Welcome to EEG Familiarity Prediction Repository

This research project is mainly concerned with:

Replicate result from this previous paper to build necessary toolset to perform additional research https://pages.ucsd.edu/~desa/EEG_Reveals_Familiarity_2021.pdf
- Simulation of Accuracy Based on the Reported Value
Apply new dataset from this paper to corrobate the findings in the first paper.

File Structures

modified-scripts
- contains the modified MatLab scripts to facilitate our experiment of strictly enforcing zero shrinkage and only measuring the training set error.
- cal_shrinkage.m
  - Enforces zero shrinkage
- check_lda_train_reg_auto.m
  - Modified bias calculation
- lda_apply_prob.m
  - Modified prediction generation processes
Reports.ipynb
- A batch report that summarized the current progress of this study.
mat_preproc.py
- A preprocessing package dedicates to structuring the multi-classes raw data into a trainable dataset in a binary classification setting.
LDA_simplest_test.m
- A simplified MatLab script directly called the modified-scripts. Design to be directly callable from python. See Reports.ipynb of how it works.