A Deeper Dive Into Multicollinearity

[#1]

High correlation in predictor variables can harm your linear regression model. This problem is called collinearity or multi-collinearity.  The first video is an overview of collinearity, how to understand it, how to detect it, how it is related to the VIF (Variance Inflation Factor) shown in JMP reports and how to interpret VIF.  The second video shows how to use Principal Components Analysis to build models that accurately account for collinearity.  The attached JMP Journal gives you the data used in the videos, so you can try the techniques.

Identifying and Understanding the Impact of Collinearity

Video Player is loading.

Play Video

Play

Mute

Current Time 0:00

/

Duration 22:28

Loaded: 0%

00:00

Stream Type LIVE

Seek to live, currently behind liveLIVE

Remaining Time -22:28

1x

Playback Rate

Chapters

• Chapters

Descriptions

• descriptions off, selected

Captions

• captions settings, opens captions settings dialog
• captions off, selected

Audio Track

• en (Main), selected

Picture-in-PictureFullscreen

This is a modal window.

Beginning of dialog window. Escape will cancel and close the window.

Text

ColorWhiteBlackRedGreenBlueYellowMagentaCyanTransparencyOpaqueSemi-Transparent

Background

ColorBlackWhiteRedGreenBlueYellowMagentaCyanTransparencyOpaqueSemi-TransparentTransparent

Window

ColorBlackWhiteRedGreenBlueYellowMagentaCyanTransparencyTransparentSemi-TransparentOpaque

Font Size

50%75%100%125%150%175%200%300%400%

Text Edge Style

NoneRaisedDepressedUniformDropshadow

Font Family

Proportional Sans-SerifMonospace Sans-SerifProportional SerifMonospace SerifCasualScriptSmall Caps

Reset restore all settings to the default valuesDone

Close Modal Dialog

End of dialog window.

Close Modal Dialog

This is a modal window. This modal can be closed by pressing the Escape key or activating the close button.

Principal Components Analysis PCA for Modeling Multivariate Relationships

Video Player is loading.

Play Video

Play

Mute

Current Time 0:00

/

Duration 10:59

Loaded: 0%

00:00

Stream Type LIVE

Seek to live, currently behind liveLIVE

Remaining Time -10:59

1x

Playback Rate

Chapters

• Chapters

Descriptions

• descriptions off, selected

Captions

• captions settings, opens captions settings dialog
• captions off, selected

Audio Track

• en (Main), selected

Picture-in-PictureFullscreen

This is a modal window.

Beginning of dialog window. Escape will cancel and close the window.

Text

ColorWhiteBlackRedGreenBlueYellowMagentaCyanTransparencyOpaqueSemi-Transparent

Background

ColorBlackWhiteRedGreenBlueYellowMagentaCyanTransparencyOpaqueSemi-TransparentTransparent

Window

ColorBlackWhiteRedGreenBlueYellowMagentaCyanTransparencyTransparentSemi-TransparentOpaque

Font Size

50%75%100%125%150%175%200%300%400%

Text Edge Style

NoneRaisedDepressedUniformDropshadow

Font Family

Proportional Sans-SerifMonospace Sans-SerifProportional SerifMonospace SerifCasualScriptSmall Caps

Reset restore all settings to the default valuesDone

Close Modal Dialog

End of dialog window.

Close Modal Dialog

This is a modal window. This modal can be closed by pressing the Escape key or activating the close button.

• JMP Documentation on correlations and multivariate techniques
• JMP Documentation on PCA
• Four half-day public course from JMP Education Analyzing and Modeling Multidimensional Data. 
• Mastering JMP video and materials on Building Predictive Models for Correlated and High Dimensional Data

[hogi]

What determines the scale (range) on the X axis in a Leverage Plot?

[Di_Michelson]

@hogi, the scale of the axes on a leverage plot is set to the data scale after the plot is created. First, find the residuals for both models, then add the mean, then scale the axes to the data. For details, see Predictum's excellent video stored at Leverage Plots 2.mp4 on Vimeo. You might also be interested in the first video of the two-part series at Leverage Plots 1 on Vimeo. 

[hogi]

thanks a lot - very illustrative