Here is given the code written and used for the bachelor's report called
Accelerating the Search for Atomic Structures with Role Model Clusters
by Anders B. Bennedsgaard
models.py includes the k-Means clustering and Role Model Search, with a linear regression and a k-fold cross validation method.
maths.py includes most or all of the mathematical methods used in the project. This includes feature-calculation and gradient descent-methods.
plot_function includes most the plot-functions used.
results_rolemodel_search.py is the main file, from where most results are calculated.