Re: Trend and Seasonality
- From: dave@xxxxxxxxxxx
- Date: 25 May 2006 11:34:00 -0700
Why don't you post the data to the List ... indicating the beginning
date .. and let's see if there are any posters out there ( besides me
! ) who can help you ...
By the way you don't want to rempve anything ...not trend .. not
seasonality ..not day-of-the-week ... not week-of-the-year ... not
UNCLEAN observations ...
You want to integrate into one equation and and all of these components
that are statistically significant ... such as
1. ONE OR MORE TRENDS
2. ONE OR MORE LEVEL SHIFTS
3. ONE OR MORE DAY-OF-THE-WEEK EFFECTS
4. ONE OR MORE WEEK=OF-THE-YEAR EFFECTS
5. ONE OR MORE HOLIDAY EFFECTS
If you don't want to release the data to the list .. just send it to me
and I will post the results ...
Nothing to lose ... All to Gain ..
Dave Reilly
Automatic Forecasting Systems
http://www.autobox.com
215-675-0652
.
- Follow-Ups:
- Re: Trend and Seasonality
- From: SayYes
- Re: Trend and Seasonality
- References:
- Trend and Seasonality
- From: SayYes
- Trend and Seasonality
- Prev by Date: Re: Trend and Seasonality
- Next by Date: Minimum number of cases for a correlation?
- Previous by thread: Re: Trend and Seasonality
- Next by thread: Re: Trend and Seasonality
- Index(es):
Relevant Pages
|