摘要
结合中国国家统计局的数据,使用基于核主成分分析与加权支持向量机的方法建立模型,找出了影响就业(失业)的22个主要指标.考虑到这些指标相互之间的相关性,使用核主成分分析与加权支持向量机的方法建模,给出了算法的步骤,构造了非线性预测函数,并对1995-2009年的城镇登记失业率进行拟合预测,得到的结果具有较高的精度.
In this paper,combining with the date of China＇s National Bureau of Statistics,we used the Kernel Principal Component Analysis and Kernel Weighted Support Vector Machine method to build the mathematical model.We found 22 key indicators which affect the employment（unemployment）.Taking into account of the correlation among these indicators,we used the Kernel Principal Component method analysis and Weighted Support Vector Machine method to build the model and give the algorithm on how to construct a non-linear prediction function,and also we used the registered urban unemployment rate during 1995-2009 to fit the predicted results and obtain high accuracy.
出处
《海南师范大学学报（自然科学版）》
CAS
2010年第4期372-374,共3页
Journal of Hainan Normal University（Natural Science）
关键词
失业率
核主成分
支持向量机
unemployment
Kernel Principal Component
Support Vector Machine
