4 个仓库
Educational resources for building mathematical models to estimate unknown parameters using observed data and priors.
Distinct from Programming Tutorials: Focuses on the methodology of probabilistic programming rather than general software development tutorials
Explore 4 awesome GitHub repositories matching education & learning resources · Probabilistic Programming. Refine with filters or upvote what's useful.
This project is a computational statistics textbook and Bayesian data analysis course. It serves as a guide for performing statistical inference and quantifying uncertainty through a probabilistic programming workflow using Python. The resource employs a computation-first pedagogy, teaching Bayesian methods and parameter estimation through executable code and simulations instead of formal mathematical notation. It provides a practical approach to implementing Markov Chain Monte Carlo sampling to estimate posterior distributions. The content covers building probabilistic models, integrating e
Provides a comprehensive guide to performing statistical inference and uncertainty quantification using a probabilistic programming workflow.
This repository contains the digital textbook and supplementary materials for probabilistic machine learning education. It provides structured text and guided study materials covering the mathematical foundations of probability and neural networks. The project emphasizes reproducibility through a collection of interactive notebooks and standalone scripts used to recreate data plots and figures from the text. These materials are hosted in external environments to allow users to execute complex machine learning code without local installation. The educational surface includes lecture slides, e
Provides educational resources and code demonstrations for implementing probabilistic programming concepts.
该项目是一个教育性的计算笔记本和教程集合,专注于贝叶斯机器学习和概率编程。它提供了一个构建预测模型的框架,通过定义参数的概率分布而不是依赖单一的点估计来表示不确定性。 该仓库作为一个统计方法库,用于估计参数分布、执行回归以及量化预测系统中的置信水平。它涵盖了一系列技术,包括高斯过程回归、马尔可夫链蒙特卡洛(MCMC)采样和变分推理,以近似复杂的后验分布。 除了核心回归和推理外,该集合还演示了如何识别高维数据集中的潜在结构,并通过概率代理建模自动化搜索最佳模型配置。这些资源被组织为分步教程,旨在促进概率模型和不确定性量化技术的实际应用。
Offers practical tutorials on defining probability distributions over model parameters to quantify uncertainty in predictive systems.
该仓库作为贝叶斯统计建模的教育资源,提供了一系列将理论概念转化为可执行 Python 代码的教学示例。它作为一个用于执行统计推断和参数估计的计算框架,旨在帮助用户通过交互式文档学习和应用概率编程技术。 该项目利用概率编程框架将统计模型定义为有向无环图,通过高级采样算法实现自动推断。通过利用哈密顿蒙特卡洛采样和自动微分,模型探索高维概率分布以生成后验样本。该实现依赖于向量化数组计算来同时处理跨数据集的复杂数学运算。 该集合涵盖了广泛的科学数据分析任务,包括构建允许跨组信息共享的贝叶斯分层模型。这些示例组织在计算笔记本环境中,该环境将叙述性文本与代码交织在一起,以记录构建、测试和验证统计假设的迭代过程。
Provides educational resources for building mathematical models to estimate unknown parameters using observed data and priors.