Huston Data for predict temperature within day

michel7

Hello!

I want to predict the temperature within a day. E.g. based on the last 7 hours predict the temperature in 2 hours.  For the training I want to use the daily data. That means I have over 45,000 examples for training.

Could you give me some suggestions on how to model this?

I know it doesn't make much sense, it's just a matter of applying the technique.

Thank you very much in advance,

Michel

MartinLiebig

Hi,

did you check the forecasting videos on academy? https://academy.rapidminer.com/catalog?query=forecasting

Best,

Martin

michel7

Hi Martin,

yes of course, several times. But i could not derive of  them my special case. Once again: I want to predict within a day , based on hourly data and  teherefore I have so many time series examples as days.

Warm regards,

Michel

MartinLiebig

Hi,

wait, so you got daily data and want to do hourly forecasts? How should this work?

Best,

Martin

michel7

Hi,

1. I have only one day data, for every minute one in my case. I can use the time series prediction based on data of a certain day . No problem. E.G based on the data of the last 15 minutes forcasting  for the +5 minute.
2. But I have several days with data (based on minutes) and  the question is: How could use all those time series data for better predicting ?

Warm regards,

Michel

michel7

Hi,

of course I have (as in the Houston example) more data per minute

1) air pressure

2) amount of precipitation

3) wind speed

4) humidity

5) wind direction

Regards,

Michel