Tools for machine learning projects
Project description
mlpj: Tools for machine learning projects
Installation of the PyPi package:
pip install -U mlpj
Contents of this repository:
- Utilities and convenience functions for various libraries and purposes:
- python_utils: for the Python standard library
- functions for basic datatypes
- functions for filepaths and temporary files
- functions on input and output streams
- functions for printing to the console
- date functions
- numpy_utils: for
numpy
- pandas_utils: for
pandas
- functions to handle dataframe columns and their names
- functions to handle undefined values
- other dataframe convenience functions (e.g. for special cases)
- fast groupby transform of multiple columns
- functions to describe the contents of dataframes and series
- many datetime convenience functions
- e.g. add missing days to a multi-index
- stats_utils: for statistical modeling
- negative-binomial (Gamma-Poisson) distributions and overdispersion estimation
- plot_utils: for
matplotlib
- histograms
- profile plots
- timeseries_utils: for timeseries models
- ml_utils: for
sklearn
and other standard machine learning libraries- types (Protocols) for sklearn estimators and transformers
- Find an enclosed estimator or transformer within a meta-estimator.
- functions to print analyses of certain kinds of trained models
- meta-estimators and meta-transformers
- python_utils: for the Python standard library
- project_utils: project management utilities
- actions_looper: Execute selected parts of your program based on persisted results of earlier steps. Together with the functionality mentioned below, this is meant as an alternative to Jupyter notebooks that integrates more seamlessly with reuse of code and test-driven development (TDD) while still being fairly interactive.
- result_display: Collect textual and numerical results and plots on HTML pages.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
mlpj-0.3.0.tar.gz
(49.8 kB
view hashes)
Built Distribution
mlpj-0.3.0-py2.py3-none-any.whl
(44.3 kB
view hashes)