mltools is a set of machine learning toolkits extracted from actual projects. Its main goal is to speed up data exploration, data extraction, data cleaning conversion, and model training, allowing machine learning engineers to focus on data analysis and model selection and evaluation. The dependent packages are mainly: numpy, pandas, sklearn, seaborn, and the statistics part has a little scipy.stats, statsmodels.stats content.
mltools has the following 4 modules
- explore module : data exploration
- feature module : feature selection
- mlcluster module : Select the best parameters and drawing for unlabeled clustering
- plot module : plotting
Download mltools package from https://github.com/danielsywang/mltools. The mltools directory under this package is the source code of the machine learning toolkits. The files in the docs directory are documents written with sphinx.
Copy the source code to the project file directory or Anaconda's \Lib\site-packages (if Anaconda is installed), and then you can use mltools .
mltools document each functional module description see Welcome to mltools 'documentation! or Welcome to mltools' documentation! , the after link is just a jump of the before link , exactly the same content.
Demonstrate how to use mltools for data exploration, data extraction, cleaning conversion, and model training.