thesis

Overview

This repository contains all the resources for my thesis on option trade classification at Karlsruhe Institute of Technology.

notes 📜	schedule ⌚	experiments 🧪	computing resources ☄️	document 🎓
See references folder. Download obsidian from obsidian.md to easily browse the notes.	Link to tasks and mile stones.	Link to weights & biases (requires login).	Link to gcp (requires login), and to bwHPC (requires login).	see releases.

Development

Set up git pre-commit hooks 🐙

Pre-commit hooks are pre-checks to avoid committing error-prone code. The tests are defined in the .pre-commit-config.yaml. Install them using:

pip install .[dev]
pre-commit install
pre-commit run --all-files

Run tests🧯

Tests can be run using tox. Just type:

tox

Acknowledgement

The authors acknowledge support by the state of Baden-Württemberg through bwHPC.

Our implementation is based on:

Gorishniy, Y., Rubachev, I., Khrulkov, V., & Babenko, A. (2021). Revisiting Deep Learning Models for Tabular Data. Advances in Neural Information Processing Systems, 34, 18932–18943.

Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A. V., & Gulin, A. (2018). CatBoost: Unbiased boosting with categorical features. Proceedings of the 32nd International Conference on Neural Information Processing Systems, 32, 6639–6649.

Rubachev, I., Alekberov, A., Gorishniy, Y., & Babenko, A. (2022). Revisiting pretraining objectives for tabular deep learning (arXiv:2207.03208). arXiv. http://arxiv.org/abs/2207.03208