https://github.com/ksharma67/partial-dependent-plots-and-individual-conditional-expectation-plots

Individual Conditional Expectation (ICE) plots display one line per instance that shows how the instance's prediction changes when a feature changes. The Partial Dependence Plot (PDP) for the average effect of a feature is a global method because it does not focus on specific instances, but on an overall average.
https://github.com/ksharma67/partial-dependent-plots-and-individual-conditional-expectation-plots

eda gradient-boosting individual-conditional-expectation linear-regression matplotlib numpy pandas partial-dependence-plot python seaborn sklearn xgboost

Last synced: 13 days ago
JSON representation

Individual Conditional Expectation (ICE) plots display one line per instance that shows how the instance's prediction changes when a feature changes. The Partial Dependence Plot (PDP) for the average effect of a feature is a global method because it does not focus on specific instances, but on an overall average.

• Host: GitHub
• URL: https://github.com/ksharma67/partial-dependent-plots-and-individual-conditional-expectation-plots
• Owner: ksharma67
• Created: 2022-11-08T16:18:25.000Z (about 2 years ago)
• Default Branch: main
• Last Pushed: 2022-11-18T17:03:49.000Z (about 2 years ago)
• Last Synced: 2024-11-06T12:48:59.456Z (2 months ago)
• Topics: eda, gradient-boosting, individual-conditional-expectation, linear-regression, matplotlib, numpy, pandas, partial-dependence-plot, python, seaborn, sklearn, xgboost
• Language: Jupyter Notebook
• Homepage:
• Size: 554 KB
• Stars: 0
• Watchers: 2
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Partial-Dependent-Plots-Individual-Conditional-Expectation-Plots

Python code snippets to perform the following task:

1. Modeling

Training a simple Linear Regression model

Training an advanced Gradient Boosting (XGBoost) Regression model

Evaluating both models and comparing them on the Validation Root Mean Squared Error metric.

2. Partial Dependence Plots

Generating the following PDPs (for both models):

a) For predictor/feature "Mfg_Year", which is indicative of the 'Age' of a vehicle.

b) For predictor/feature "HP", which is indicative of the (horse) power of the vehicle's engine. 

c) For predictor/feature "KM", it indicates the vehicle's (accumulated) Kilometers on the odometer

3. Individual Conditional Expectation Plots

Generating ICE Plots (for both models) on the same predictors as above. The ICE Plots were generated for 10 unique points.