Classification results using isolated added noise to calculate features

Threshold ρ_T

(Left) Confusion matrix showing the number of correct predictions (diagonal elements, green) and incorrect predictions (off diagonal elements, pink). Hover over boxes to see the precise number (out of 250). Click on the boxes to bring up the associated Betti curves (Right). Betti curves shown for dimensions 1-4. Drag the threshold slider to see results for different threshold values ρ_T. Current threshold value shown as vertical dotted line in the Betti curve plots. NOTE: the dashed line and confusion matrix correspond to the indicated threshold ρ_T, but by construction the isolated added noise network will have persistent homology calculated until density 1-ρ_T.