This repository includes assignments, notes, and calculations for assignments submitted to the Machine Learning for Biosignals Analysis course. It will be updated periodically to reflect the addition of other assignments, courses, and projects for the duration of the course.
Local branches are intended to be structured as follows:
~/.../OneDrive/.../BMME 890-014 ML
/bmme_890_yokoyama/
this repository
/machine-learning-BMME890/
fork of course materials repository
forks of other files and folders, as necessary for other assignments
This repository shall be structured as follows:
./
HW-nn/
Submissions, screenshots, code dependencies, and other files for homework # n
Notes/
Lecture notes from the course
Datasets/
Datasets downloaded for use in assignments and projects
This assignment involved the creation of a text file using a Unix console, creation and manipulation of a Git repository, and the merging of third-party edits.
/HW-01/testMe.txt - text file that was appended with entries by multiple users (Katie Heath and Keerthi Anand).
This assignment involves the manipulation of the Titanic survivors' dataset at Kaggle using Pandas, as well as a submission to Kaggle.
/HW-02/titanic-analysis.ipynb - Jupyter notebook describing processing steps for assignment 2A.
/HW-02/test-output.csv - CSV outputted dataset from assignment 2A; raw file submitted to Kaggle.
/HW-02/linAlg-recap.ipynb - Jupyter notebook with responses for tasks in assignment 2B.
This assignment revisits HW02, but by actually attempting to apply multiple classification algorithms (supervised) to identify survivors of the Titanic.
/HW-03/MLtitanic.ipynb - Jupyter notebook with steps to recreate analysis performed to create test-output.csv.
/HW-03/test-output.csv - CSV outputted dataset from assignment 3; raw file submitted to Kaggle.
The student-led lecture on principal component analysis (in collaboration with Katie Heath) is stored in this directory.
Basic concepts behind PCA is described, then demonstrated using datasets for the P300 (event-related potential)-based neurally controlled keyboard as published at http://bnci-horizon-2020.eu/database/data-sets
The P300-based paradigm is a part of the EU-funded open database for brain-machine interfaces, which was made possible by the BNCI Horizon 2020 initiative. The specific application was first published as "Attention and P300-based BCI performance in people with amyotrophic lateral sclerosis" by Andrea Riccio et al (2013, Front. Hum. Neurosci.).
/HW-99/description.pdf - PDF describing the dataset, as provided by the BNCI Horizon 2020 consortium.
/HW-99/PCA-neuro.ipynb - Jupyter notebook + Rise presentation to explain PCA and its application in the Riccio paper.
/HW-99/protocol.pptx - PowerPoint document with original geometry objects for the visual summaries of the study protocol, as seen in the Jupyter notebook. Each image is saved as protocol-1.png and protocol-2.png in order of appearance.