Studies On Consumer Confidence Index

In this paper, Time series of Consumer Confidence Index,ten years of monthly consecutive data of China’s Taiwan province,were used to explore the variation correlation among the six sub-indexes.Firstly, by descriptive statistics, the paper describe the fundamental changes to the form of six sub-index sequences, and then using a combination of the classic model of analysis and time series mining to analyze the changes in the characteristics of the six sub-indexes. Of time series similarity search process, the crucial-boundary-point was found to calculate the distance of two time series when to look for the similarity of them, the method which was used to calculate the distance was a new trail of this paper.Second, found by association rule mining of time series and similarity search, Family Economy sub-index and Regional Economy sub-index is most closely related to each other, the change patterns of this two series is the closest among all of the six indexes. The change pattern of the Investment sub-index is different from other sequences, and Purchase sub-index also shows the unique characteristics.Last, through six sub-indices of the confidence index of the ARIMA model, VAR model analysis, the author found that the sub-index is mainly affected by its own one-month-lag impact. The Investment sub-index significantly different from several other sub-indexes, the Investment sub-index is close to the Random Walk Model.The significant results of the paper is giving the in-depth study of the six sub-indexes, and giving the recommendations of the survey and construction of the six indexes.