| Determine Data relationship [message #48718] |
Tue, 16 May 2006 10:35  |
daevu
Messages: 6
Registered: April 2006
|
Junior Member |
|
|
Folks,
How would you approach this problem
I have a dependent variable x (in this case the probability of
occurrence of a plant disease) which can be explained by a bunch of
independent variables x(i) ( in this case, weather data,eg
Tair,Precip,Wind,soilmoisture, etc).I have a large dataset of plant
disease probability and the corresponding weather data.
How would I build a model of this relationship, like y=f(x(1-i)),
linear, cubic, polynomial?
- neuronal networks would be an approach but there is no tool for it in
IDL (as far as I know)
- wavelets: could this used for it? if yes, any hints?
Thanks four your comments in advance...
D.
|
|
|
|
|
|
|
|
| Re: Determine Data relationship [message #48754 is a reply to message #48718] |
Thu, 18 May 2006 22:23  |
daevu
Messages: 6
Registered: April 2006
|
Junior Member |
|
|
wem schrieb:
> daevu wrote:
>> Folks,
>>
>> How would you approach this problem
>>
>> I have a dependent variable x (in this case the probability of
>> occurrence of a plant disease) which can be explained by a bunch of
>> independent variables y(i) ( in this case, weather data,eg
>> Tair,Precip,Wind,soilmoisture, etc).I have a large dataset of plant
>> disease probability and the corresponding weather data.
>> How would I build a model of this relationship, like y=f(x(1-i)),
>> linear, cubic, polynomial?
>> - neuronal networks would be an approach but there is no tool for it
>> in IDL (as far as I know)
>> - wavelets: could this method used for it? if yes, any hints?
>>
>> Thanks for your comments in advance...
>>
>> D.
>
> I think this depends on what you are trying to accomplish with the model.
> Are you trying to predict new, future results, or trying to visualize
> the data, or maybe guessing results from data from the past where no new
> information can be gathered about, or ...?
>
Well I would like to predict, new future results, where I will predict
the variable x.
|
|
|
|
| Re: Determine Data relationship [message #48835 is a reply to message #48718] |
Mon, 22 May 2006 07:54 |
Predictor
Messages: 1
Registered: May 2006
|
Junior Member |
|
|
I do not know IDL, but judging from this document:
http://www.ittvis.com/idl/pdfs/IDL63_FuncSum.pdf
...I'd be surprised if one couldn't cobble together a logistic
regression. If IDL isn't providing what you need, though, I'd suggest
considering another tool.
-Will Dwinnell
http://will.dwinnell.com
daevu wrote:
> How would you approach this problem
>
> I have a dependent variable x (in this case the probability of
> occurrence of a plant disease) which can be explained by a bunch of
> independent variables x(i) ( in this case, weather data,eg
> Tair,Precip,Wind,soilmoisture, etc).I have a large dataset of plant
> disease probability and the corresponding weather data.
> How would I build a model of this relationship, like y=f(x(1-i)),
> linear, cubic, polynomial?
> - neuronal networks would be an approach but there is no tool for it in
> IDL (as far as I know)
> - wavelets: could this used for it? if yes, any hints?
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
