use bayesian model selection instead of heuristics for .guess

描述

此议题没有描述。

贡献者指南

技术栈

python

领域

machine learning

议题类型

feature

难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。

4

预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。

over 1 week

活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。

stale

清晰度议题是否清楚说明期望改动、验收标准和下一步。

unclear

前置要求

Bayesian statisticsPythonbayeslite codebase

新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。

15

研究方向

The issue proposes replacing heuristic model selection with Bayesian model selection for the .guess method. First, locate the current heuristic implementation in the bayeslite source code (likely in Python). Then, research Bayesian model selection techniques such as Bayes factors or WAIC that can be applied. The implementation should modify the .guess method accordingly. Since the issue body is empty, the contributor should review any existing comments and the codebase structure to understand the current heuristic.