A small Python library which implements an outlier detector based on persistent homology.

You can import the module as

import mknn

This will give you access to the main class, Filtration. If you have already parsed your data into an array of shape (N,d), where N is the number of points and d the dimension of the ambient space, then you can construct a Filtration instance like so

filt = mknn.Filtration(data = data)

You can find an example case here. This file contains 42 points in two dimensions. The dataset has two main clusters and two outliers,

.

The two outliers are the points labeled 1 and 2.

To run the detector on this set, first you should store the data in a numpy array as follows

import numpy as np
data = np.genfromtxt("example_data.csv")

Then, you build a Filtration instance:

import mknn
filt = mknn.Filtration(data = data)

Next, instruct the Filtration instance to compute its persistent homology

filt.build_complex()
filt.compute_persistent_homology()

Finally, you can see the persistence plot by calling plot_persistence on filt, which takes a string that will be used as the name of the file in which the plot will be stored.

filt.plot_persistence("persistence_example.png")

.