CAJ | 학술논문

运用人工神经网络的理论和方法，构建BP神经网络，评价2009年甘肃省县域城市化水平，将87个县域城市化水平分为5级。对频数分布特征、变异系数、威廉森系数和最大与最小系数的分析表明，甘肃省县域城市化空间分异显著。具体表现为：呈正偏态分布，第三、四级别的县市比例较大；城市化水平发展不均衡，呈现西北-东南差异；经济区内部差异大，表现为西北高、东南低的趋势。利用Spearman’s rho相关分析得出影响城市化水平的因素及相关度。
운용인공신경망락적이론화방법，구건BP신경망락，평개2009년감숙성현역성시화수평，장87개현역성시화수평분위5급。대빈수분포특정、변이계수、위렴삼계수화최대여최소계수적분석표명，감숙성현역성시화공간분이현저。구체표현위：정정편태분포，제삼、사급별적현시비례교대；성시화수평발전불균형，정현서북-동남차이；경제구내부차이대，표현위서북고、동남저적추세。이용Spearman’s rho상관분석득출영향성시화수평적인소급상관도。
Choosing 87 county areas in Gansu Province as study object, this paper first selects 17 representa- tive indicators from the aspects of space concentration level, economic progress level, social development level and infrastructural facility construction level, constructs index system to evaluate urbanization level by using artificial neural network theory, based on statistic data of Gansu Province in 2009. Then, urbanization level of 87 county ar- eas are classified into five degrees. Moreover, the paper analyzes frequency distribution features and calculates vari- ation coefficient, William coefficient, maximal and minimal coefficient, finding： 1 ） Its frequency distribution has positive skewness features, and a bigger proportion of counties in the third and fourth degree; 2） The urbanization level development is uneven, and it is declined from northwest to southeast ; 3 ） Internal differentiations of five eco- nomic regions declined from northwest to southeast. Finally, the Spearman＇ s rho correlation analysis indicated that the level of economic growth is the greatest impacting factor of urbanization level, which is also the powerful driving force; The proportion of non-agricultural population, per capita GDP, per capita retail sales of social consumer goods and the number per million people own a mobile phone are the most relevant factors of urbanization level, meanwhile, natural population growth rate and the number of students per million people in the school are negative- ly correlated with urbanization level.