kmedoids_model

Builtin index · Statistics & ML

result = kmedoids_model(X, k, [seed=])

Input: X, k, seed · Output: result

What it is

PAM clustering on the same X/k/seed= shape as kmeans_model — every one of the k centres is an actual row of X, never an average.

Overloads

One call shape — this function is not overloaded.

Code	Input	Output	Description
`result = kmedoids_model(X, k, [seed=])`	`X` —not documented yet `k` —not documented yet `seed` optionalnot documented yet	`result` Record or vector	Returns a fitted model Record whose `.predict` gives a length-`M` integer label vector for an `M`-row input.

Examples

From the documentation, and run by the build -- the output and the figure below are what the engine actually produced. Simplest first.

km = kmedians_model(G, 3, seed = 5)
kd = kmedoids_model(G, 3, seed = 5)
gm = gmm_model(G, 3, seed = 5)
print("kmedians label for row 0: {km.predict(G)[0]}")
print("kmedoids label for row 0: {kd.predict(G)[0]}")
print("gmm log-likelihood: {round(gm.log_likelihood, 1)}")

gb = gradient_boosting_model(X, y, 20, learning_rate = 0.1, max_depth = 3)
print("gradient boosting r2: {round(gb.score(X, y), 3)}")

sv = svm_model(Xc, yc, kernel = "linear", C = 1.0)
lg = logistic_model(Xc, yc)
print("svm accuracy:      {round(sv.score(Xc, yc), 3)}")
print("logistic accuracy: {round(lg.score(Xc, yc), 3)}")

svr = svr_model(X, y)
print("svr r2: {round(svr.score(X, y), 3)}")

output

kmedians label for row 0: 1
kmedoids label for row 0: 1
gmm log-likelihood: [-448.7, -448.7, -448.7, -448.7]
gradient boosting r2: 0.953
svm accuracy:      0.995
logistic accuracy: 0.995
svr r2: 0.596

from statistics-ml

How it works

kmedoids_model(X, k, [seed=]) — PAM (Partitioning Around Medoids). The property that actually distinguishes this from kmeans_model/kmedians_model above: every cluster center (medoid) is an ACTUAL row of X, never a computed average or median — see kmedoids_fit's own doc comment for the swap procedure. centers holds those real rows (so .predict works exactly like kmeans_model's nearest-centroid assignment, unchanged, since the "real point" property only matters at fit time); medoid_indices additionally exposes which row of the original X each medoid came from, for scripts (and this pass's own tests) that want to check that property directly rather than re-deriving it.

From the note above this builtin in the interpreter, where the reasoning was written down as it was built.