y*********m
发帖数: 33 | 1 Sample Problem：
I would like you to build a model that predicts the electricity price every
hour into the future for 1 week into the future for the following 1 data
streams (the resulting prediction will be a matrix of 1x168):
HOUSTON2010 NORTH2010 SOUTH2010 WEST2010 (or 11/12)
The actual real time electricity price you can find on http://www.ercot.com/content/cdr/html/real_time_spp for HB_Houston. If you have a hard time getting connected to this also check out http://www.ercot.com/mktinfo/rtm/ for HB_houston.
Limitation and restriction are:
· You can use any additional data source other than what you can find
on http://www.ercot.com/ I encourage you to improve your model quality by thinking what are some other data sources that you want to link to or use in your model building.
· I would prefer if you code this up in matlab (if you don’t have
access to matlab to do this let me know, then you can pick one of the
following: R, SAS, or C#)
· I would prefer if you make the links real-time to get the data and
predict future values (but this is not necessary)
· You can use as much or as few past observations that you wish
· You can use any modeling technique that you think is suitable
. You can use and other data stream in your model
- There are no limitation on time; however, please do not spend
excessive amount of time on it.
Results:
· I like a matlab m-file so I can see what you did
· Summary of your solution method
· Any statistics that you think would illustrate your model quality |
|
