Code and data sets for "Topological Machine Learning for Multivariate Time Series" (https://arxiv.org/abs/1911.12082)

Folder structure:

Occupancy_Test1: Contains data set and codes for Test 1 (predicting future room occupancy)

Occupancy_Test2: Contains data set and codes for Test 2 (predicting past room occupancy)

Order of running code:

1. dataCleaning.py

Python code for data preparation and cleaning.

2. pointCloud.py

Python code for converting multivariate time series to point cloud data.

3. predictionColumn.py

Python code to output a column containing room occupancy status (0 or 1) of each time window.

4. TDA_distmatrix.R

R code to calculate Wasserstein distances between persistence diagrams.

5. knn.py

Python code to run the k-nearest neighbors algorithm for binary classification.