RapidMiner

Which operators or combination of operators should I use?

Regular Contributor

Which operators or combination of operators should I use?

Dear Mr. and Mrs Rapidminer, hello Community,

 

I have a set of data as follows:

[...]

01.01.2017, 1,2,3,0,4,5,6

02.01.2017, 6,2,3,0,4,5,1

03.01.2017, 5,2,3,6,4,0,1

 

...and so on.

No number in a row accours twice.

 

Now I want to analize this and maybe predict a new sequence for the future. Lets say: 04.01.2017.

I have seen some videos from Thomas Odd (from 2010) about financial forcasts and I have seen the general sequence pattern operator. But nothing seems to fit to my problem exactly.

 

Which methods would you suggest? Is this problem based in neural networks or in sequence analytics?

Do regression test make sense in such cases?

 

What would be useful practical methods/operators in RapidMiner to analyse this data over time and maybe predict something?

 

Thank you.

 

2 REPLIES
Community Manager

Re: Which operators or combination of operators should I use?

Since this is a time series, I'd use the Series Extension. You'll need to use the Sliding Window Validation operator and a Windowing operator too boot. 

Regards,
Thomas
LinkedIn: Thomas Ott
Blog: Neural Market Trends
Regular Contributor

Re: Which operators or combination of operators should I use?

Thank you T-Bone for the information!

 

I have past some statistics exams at university. But it was long ago and I forgot nearly everything. 

 

Are there any known additional resources for this kind of problems (tutorials, videos, example chains etc.)?

 

Thank you in advance.