Skip to content

anschmieg/Thesis_public

Repository files navigation

Thesis Project: Machine Learning in Credit Decision Modelling

How to use

Document/Document_main.pdf is the latest version of the thesis. The same is rendered as a webpage in Document_main.html, by automatic conversion - may contain errors.

script/thesis.py can be run as-is. It contains the relevant calculations for the thesis. Interactive execution as IPython cells is recommended.

The remaining files and folders are best ignored.

Repository Contents

(May be outdated)

  • Document folder contains TeX files, figures and tables. Renders to HTML using pandoc.
    • main.tex: root TeX file, includes all others.
    • Document_*: pandoc conversions of main.tex.
    • chapters: chapters to include.
    • resources: contains figures, tables, fonts, bibliography etc.
  • Script folder contains the Python script and its dependencies. prepare.py and thesis.py write to Document/resources, as these outputs are required by the TeX document.
    • thesis.py: main script.
    • *scorecard: contains the scorecard and some evaluation scripts.
    • prepare.py: prepares dataset for processing by thesis.py.
      • requires data-raw.csv
      • outputs data.parquet
    • models.py: trains the models and writes them to models/*.
    • Methods.py: outsourced functions, esp. visualization-related.
    • environment.yaml: conda environment dump.
    • models: Pickle files of the ML models and scorecard predictions. Required by thesis.py.
    • depreceated: files that are no longer required are kept here to easily revert, just in case.
  • OnlineSources folder contains PDF prints of online sources

About

Bachelor's Thesis - stripped of sensitive data

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published