依据可持续发展、生产资源密集度、区位优势、关联带动这四个原则构建了一个能源型城市主导产业选择的两层指标体系.在此基础上,为解决PCA不适合多指标综合分析中非线性主成分分析的问题,采用KPCA建立了一个能源型城市主导产业选择的理论模型.最后以陕西省铜川为例,证明了KPCA不仅适用于非线性问题,而且具有优秀的主元提取功能.
Focusing on the selection of dominant industries for energy cities, this paper puts forward a two-layer index sytern based on the four principles: continuous development, manufacturing resource density, position strength and industry correlation. And then a KPCA modeling of selecting dominant industries for energy cities is proposed in order to overcome the defect of classical PCA, that is, PCA is unable to settle the non-llnear problem in multi-attribute evaluating analysis. Finally, a case of Tongchuan in Shanxi Province demonstrates that KPCA is not only suitable for non-linear problems, but also an excellent method of extracting feature.