Multivariable time series forecasting

cmLscmLs MemberPosts:7Contributor I
edited December 2018 inHelp

Hello all,

I am doing a time series project in rapidminer. I am trying to forecast future with multiple independent variables.

However, i could not add the independent variables to the process. When the independent variables are windowed, they all have same value.

I tried to use "MultivariateSeries2WindowExamples" and many other tools, but could not manage it.

Can somebody help me in these regard? My process and data is attached below.

Also anothor problem i encountered is to usage of optimize parameters tool in order to optimization of horizon, and windowing size and step size, but despide all the waiting RM could not respond.

I'd be very grateful if someone could provide me any information.

Thanks in advance.







futureMonths
15


horizon
1


windowSize
6



<打开ator activated="true" class="process" compatibility="6.0.002" expanded="true" name="Process">

<打开ator activated="true" class="read_excel" compatibility="8.1.000" expanded="true" height="68" name="Read Excel" width="90" x="45" y="85">

















<打开ator activated="true" class="subprocess" compatibility="8.2.000" expanded="true" height="82" name="Set Predictions_Params" width="90" x="179" y="85">

<打开ator activated="true" class="set_macro" compatibility="8.2.000" expanded="true" height="82" name="Set Window_Size" width="90" x="45" y="34">



<打开ator activated="true" class="set_macro" compatibility="8.2.000" expanded="true" height="82" name="Set Horizon" width="90" x="179" y="34">



<打开ator activated="true" class="set_macro" compatibility="8.2.000" expanded="true" height="82" name="Set Future_Years" width="90" x="313" y="34">










< portSpacing端口= " sink_out 2”间隔= " 0 " / >


<打开ator activated="true" class="set_role" compatibility="5.3.013" expanded="true" height="82" name="Set Role" width="90" x="112" y="289">




<打开ator activated="true" class="select_attributes" compatibility="8.2.000" expanded="true" height="82" name="Select Attributes" width="90" x="313" y="289">




<打开ator activated="true" class="filter_examples" compatibility="6.4.000" expanded="true" height="103" name="Filter Examples" width="90" x="447" y="85">



<打开ator activated="true" breakpoints="after" class="series:windowing" compatibility="5.2.000" expanded="true" height="82" name="Windowing for Training" width="90" x="648" y="238">

<参数键= " create_label " value = " true " / >



<打开ator activated="true" class="series:sliding_window_validation" compatibility="7.4.000" expanded="true" height="124" name="Validation" width="90" x="782" y="34">




<打开ator activated="true" class="linear_regression" compatibility="8.2.000" expanded="true" height="103" name="Linear Regression" width="90" x="112" y="85"/>







<打开ator activated="true" class="apply_model" compatibility="8.2.000" expanded="true" height="82" name="Apply Model (2)" width="90" x="112" y="34">


<打开ator activated="true" class="performance_regression" compatibility="8.2.000" expanded="true" height="82" name="Performance" width="90" x="380" y="34">














<打开ator activated="true" breakpoints="after" class="series:windowing" compatibility="5.2.000" expanded="true" height="82" name="Windowing for Application" width="90" x="849" y="493">



<打开ator activated="true" class="extract_macro" compatibility="8.2.000" expanded="true" height="68" name="Extract Example Count" width="90" x="983" y="493">



<打开ator activated="true" class="filter_example_range" compatibility="8.2.000" expanded="true" height="82" name="Filter Example Range" width="90" x="1117" y="493">



<打开ator activated="true" class="remember" compatibility="8.2.000" expanded="true" height="68" name="Remember" width="90" x="1251" y="493">


<打开ator activated="true" class="loop" compatibility="8.2.000" expanded="true" height="82" name="Loop" width="90" x="983" y="238">


<打开ator activated="true" class="recall" compatibility="8.2.000" expanded="true" height="68" name="Recall" width="90" x="45" y="136">


<打开ator activated="true" class="apply_model" compatibility="7.1.001" expanded="true" height="82" name="Apply Model" width="90" x="179" y="34">


<打开ator activated="true" class="multiply" compatibility="8.2.000" expanded="true" height="103" name="Multiply" width="90" x="447" y="30"/>
<打开ator activated="true" class="materialize_data" compatibility="8.2.000" expanded="true" height="82" name="Materialize Data (2)" width="90" x="179" y="187"/>
<打开ator activated="true" class="generate_attributes" compatibility="6.4.000" expanded="true" height="82" name="Increase Date (2)" width="90" x="380" y="187">




<打开ator activated="true" class="set_role" compatibility="5.3.013" expanded="true" height="82" name="Set Role (2)" width="90" x="179" y="340">



<打开ator activated="true" class="select_attributes" compatibility="8.2.000" expanded="true" height="82" name="Select Attributes (2)" width="90" x="313" y="340">





<打开ator activated="true" class="rename" compatibility="8.2.000" expanded="true" height="82" name="Rename" width="90" x="447" y="340">










<打开ator activated="true" class="remember" compatibility="8.2.000" expanded="true" height="68" name="Remember (2)" width="90" x="581" y="340">


















<打开ator activated="true" class="append" compatibility="8.2.000" expanded="true" height="82" name="Append" width="90" x="1184" y="136"/>
<连接from_op = "读取Excel“from_port =“输出”to_op="Set Predictions_Params" to_port="in 1"/>




















Data.csv 496B
Tagged:

Answers

  • cmLscmLs MemberPosts:7Contributor I

    Hi everyone again,

    I found similar question on this posthttps://community.www.turtlecreekpls.com/t5/RapidMiner-Studio-Forum/Prediction-with-several-attributes/m-p/47321#M30517

    However, my question still stands. After the append except for the label other variables stays same value.

    Can anyone help me on this issue.

    Thanks

  • sgenzersgenzer Administrator, Moderator, Employee, RapidMiner Certified Analyst, Community Manager, Member, University Professor, PM ModeratorPosts:2,959Community Manager

    tagging@tftemme...

  • elham_calmelham_calm MemberPosts:18Contributor I
    Hello,
    You can use free modeling of rapid miner and see path of this way...
    Best regards.
    @elham_calm
  • SGolbertSGolbert RapidMiner Certified Analyst, MemberPosts:344Unicorn

    Hi,

    As far as I know the series extension only support univariate models. You can try a windowing approach, aggregating data in each window and then applying a "normal" model like decision trees. But that's definitely different from applying multivariate time series.

    If you really need the multivariate time series approach, your option is to use one of the scripting operators. I have done so successfully with Execute R and the R package "vars".

    I hope that VAR models appear on the series extension someday!

  • MartinLiebigMartinLiebig Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University ProfessorPosts:3,404RM Data Scientist

    Hi,

    我不同意。使用窗口如聚合together with a multi-variate model like an SVM is mult-variate time series forecasting. It often also yields better results than multi-variate expansions of ARIMA.

    BR,

    Martin

    - Sr. Director Data Solutions, Altair RapidMiner -
    Dortmund, Germany
  • SGolbertSGolbert RapidMiner Certified Analyst, MemberPosts:344Unicorn

    Hi Martin,

    I don't think it's time series from a model point of view (i.e. model with regressor and noise terms). I think the time series approach could be used if the interest lies on the model itself and its coefficients. Otherwise a more black box approach with complex models should give better predictions.

    Regards,

    Sebastian

  • MartinLiebigMartinLiebig Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University ProfessorPosts:3,404RM Data Scientist

    Hi@SGolbert,

    ahh, good. so we are on the same page.

    BR,

    Martin

    - Sr. Director Data Solutions, Altair RapidMiner -
    Dortmund, Germany
Sign InorRegisterto comment.