Adaptive Rejection Sampler

关于

This skill offers comprehensive procedural guidance for building Adaptive Rejection Sampling (ARS) systems, focusing on the construction of piecewise linear envelopes for log-density functions. It helps developers navigate the mathematical foundations of log-concave sampling while prioritizing performance optimization, numerical stability, and robust verification strategies. By addressing common pitfalls like infinite loops and poor initialization, the skill ensures that sampling implementations are statistically sound and computationally efficient across various environments including R and Python.

主要功能

• Procedural approaches for log-concavity verification and numerical tolerance
• Comprehensive verification strategies including statistical property testing
• Mathematical guidance for log-concave function envelope construction
• Guidance on adaptive initialization for shifted distributions and edge-case handling
• Optimization strategies for iteration limits and performance-first development
• 16 GitHub stars

使用场景

• Implementing custom statistical samplers from scratch in Python or R
• Validating the statistical integrity and quality of generated probability distribution samples
• Optimizing existing sampling algorithms to prevent timeouts and infinite loops