mlr3proba: an R package for machine learning in survival analysis.

Area of Science:

• Computational statistics
• Machine learning
• Bioinformatics

Background:

• Machine learning (ML) is widely used, necessitating robust implementation interfaces.
• Existing R libraries offer limited support for survival analysis, a critical method in many disciplines.
• Survival analysis is vital in medicine, bioinformatics, economics, and engineering.

Purpose of the Study:

• To introduce mlr3proba, a novel R package.
• To provide a comprehensive machine learning interface specifically for survival analysis.
• To integrate survival analysis capabilities with existing mlr3 infrastructure.

Main Methods:

• Development of the mlr3proba package in R.
• Integration with mlr3's general model tuning and benchmarking facilities.
• Focus on providing a systematic infrastructure for survival modeling.

Main Results:

• mlr3proba offers extended support for survival analysis within the mlr3 ecosystem.
• The package facilitates comprehensive machine learning model implementation for survival data.
• It enables systematic evaluation and benchmarking of survival models.

Conclusions:

• mlr3proba addresses the need for specialized survival analysis tools in machine learning.
• The package enhances the capabilities for survival modeling and evaluation.
• It provides a valuable infrastructure for researchers in fields utilizing survival analysis.