dobriban/Topics-In-Modern-Statistical-Learning
Materials for STAT 991: Topics In Modern Statistical Learning (UPenn, 2022 Spring) - uncertainty quantification, conformal prediction, calibration, etc
When you use machine learning models, it's not enough to just get a prediction; you need to understand how certain or uncertain that prediction is. This project helps you explore and apply advanced statistical methods to quantify this uncertainty. It provides a curated set of research papers and discussions on techniques like conformal prediction, which allow you to generate reliable prediction intervals or sets, rather than just point estimates. This resource is ideal for researchers, statisticians, and data scientists who build and deploy machine learning models in critical applications where understanding prediction reliability is paramount.
176 stars. No commits in the last 6 months.
Use this if you need to reliably quantify the uncertainty of your machine learning model's predictions, ensuring they are valid even when the underlying data distribution shifts.
Not ideal if you are looking for an off-the-shelf software tool for basic machine learning model training or if you only need a single best prediction without any measure of confidence.
Stars
176
Forks
19
Language
—
License
—
Category
Last pushed
May 17, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/dobriban/Topics-In-Modern-Statistical-Learning"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
zillow/quantile-forest
Quantile Regression Forests compatible with scikit-learn.
valeman/awesome-conformal-prediction
A professionally curated list of awesome Conformal Prediction videos, tutorials, books, papers,...
yromano/cqr
Conformalized Quantile Regression
henrikbostrom/crepes
Python package for conformal prediction
xRiskLab/pearsonify
Lightweight Python package for generating classification intervals in binary classification...