全部版块 我的主页
论坛 数据科学与人工智能 数据分析与数据科学 SAS专版
1249 3
2012-08-13
I was trying to compare the perfomance of different smoothing methods in the time series forecasting system (TSFS). However, I could not select the Seasonal soothing, Winter method-additive, and Winter method-multiplicative for the series I was investigating. These three methods were all grey out, I could only select the other methods.

When I tried fit models from list, I got the following notice: "error: model fitting failed. this model is not allowed for this series."

I knew that all the observations in the series should be strictly positive if I want to fit a seasonal smoothing model, and they are all positive. Anyone know what the problem is.
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

全部回复
2012-8-14 09:15:43
waiting for answer!
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

2012-8-17 10:52:41
I solve the problem by myself finally, let share with you.
For seasonal data, the date variable should contain NO gaps. That is, when you get some data from the internet, it always provides those available data with the corresponding date value. For instance, the available data provided in stock market is only in week day, and you want to perform smoothing method on daily basis, then you have to add those date value covering the weekends with a corresponding reasonable stock price before using TSFS.
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

2012-8-31 07:56:06
You've already solved your problem.
By the way why don't you use Eviews to deal with time series data ?  
I guess Eviews is one of the best specialized program to deal with time series data.
二维码

扫码加我 拉你入群

请注明:姓名-公司-职位

以便审核进群资格,未注明则拒绝

栏目导航
热门文章
推荐文章

说点什么

分享

扫码加好友,拉您进群
各岗位、行业、专业交流群