HoWi ML is a top-level machine learning package for prototyping and comparison between different scikit-learn, MLP, LSTM and GRU model architectures. It originates from a master thesis focusing on the use of Machine Learning regression models for the oil and gas domain, found here: https://github.com/hermanwh/master-thesis
The package is published on PyPi (https://pypi.org/project/howiml/). To install, do the following:
- Install Python 3.6
- Create a new project folder
- Create a new virtual environment
- pip install howiml
Additional packages like Tensorflow, Keras etc. are automatically installed. This takes some time (approx. 5min), so be patient.
Code documentation is available in the "docs" folder
Two examples using the stateless (default) and stateful module are seen in the top-level repository ("example_stateful.ipynb" and "example_stateless.ipynb", respectively).
Some features of the package are:
- Stateless top-level module with most required functionality to define and compare machine learning regression models
- Similar, stateful top-level module for inexperienced users
- A lot of underlying functionality for more advanced users, available from howiml.utils
Usage is as follows:
- Make sure your dataset is available in .csv format, with column names in the first row and each data row in subsequent rows
- Define the required metadata for your dataset. It is suggested that you implement a local config file and import this in your project, e.g. configs.py with methods to extract all the same metadata as seen defined in the notebook examples