新可変型季節性指数とその評価に関する研究

抄録

In this paper, a new extraction method of the variable seasonal index was developed which based on learning the extraction methods of the fixed and the variable seasonal index. The new method calls Link Relative Moving (LRM) method. We tested that the variable seasonal index of LRM method used Monte Carlo Experiment, and the result was compared the Economic Planning Agency (EPA) method.
We studied the influence of the different seasonal adjustment index on cyclical fluctuation and the irregular fluctuation. We evaluated that total effect of the seasonal index on LRM method, EPA method and Link Relative method. After decision of period by power-spectral analysis, we compared the irregular fluctuation results on LRM method, EPA method and Link Relative method. In conclusions are; firstly, LRM method is a valuable method among extraction methods of the variable seasonal index; secondly, the result of Monte Carlo Experiment is stabilized in seasonal index of LMR and EPA methods, and the extract power of seasonal index is stronger in LMR method than EPA method; thirdly, the irregular fluctuation is smaller in seasonal adjustment index of LMR method than EPA method and Link Relative method.