🎉 🎉. RAPIDMINER 9.8 IS OUT!!! 🎉 🎉

RapidMiner 9.8 continues to innovate in data science collaboration, connectivity and governance

CLICK HERE TO DOWNLOAD

[New Extension] How to do Time Series Forecasting with just 1 Operator!

mschmitzmschmitz Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 2,685  RM Data Scientist
edited July 9 in Knowledge Base

Today I released a new extension: Forecasting

Forecasting is targeted to make the forecasting of time series easy. It provides two operators, which forecast the next steps of a time series. The extension is designed to remove the complexity and ensuring good practices in hyper parameter tuning and validation.

The operator “Forecast Univariate” allows you to use statistical methods like ARIMA, Holt-Winters and Functional Seasonal Decomposition forecasting with one simple operator. A standard analysis of a data set looks like this:

 

As you see we only provide the data set to be forecasted and all the rest is done in this operator. The operator provides three results.

The forecast of the next n rows:


The performance of the operation validated using Sliding Window Validation:


And the back tested data, so that you can easily compare label and forecast.

 [

The same interface and results are returned by the new operator “Forecast (Multivariate)”. Contrary to the univariate case it can use additional attributes as depended variables. This allows you to use for example the diesel price to forecast the gas price and vice versa. All the complexity of Feature Aggregation, Optimization of Hyperparameters and Validation is wrapped into the operator. You can just use it out of the box!

Currently the operator uses Decision Trees and Linear Models to forecast the future. Other machine learning models are added shortly.

 

Custom Operators

This extension is based on the custom operators’ extension. Custom operators are operators which created by using other operators. You can easily turn any workflow into an operator and ship it to your colleagues. This also means you can investigate our new operators and see what we are doing!


- Head of Data Science Services at RapidMiner -
Dortmund, Germany
David_Ajacobcybulskihbajpaisara20yyhuangPavithra_RaolionelderkrikorLeMarcwufutura

Comments

  • websiteguywebsiteguy Member Posts: 24 Contributor II
    edited August 7
    @mschmitz - am getting ??? for results - Edit, sorry think its date formatting issue, will investigate (I enabled log) gave me a clue. 
  • mschmitzmschmitz Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 2,685  RM Data Scientist
    i think you expirience 2 effects here:

    First Data Points cannot be predicted
    The first data points simply cannot be predicted, because there is no proceeding data. We still show the "original data", but we have no forecast. If you scroll down, you should start to see forecasts, don't you?

    Unequal Time Stamps:
    If your data as time stamp which are not-equalized (like 5AM, 7AM, 9AM), then we opt to replace your date attribute with and simple id. This then causes the use of those ids fore Reference and Forecast ID. This happens also if you have monthly or yearly data, since months have a different length and years can have leap years.

    Best,
    Martin

    - Head of Data Science Services at RapidMiner -
    Dortmund, Germany
  • websiteguywebsiteguy Member Posts: 24 Contributor II
    mschmitz  Hi - getting this error, not sure if i mangled something, i did delete and reinstall all exstensions did not fix... was working before last update i think??
  • mschmitzmschmitz Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 2,685  RM Data Scientist
    Hi,
    this looks like your time series attribute is not available?

    Is it possible to share data and process so i can debug?

    Best,
    Martin
    - Head of Data Science Services at RapidMiner -
    Dortmund, Germany
  • mschmitzmschmitz Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 2,685  RM Data Scientist
    i figured out what likely causes this error. If your time series attribute is not of role regular, than this happens. Please set it to regular. I will release an update to the extension later this week which covers for this.

    Best,
    Martin
    - Head of Data Science Services at RapidMiner -
    Dortmund, Germany
  • websiteguywebsiteguy Member Posts: 24 Contributor II
    @mschmitz thanks for reply. And heads up.. I think it may also have something to do with my trying to convert a UNIX time stamp back in to another format of date, and mangling it in the process.
    mschmitz
  • SkyTraderSkyTrader Member Posts: 88 Contributor I
    Hi Martin @mschmitz, how do we access the new extension? I can't see it in search or in the marketplace.

    Thanks. 
  • mschmitzmschmitz Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 2,685  RM Data Scientist
    - Head of Data Science Services at RapidMiner -
    Dortmund, Germany
  • SkyTraderSkyTrader Member Posts: 88 Contributor I
    Thanks for the link Martin, @mschmitz

    I also got this operator warning - pls see image:



    So I went back to the marketplace and downloaded a suggested operator. Now when I double click on the Forecast univariate operator I don't see this sub process?

    I then attached the process to a Retrieve operator with my usual Dow Jones Excel file and after 15 mins there was no result. So I downloaded just a basic 100 row FTSE file and after 8 mins it's still processing after 10 mins. Is this normal?

    Cheers,
Sign In or Register to comment.