FeatureValueImpactPlot

FeatureValueImpactPlot[model,data]

plots the impact of the value of a given feature in data on the result of model.

FeatureValueImpactPlot[model]

estimates the feature value impact using synthetic data.

FeatureValueImpactPlot[modelfname,]

plots only the impact of the specified feature fname.

FeatureValueImpactPlot[modelfnameclass,]

plots only the impact on the classification class.

Details and Options

Examples

open allclose all

Basic Examples  (2)

Train a linear regression on a sample dataset:

Visualize the impact of the input feature on the average model result:

Compare the plot with the numerical deviation between the average and the prediction:

Train a classifier to identify an Iris species from some features of the plant:

Visualize how a feature value affects the odds of each class:

Scope  (4)

Visualize the cumulative impacts on a prediction:

Visualize the cumulative impacts on a classification:

Use synthetic data to compute the impacts:

Visualize the impact on a specific classification:

Visualize the impact on a list of classes:

Options  (12)

AspectRatio  (1)

Choose the ratio of height to width from the actual plot values:

AxesLabel  (1)

Specify a label for each axis:

AxesStyle  (1)

Change the style for the axes:

Frame  (2)

Draw a frame around the plot:

Draw a frame on the left and right edges:

FrameLabel  (2)

Place a label along the bottom frame of a plot:

Place labels on each of the edges in the frame:

Filling  (1)

Use symbolic or explicit values for "stem" filling:

MaxPlotPoints  (1)

PlotLabel  (1)

The plots are labeled by default:

Turn off the labeling:

PlotLegends  (1)

By default, a legend based on the output class is used:

Use no legend:

Create a legend with placeholder text:

Create a legend with specific labels:

PlotStyle  (1)

By default, different styles are chosen for multiple curves:

Explicitly specify the style for different curves:

Applications  (1)

Analyze how different models are affected by feature variations:

Train a linear model to predict a value given by the nonlinear combination of two features:

As expected, the impact of each feature depends only on that feature value:

Train a nonlinear model on the same data:

Now the same feature value can have different impacts depending on the other feature in the example:

Wolfram Research (2022), FeatureValueImpactPlot, Wolfram Language function, https://reference.wolfram.com/language/ref/FeatureValueImpactPlot.html.

Text

Wolfram Research (2022), FeatureValueImpactPlot, Wolfram Language function, https://reference.wolfram.com/language/ref/FeatureValueImpactPlot.html.

CMS

Wolfram Language. 2022. "FeatureValueImpactPlot." Wolfram Language & System Documentation Center. Wolfram Research. https://reference.wolfram.com/language/ref/FeatureValueImpactPlot.html.

APA

Wolfram Language. (2022). FeatureValueImpactPlot. Wolfram Language & System Documentation Center. Retrieved from https://reference.wolfram.com/language/ref/FeatureValueImpactPlot.html

BibTeX

@misc{reference.wolfram_2024_featurevalueimpactplot, author="Wolfram Research", title="{FeatureValueImpactPlot}", year="2022", howpublished="\url{https://reference.wolfram.com/language/ref/FeatureValueImpactPlot.html}", note=[Accessed: 15-October-2024 ]}

BibLaTeX

@online{reference.wolfram_2024_featurevalueimpactplot, organization={Wolfram Research}, title={FeatureValueImpactPlot}, year={2022}, url={https://reference.wolfram.com/language/ref/FeatureValueImpactPlot.html}, note=[Accessed: 15-October-2024 ]}