ILest: Ionic Liquids Estimation

UW DIRECT Project on Ionic Liquids Machine Learning. The package contains tools and wrappers of existing Python infrastructure for data analysis (http://pandas.pydata.org/) and (http://scikit-learn.org/stable/) for the purpose of estimating desired properties of ionic liquid binary mixtures.

Contributors

Joseph Kasper, Hongbin Liu, Moke Mao, Sarah Floris

Documentation

See the examples directory

License

This software is released under the MIT license since it is a permissive free software license that has excellent compatibility with other licenses. See the LICENSE file for more information.

Code Structure

|---ionic_liquid (master)
    |---ionic_liquid
        |---__init__.py
        |---version.py
        |---main.py
        |---util.py
        |---datasets
            |---compoundSMILES.xlsx
            |---compounddata.xlsx
        |---examples
            |---Example_Workflow.ipynb
        |---method
            |---__init__.py
            |---method.py
        |---visualization
            |---__init__.py
            |---core.py
            |---plot.py
        |---test
            |---test_utils.py
            |---test_method.py
            |---test_utils.py
        |Interface.ipynb
    |---doc
        |---overview.md
        |---functional_spec.md
        |---code_struct.md
        |---tutorial.md
        |---runcell.png
        |---model_train.png
        |---model_read.png
    |---README.md
    |---LICENSE
    |---setup.py

Directory Summary

datasets contains the downloaded ionic liquids data.
methods contains the regression model used in this work.
visualization contains the plot function.
test is the folder for unit test.
Interface.ipynb is a portable entrance of the interface widgets.
doc contains documents, tutorial is also available in this directory.
LICENSE MIT license

Graphic User Interface (GUI)

The GUI provide a handy interface to choose the data set and the size of training and testing. After training the data, the regression model can be saved for the furure usage.

The GUI also provide a handy interface to load the regression model and predict the electrical conducitivity under different conditions.