Modeling of Trends in Time Series based on Maximum Autocorrelation Factors
| Authors | |
|---|---|
| Year of publication | 2008 |
| Type | Article in Proceedings |
| Conference | XVIII. letná škola biometriky BIOMETRICKÉ METÓDY A MODELY V PÔDOHOSPODÁRSKEJ VEDE, VÝSKUME A VÝUČBE |
| MU Faculty or unit | |
| Citation | |
| Field | Applied statistics, operation research |
| Keywords | Maximum autocorrelation factor; Trend; Time series; Ozone; |
| Description | Description of trends in concurrent time series can be based on a nonparametric method suggested by P. Switzer [Analysis of Monitoring Data for Time Trends Using Maximum Autocorrelation Factors. International Association of Mathematical Geology, IAMG06, Liege. 2006]. In this approach, trends in individual time series are expressed as linear combinations of a chosen number of maximal autocorrelation factors an analogy of principal components focusing on the autocorrelation instead of the variance of a component. This paper describes basic ideas and chosen properties of the maximum autocorrelation factors analysis. An application of the considered method in analysis of trends of ozone time series at chosen European monitoring stations is also presented. |
| Related projects: |