Questions tagged [partial-dependency-plots]
The partial-dependency-plots tag has no usage guidance.
14 questions
2
votes
1
answer
43
views
Understanding negative partial dependency values
What does a negative partial dependency value (centered/mean value), such as a change from -0.8 to -0.7 with an increase in input, indicate for a parameter?
Does this mean the parameter has a negative ...
- 23
0
votes
0
answers
30
views
Partial dependence plot from scratch - XgBoost depth 1
There are several packages that introduces enhanced explainability to a ML model like XGBoost, for example PiML. And there is a lot of other resources online about different aspects of ...
- 243
1
vote
0
answers
22
views
What is the connection of a partial regression plot and a partial dependence plot?
We have a partial regression plot as defined here: https://en.m.wikipedia.org/wiki/Partial_regression_plot
It's usefulness is discussed in What does an Added Variable Plot (Partial Regression Plot) ...
- 1,783
0
votes
0
answers
22
views
Derive elasticities from partial dependences
I am producing a partial dependences plot based on a random forests (RF) binary classification model, which yields predicted probabilty of observation $i$ belong to class A for the outcome variable Z. ...
- 1
0
votes
1
answer
59
views
Suspicious Partial Dependence Curves from a Random Forest Model
I have created several Partial Dependence Plots (PDP) for some of the top variables in my RF model. Overall, the plots seem to make sense; however, for the 'Bog" class, the probability for all ...
- 11
0
votes
0
answers
38
views
Measuring features effect and importance in Partial Least Square (PLS) regression
Context: it is possible to assess features importance and effect for a model using model-independent scoring techniques such as Partial Dependence (PD) profile, Acculumated Local Effect (ALE) profile, ...
- 101
1
vote
0
answers
35
views
Interpretation of partial dependence plot for explanatory variables with different distributions
I've fit a random forest regression model to predict a continuous landscape metric y. One of the explanatory variables in the model is continuous (Longitude), while another is categorical (Class) with ...
- 11
4
votes
1
answer
387
views
Why don't partial dependence plots match model predictions?
Background
My training is in statistics, but I'm interested in machine learning. My models involve nonlinear relationships between 2-5 predictors and a single response (all variables continuous). ...
- 20.4k
1
vote
0
answers
46
views
Is it right to have large partial dependence value?
I am trying to plot the Partial Dependence of built form against energy consumption (kwh/year) in python. As shown in the attached chart, so it returned a plot with huge partial dependence on the y-...
- 11
1
vote
1
answer
501
views
Partial dependency for random forest when predictive accuracy is bad
I have a relatively simple question which is just that if I run a random forest regression model and then use partial dependency plots to understand the influence the features have on prediction of ...
- 11
2
votes
1
answer
238
views
Partial Dependence Plots for Linear Mixed-Effects Models?
I am currently using R and am trying to figure out a way to use partial dependence plots for mixed-effects models. I have heard that PDP plots are model-agnostic, but I have never heard of any PDP for ...
- 133
4
votes
3
answers
1k
views
Understanding the behavior of a neural network when extrapolated
I trained a Multilayer Perceptron to predict a variable Y based on a set of predictors. Then I decided to test it on unseen data outside of the training range. I am aware of (some of) the implications ...
- 327
6
votes
1
answer
4k
views
Extrapolation using machine learning models under specific assumptions
I have a problem that requires inherently extrapolation. I am aware that this a crucial matter with most (if not all) machine learning models.
Yet, given the physical phenomenon underlying the ...
- 327
2
votes
0
answers
798
views
Using bootstrap confidence intervals for partial dependence plots
Core Question
Bootstrapping could be used to calculate confidence intervals (CIs) for Partial Dependence Plots (PDPs) from a random forest. But would this be appropriate?
Example
I have a dataset with ...
- 21