Due to recent updates, all users are required to create an Altair One account to login to the RapidMiner community. Click the Register button to create your account using the same email that you have previously used to login to the RapidMiner community. This will ensure that any previously created content will be synced to your Altair One account. Once you login, you will be asked to provide a username that identifies you to other Community users. Email us at Community with questions.

"[WRONG CALCULATION?] Linear Regression Output"

SharltoSharlto Member Posts: 3 Contributor I
edited June 2019 in Help
Hi guys,

I'm a new user of RapidMiner and a newbie in data mining.
I try to built a linear regression model and compared the result it with R.
I quite shock when I get a different result from R for Std. Error and t value.

I have change feature selection to: none and ridge to 0

Is there anybody can help me to explain it?
Am I wrong to read the output?

Below is result from RM
Attribute: AGST
Coefficient: 0.635
Std. Error: 0.031
Std. Coefficient: 0.026
Tolerance: 1
t-Stat: 20.221
p-Value: 0

Result from R
AGST
Estimate: 0.6351
Std. Error: 0.1509
t value: 4.208
Pr(>|t|): 0.000335
Tagged:

Answers

  • earmijoearmijo Member Posts: 271 Unicorn
    If you want us to help you, provide the dataset or a subset of the dataset so we can replicate the problem.
  • SharltoSharlto Member Posts: 3 Contributor I
    Hi earmijo,

    It's interesting that there is another people experiences a same thing like me but he was using SPSS
    http://www.researchgate.net/post/Why_are_the_output_values_for_simple_linear_regression_using_RapidMiner_different_from_other_software

    I think we can use any data as dataset and compare the linear regression result with another statistic software.
    I'm afraid the problem is in how RapidMiner do the calculation for Standard Error and t-Stat

    I want to attached my dataset but don't know how to attach my file.
  • MartinLiebigMartinLiebig Administrator, Moderator, Employee, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 3,529 RM Data Scientist
    could you please try out vector linear regression instead of linear regression?

    Edit: This is not returning T-test values :/
    - Sr. Director Data Solutions, Altair RapidMiner -
    Dortmund, Germany
  • SharltoSharlto Member Posts: 3 Contributor I
    Hi Martin,

    I have tried using vector linear regression as you suggested but the problem is not in the coefficient. The problem is not in the coefficient but in Std Error, T Stat and P-Value.

    Have you try to compare the output result (Std Error, T Stat and P-Value) from RapidMiner and another Statistic package?

    Anyway how to upload a dataset?
Sign In or Register to comment.