Initial_TDA_exploration

UofSC workspace

Codes

V1

1. Just connecting the dots

V2.1

1. Trying to implement TDA using a non-linear Algebra approach.
2. pointInside taken from https://github.com/sasamil/PointInPolygon_Py

V2.2

1. loop v2 to get the barcode plot
2. on the data taken from the smith paper

V2.3

1. plots the bar plot and connects line for Mackey glass data

V3.1

1. Trying to build a function that tracks the number of 1-simplexs over a time-series signal.
2. Start out on the high-rate data from the air force.

V3.2

1. run a window over the Air Force data

V4.1

1. A copy of V3.1, but uses the combinatorial computational approach rather then the polygram method.

V4.2

1. builds a plot line by line, looking for the persistence

V5.1

1. Gave up on writing my own code, looking at scikit-TDA and Risper
2. https://github.com/scikit-tda/ripser.py

V5.2

1. Using example codes I got from https://persim.scikit-tda.org/en/latest/notebooks/Persistence%20images.html to map out persistence images.

V5.3.

1. Trying to merge what I learned form 5.2 and merge that into 5.1.

V5.4.

1. Copy of 5.3, but in a loop so save all the images to find one that looks cool

V5.4.1

1. Copy of V5.4, but in a loop with set axis (ax) sizes for the figure.

5.5

1. A copy of 5.3, but no plotting and just looking at timing.

6.1

1. A copy of 5.3, but trying to add more dimensions to the problem.
2. This is sed up to just give the timing for ripser to the console.
3. This as 2 embedded dimensions. So it returms H_0, H_1, and H_2.

6.2

1. A copy of 5.5
2. added maxdim to "tda_results = ripser.ripser(data,metric='manhattan',maxdim=2)".
3. added a 3rd dimension to the dataset.

6.3

1. Copy of 6.1
2. Added a 3rd dimension.

6.4

1. Copy of 6.3
2. This code is designed to run with a memory profiler.
3. I used memory_profiler from https://towardsdatascience.com/profile-memory-consumption-of-python-functions-in-a-single-line-of-code-6403101db419 and https://anaconda.org/conda-forge/memory_profiler
4. I tried to use spyder-memory-profiler but could not get it to work.
5. I created a python 3.7 environment in Anaconda, not sure that I needed to, and ran the code from the command line using ipython v6.4_processing_code.py.

7.1

1. Working on a super simple TDA with Ripser trying to understand birth and death.
2. Need to figure out if epsilon is radius or diameter.
3. Neither the birst or death really makes sense.
4. You cna get an H_2 from 2D data, found on on 500 points.