Python:没有机器学习的Gridsearch? [英] Python: Gridsearch Without Machine Learning?

问题描述

我想优化一个具有多个可变参数的算法 作为输入.

I want to optimize an algorithm that has several variable parameters as input.

对于机器学习任务，Sklearn使用 gridsearch提供超参数的优化. > 功能.

For machine learning tasks, Sklearn offers the optimization of hyperparameters with the gridsearch functionality.

Python中是否存在标准化的方法/库，可以优化超参数，而不仅限于机器学习主题?

Is there a standardized way / library in Python that allows the optimization of hyperparameters that is not limited to machine learning topics?

推荐答案

您可以创建自定义管道/估算器(请参见链接

You can create a custom pipeline/estimator ( see link http://scikit-learn.org/dev/developers/contributing.html#rolling-your-own-estimator) with a score method to compare the results.

ParameterGrid 可能也对您有帮助.它将自动填充所有超参数设置.

The ParameterGrid might help you too. It will automatically populated all the hyper-parameters settings.

这篇关于Python:没有机器学习的Gridsearch?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！