Forecast Modeling with different time period datas

2014 10 20
1 个回答
回答已采纳
更新时间：2014 10 21
89 次查看（30 天）

0 个投票

I want to model electricity price forecasting. However I have different time period datas. For example; I have hourly and daily datas for different variables. I do not know how to combine them in a model? Do you have any idea?

请先登录，再回答此问题。

 采纳的回答

Roger Wohlwend
Roger Wohlwend 2014-10-21

0 个投票

If you want to forecast hourly price movements then you have to ignore the daily data and work with the hourly data only. However, if you intend to forecast prices on a daily basis, use the whole dataset. For the hourly data use the latest data point of each day only. If you have the Financial toolbox, create a fints object for each time series and merge them with the function merge. In such a way it is quite easy to combine data with different frequencies.

1 个评论
显示 -1更早的评论 隐藏 -1更早的评论

Adam
Adam 2014-10-21
Hi Roger. Thank you for your reply at first. I wish to ignore all daily variables however hourly data and daily data are related each other. I have to know what is the effect each other to the model. Morever I have approx. 12 different data type and I am thinking to group them but I do not how to do it. For example I have x,y,z,t datas. I have to analysis firstly z,t data because I know its results effect to x more relaible. Than make analysis with others.
I am thinkinh of state space form of equation. For example x,y and z are the hourly datas however t is daily data. I can build a model with x,y,z and put the daily t data into the equation as a trend function in state-space form but I am really confused how to do it.

请先登录，再进行评论。

更多回答（0 个）

类别

在 帮助中心 和 File Exchange 中查找有关 Time Series Events 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by