很好应用前景。

Analysis on County Development Level of Urbanization by Using BP Neural Networks in Shaanxi Province

BP neural network model for the county development level of urbanization was constructed by using the theory and method of BP neural network.The comprehensive evaluation was made on the urbanization development level of 83 counties in Shaanxi Province in 2008.The results showed that there was significant regional difference of county development urbanization level of Shaanxi Province and the evaluation results were basically accordant with the judgment of experts.Based on the evaluation results,combined with natural breaks method in ARCGIS9.2,the score values were divided into 5 categories.The results showed that the county development level of urbanization was highest in the north of Shaanxi,followed by Guanzhong region,and that in Southern Region of Shaanxi was lowest.BP neural network is simple and feasible for assessing the county development level of urbanization.It avoided the subjectivity of determinming the weight of indices artificially and had a good application foreground.