on 03-20-2008 12:43 PM
Hello,
I have a query.
Our sales data has some seasonal effect.
Say I had a festival in the month of October last year, but this year it is coming in the month of November.
How can I model this shift behaviour using the forecast techniques?
Your help will be greatly appreciated.
Regards,
Rajesh
In any forecast model that uses seasonal component, there is an ability to be able to dynamically calculate the seasonal length.
eg. in strategy 35 or 56, there is a field called "variation" next to period. See the help text below. You may want to set it to 1 or 2. I think this should meet your requirement.
"Variation of Time Period Length
You have selected an automatic model selection procedure . These methods take seasonal effects into account. If you do not make an entry in this field, the system uses a fixed number of periods as defined in the Seasonal Periods field. If you make an entry in the Variation Length field, the system varies the length of the seasonal period and chooses the period that results in the highest autocorrelation value. The value in the field determines how much the period length is varied to either side of the value given in Seasonal Periods .
Example
The following entry was made in the profile:
Seasonal Periods: 4 Variation Length: 2
The system looks for the optimal period length from a value of 2 (4-2) to 6 (4+2)."
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
9 | |
4 | |
3 | |
2 | |
2 | |
1 | |
1 | |
1 | |
1 | |
1 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.