摘要
首先利用STATA进行偏相关分析剔除无关变量后再运用k-means均值聚类法将地区进行分类,然后将面积加权平均求出的各个类别的平均温度以及用变异系数法求出的各个类别的权重进行加权平均,最后用SPSS做相关性和偏相关性检验,得出序列非平稳季节波动结论后采用ARIMA程序进行一阶季节性差分时间序列分析,得到了整个南极洲陆地的平均温度以及温度随时间变化的规律,本文的特色是对每个模型的合理性都进行了检验.
The paper, firstly uses STATA to do partial correlation analysis to eliminate the irrelevant variables and uses the k- means average clustering method to categorize area. Secondly, it weights the average temperature of each category and uses the variation coefficient method to calculate the weight of each category. Thirdly, using SPSS to do correlation and partial correlation test, it gets the se- quence of non-stationary seasonal fluctuations conclusion. Lastly, using seasonal ARIMA procedures to analyze one-order seasonal different time series, it gets the average temperature of the entire conti- nent as well as the regularity of temperature change over time. What features this paper is that the ra- tionality of each model is tested.
出处
《兰州文理学院学报（自然科学版）》
2016年第5期32-38,64,共8页
Journal of Lanzhou University of Arts and Science(Natural Sciences)
基金
省级大学生创新项目(AH201410378555)
国家自然科学项目(11301001)
安徽财经大学教研项目(acjyzd201429)
