Kindly find the attachment of the temperature and pressure data .
I have a temperature and pressure data for a day 20140817.
I need to predict the temerature and pressure for the next day 20140818.
How to choose target data?
can anyone help me how to do it in Matlab and send me the code if have.
thanking you
Hi Manoj,
Well your problem is not so complex so i would recommend using the neural network tool box for this. As a lot of consistent training data is present for you to work with the network, weight control and optimization is not required.
Please follow the following steps-
Prepare data by normalizing it (for each variable divide the whole series by the their maximum values respectively)
Use the command 'nntool' in MATLAB to start the tool
Use the interactive tool and aim for highest regression
Then simulate the targets to get the outputs from the devised network.
I hope this helps. Any further queries are welcome.
Thanks and regards,
Atiya
Thank you Md.Belayat Hossain
I need a matlab code for implementing my temperature data.
Your data is for one day vs height .You will require data over some days to use
MA to predict or some empirical recursive relation f(time) or a mathematical
differential equation model to predict.
Cheers
To use Neural network you need . Training data . If you want to use
this single data set say for predicting temp based on height and pressure
then some of the data sequences can be used for this . Hopefully
if the fit of ANN is not too tight it will predict the temp(output)
given height , pressure..
Good luck
Cheers
Temperature and pressure that you want to predict are neural network’s targets. Your data are good just for predicting one day’s temperature (20140817) but if you want to predict whole days of year/season/ month, you should collect temperature data for some days of year/season/ month to training your network ( for example average temperature 5 days of month) .You should define some input to show characteristic of the day ( for example inputs: 1st year(for more than one year), 2nd month, 3rd day).
If you want to predict the temperature and pressure then you are left with only one input, height! Unless both temperature and pressure are a direct function of height(which I am sure they are not), you can not do this. But for the sake of discussion if it was, you only need LSQ fit between height and the desired parameter; temp or pres. There is no need for aNN.
Please see if the simple Scilab 5.3.3 code (this is open source similar to Matlab)
is useful. The main problem of what is your input and output as already pointed by others is something to be decided. by you. as well as training set /hidden nodes
etc perhaps by experimentation.
Good luck
Cheers
Use a long time line data for prediction. You may use some Classification by Decision Tree
Hi Manoj,
Well your problem is not so complex so i would recommend using the neural network tool box for this. As a lot of consistent training data is present for you to work with the network, weight control and optimization is not required.
Please follow the following steps-
Prepare data by normalizing it (for each variable divide the whole series by the their maximum values respectively)
Use the command 'nntool' in MATLAB to start the tool
Use the interactive tool and aim for highest regression
Then simulate the targets to get the outputs from the devised network.
I hope this helps. Any further queries are welcome.
Thanks and regards,
Atiya
- Badges
- Science topic