Can this skill handle high-dimensional data?

Yes, it provides specific workflows for using penalized Cox models (Coxnet) and ensemble methods that are well-suited for high-dimensional datasets where features outnumber observations.

Is scikit-survival compatible with scikit-learn pipelines?

Yes, this skill provides implementation patterns for integrating survival estimators directly into standard scikit-learn Pipelines, GridSearchCV, and cross-validation workflows.

How does this skill evaluate model performance?

It utilizes survival-specific metrics including Harrell's and Uno's Concordance Index (C-index), Brier scores, and time-dependent AUC to account for the presence of censored observations.

What are competing risks in survival analysis?

Competing risks occur when multiple types of events can happen, and the occurrence of one prevents the others. This skill includes tools to estimate cumulative incidence in these scenarios.

What is scikit-survival used for?

It is a Python library built on top of scikit-learn specifically for survival analysis, which involves predicting the time until an event occurs while accounting for censored data.

Scikit-Survival Analyst

Name: Scikit-Survival Analyst
Author: ricable

byricable

•

データサイエンスとML

Performs comprehensive survival analysis and time-to-event modeling using the scikit-survival library in Python.

概要

This skill empowers Claude to perform advanced survival analysis, focusing on the unique challenges of censored data where event times are only partially known. It provides expert guidance on fitting Cox proportional hazards models, implementing ensemble methods like Random Survival Forests and Gradient Boosting, and training Survival Support Vector Machines. Whether you are conducting clinical research, analyzing equipment failure, or predicting customer churn, this skill offers standardized workflows for data preprocessing, non-parametric estimation, and rigorous model evaluation using specialized metrics like the concordance index and Brier score.

主な機能

Implementation of standard and penalized Cox Proportional Hazards models
Non-parametric estimation via Kaplan-Meier and Nelson-Aalen curves
Ensemble survival modeling with Random Survival Forests and Gradient Boosting
Handling of censored data and competing risks analysis
Advanced evaluation using Harrell's and Uno's Concordance Index
1 GitHub stars

ユースケース

Predicting patient outcomes and survival probabilities in medical research
Analyzing time-to-failure for predictive maintenance in industrial settings
Modeling customer retention and churn duration for subscription services

概要

主な機能

Implementation of standard and penalized Cox Proportional Hazards models
Non-parametric estimation via Kaplan-Meier and Nelson-Aalen curves
Ensemble survival modeling with Random Survival Forests and Gradient Boosting
Handling of censored data and competing risks analysis
Advanced evaluation using Harrell's and Uno's Concordance Index
1 GitHub stars

ユースケース

Predicting patient outcomes and survival probabilities in medical research
Analyzing time-to-failure for predictive maintenance in industrial settings
Modeling customer retention and churn duration for subscription services