Bgolearn 1.2.0

A Bayesian global optimization package for material design

[](https://pypi.org/project/Bgolearn/) # Python package - Bgolearn


## 为材料设计而生！ ## （ A Bayesian global optimization package for material design ）Version 1, Jul, 2022

Reference paper : V. Picheny, T. Wagner, and D. Ginsbourger. “A Benchmark of Kriging-Based Infill Criteria for Noisy Optimization”. In: Structural and Multidisciplinary Optimization 48.3 (Sept. 2013), pp. 607–626. issn: 1615-1488.

Written using Python, which is suitable for operating systems, e.g., Windows/Linux/MAC OS etc.

## Content Bgolearn guides subsequent material design based on existed experimental data. Which includes: 1.Expected Improvement algorithm, 2.Expected improvement with “plugin”，3.Augmented Expected Improvement，4.Expected Quantile Improvement，5.Reinterpolation Expected Improvement， 6.Upper confidence bound，7.Probability of Improvement，8.Predictive Entropy Search，9.Knowledge Gradient, a total of nine Utility Functions. Predictive Entropy Search，Knowledge Gradient are implemented based on Monte Carlo simulation.（贝叶斯优化设计，根据已有的实验数据对后续材料设计作出指导，本算法包共包括：期望最大化算法，期望最大化算法改进（考虑数据噪声），上确界方法，期望提升方法，熵搜索，知识梯度方法等在内的共计9种贝叶斯采样方法。其中熵搜索和知识梯度方法基于蒙特卡洛实现）

## Installing / 安装 pip install Bgolearn

## Updating / 更新 pip install –upgrade Bgolearn

## Running / 运行 ### Ref.https://github.com/Bin-Cao/Bgolearn/blob/main/Template/demo.ipynb

## Utility Function 效用函数: + 1:Expected Improvement + 2:Expected improvement with “plugin” + 3:Augmented Expected Improvement + 4:Expected Quantile Improvement + 5:Reinterpolation Expected Improvement + 6:Upper confidence bound + 7:Probability of Improvement + 8:Predictive Entropy Search + 9:Knowledge Gradient

## About / 更多 Maintained by Bin Cao. Please feel free to open issues in the Github or contact Bin Cao ([e-mail]) in case of any problems/comments/suggestions in using the code.