cluster_assignments.md

Cluster assignments

Predicted cluster assignments result is saved in the cluster_assignments.txt file.
Default cluster method is using K-Means on the latent features saved in the feature.txt file.
You can use other clustering method with the latent features.

Compare cluster assignments with ground truth labels.txt

from sklearn.metrics import confusion_matrix
from scale.plot import plot_confusion_matrix

ref, classes = read_labels('../data/labels.txt') 
assignments, assign_classes('../output/cluster_assginments.txt')

cm = confusion_matrix(assignments, ref)
plot_confusion_matrix(cm, classes, assign_classes, title='', normalize=False, figsize=(3,3), show_cbar=True)