Analysis and normalization of instantaneous data

student_compute

Hello

I am involved with my exams. Therefore, my time is small.

But I've been looking at your offer all afternoon and running on my arima data. And I chose different values ​​for p, q. But the prediction is not right. I've got a hundred. I entered 80 data as arima input.

And I decided. Twenty next data to predict. But the results are very different to the 20 data I already have. My data is given in this twenty first, ascending ascending and then rising again. But in the twenty output data generated by the arima forecast, the data is declining.

Is Arima not my data prediction solution? Do you have any other suggestion?

thanks for your help

hughesfleming68

@student_compute. If you are having difficulty learning the material on your own then your best option is to get some extra tutoring form your professor. Your problems with Rapidminer come from a lack of understanding of the basics. You have to put in the time. It is possible that your forecast is the best that you can do with your data. Predicting the future by looking at the past does not always workout the way we would like.

student_compute

Hello

Thank you very much for all your help.

sorry for taking your time.

Unfortunately my teacher does not help.

And I inevitably raised the problem with the RapidMiner forum. Because there really is no one to help.

Anyway, thank you again

[Deleted User]

@student_compute
Hi
I suggest you to search in community you can find the answer of your question easily. 
I think your training data is different with the test. 
mbs