Control Weekly Seasonality
Hi.
I have a question about weekly statistic forecast model.
In this model, can we control the difference of weeks in another year?
Suppose that, some food's demand will increase dramatically on 01/01(New Year).
・2014/01/01 - Week 1 FY14
・2015/01/01 - Week 52 FY14
In this case, I think we cannot control the Weekly Seasonality.
This case will happen in another situation(ex) Christmas, Halloween).
Answers
-
Do you want to normalize it or just calculate the seasonality within the same Year? Share your use case here with examples, will try our best to see if that can be achieved.
Misbah
0 -
It seems that 1 Jan is on 1st week in every year.
And you can set a flag on the specific week in this way.
You can use these flags as dummy variables in your statistic model.I hope it helps,
Taichi
>>>>
I have changed the end of fiscal year setting.
The week of 1 Jan moves, and It still works.
0 -
Hi Misbah.
I just want to control the weekly seasonality within different year.In this below case, the sales increase in 1/1-8, and decrease 1/9-16.
I think This is Seasonality.
But when we consider the weekly seasonality, can we catch it?
Week 1,2 are different in 2019 and 2020.
Day 1/1 1/2 1/3 1/4 1/5 1/6 1/7 1/8 1/9 1/10 1/11 1/12 2019 week w1 w1 w1 w1 w1 w1 w1 w2 w2 w2 w2 ・・・ Sales 100 200 300 400 500 600 700 700 600 500 400 2020 week w52 w52 w52 w52 w52 w52 w1 w1 w1 w1 w1 ・・・ 100 200 300 400 500 600 700 700 600 500 400
Hi Amaya.
OK, I understood the case.
But I think we can not control the another situation like public holidays.
There are many public holidays in one year and they are in different week each year.
Thanks for replay!p.s
January sun mon tue wed thur fri sat 2019 week1 1 2 3 4 5 6 7 2020 week1 1 week2 2 3 4 5 6 7 8 0 -
Amaya
Hi, I have an another question.
How can we add dummy variables into Seasonality model?
Could you give me ideas?0
