|
Abstract:
|
Seasonal adjustment is the process of accounting for regular seasonal patterns in a time series. By removing the seasonal effect from the series, a better understanding of the underlying dynamics may be revealed. Weather effects can contribute to a seasonal pattern (for example, agricultural series are likely to experience different levels of activity across seasons that happen to correlate to months). Weather effects can vary greatly from year to year, which is generally not the case for the regular seasonal patterns that are typically handled with seasonal adjustment. An unaccounted-for weather effect may end up being dismissed as an outlier, so the ability to accommodate a weather effect for seasonal adjustment could have some interpretative value. We examine weather data in an attempt to find weather regressors that can help achieve this.
|
