https://github.com/jbolns/atheoretical
Data analysis > VBA-driven automated regression (supervised ML) in Excel.
https://github.com/jbolns/atheoretical
machine-learning supervised-machine-learning vba vba-excel
Last synced: about 1 month ago
JSON representation
Data analysis > VBA-driven automated regression (supervised ML) in Excel.
- Host: GitHub
- URL: https://github.com/jbolns/atheoretical
- Owner: jbolns
- Created: 2022-09-28T13:50:52.000Z (over 2 years ago)
- Default Branch: master
- Last Pushed: 2022-10-06T18:03:57.000Z (over 2 years ago)
- Last Synced: 2025-01-18T14:46:21.488Z (3 months ago)
- Topics: machine-learning, supervised-machine-learning, vba, vba-excel
- Language: VBA
- Homepage:
- Size: 386 KB
- Stars: 1
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Atheoretical
A set of VBA scripts to rapidly produce supervised ML analyses in Excel.
Helpful when you run need to run regressions/supervised ML in Excel on a periodical basis and you want to save yourself some time.
Very early stages. Expect bugs. Loads of them.
### Notes
- The *mainFiles* folder contains an Excel document with clear guidance, preloaded scripts, and pre-populated (dummy) hyper-parameters.
- The *mainFiles* folder also contains a (zipped) dataset that can be used for testing (due credits given in Excel file).
- The individual modules/scripts are given as *.bas* files, which can be imported easily into Excel's VBA console.
- Eventually, the file should be able to handle any *clean* dataset made of quantitative continuous variables. It will take a while until things run smoothly, though.
### Pending:
- [ ] Documentation. The Excel file includes guidance, but more detailed documentation would not hurt.
- [ ] Stress testing. The file currently works fully with the sample dataset, but bugs are incredibly likely.
- [ ] Code cleaning. There are still macros that use repeated code sequences.
## Usage/Licensing
I honestly have not thought about this just yet. I *really* do not think the files are yet at production or even pilot-testing level. So, for now, let's say it's all rights reserved.