选取福建省23个市辖区及地级市为样本,选取11个反映区域经济发展水平的评价指标,采用Matlab7.1建立SOM网络模型进行区域经济发展差异分类,得出23个市辖区及地级市的分类结果.结果表明:利用SOM网络模型可将23个市辖区及地级市划分成六类:高度发达型、发达型、较发达型、一般发达型、次发达型及欠发达型,其中前三类城市均分布于东南沿海地区,后三类城市大部分位于内陆地区;分类结果与研究区实际情况较相符,较合理地反映区域经济发展差异的情况;SOM网络模型进行区域经济发展差异的分类具有简便、实用的特点,且避免了人工确定指标权重的主观性,具有良好的应用价值.
The study selects 23 cities as the research samples and 11 evaluating indicators reflecting the level of regional economic development,and sets up the SOM model for the difference of regional economic development classification by using Matlab7.1,with the result of 23 jurisdictional and prefecture-level cities' classification.The result shows that by using the SOM model can classify 23 jurisdictional and prefecture-level cities into 6 models:Highly developed,the More Developed,developed,General developed,Sub-develop and Less-develop,among which the first three types are all distributed throughout the southeast coastal areas and the last three ones almost are located in inland areas.This classification result is basically similar to the expert's judgment which can get reasonable reflection of the difference of regional economic development.The evaluation result also indicates that the application of SOM network in the difference of regional economic development classification is simple,precise and feasible as it avoids the subjectivity of artifical determine the index weightwith a good application value.