KNN parameters and neighbor inspection
Inspect noncontiguous class IDs, vote fractions, nearest row indices and distances using complete deterministic input.
// Deterministic KNN contract example: noncontiguous classes and exact ties.
import aner.ml
import aner.tensor
let features: Tensor = [[0.0, 0.0], [0.0, 2.0], [2.0, 0.0],
[10.0, 10.0], [10.0, 12.0], [12.0, 10.0]]
let labels: Tensor = [[10.0], [10.0], [20.0], [30.0], [30.0], [40.0]]
let queries: Tensor = [[1.0, 0.0], [11.0, 10.0]]
let model = ml.fit_knn(features: features, labels: labels, k: 3)
let predicted = model.predict(input: queries)
let probabilities = model.predict_proba(input: queries)
let known_classes = model.classes()
let neighbors = model.neighbor_indices(input: queries)
let distances = model.neighbor_distances(input: queries)
print(tensor.rows(known_classes))
print(tensor.value(known_classes, row: 0, column: 0))
print(tensor.value(predicted, row: 0, column: 0))
print(tensor.value(predicted, row: 1, column: 0))
print(tensor.value(probabilities, row: 0, column: 0))
print(tensor.value(probabilities, row: 0, column: 1))
print(tensor.value(neighbors, row: 0, column: 0))
print(tensor.value(neighbors, row: 0, column: 1))
print(tensor.value(neighbors, row: 0, column: 2))
print(tensor.value(distances, row: 0, column: 0))
print(tensor.value(distances, row: 0, column: 1))
// Two equally distant neighbors vote for different labels; smallest ID wins.
let tied = ml.fit_knn(features: [[0.0], [2.0]], labels: [[20.0], [10.0]], k: 2)
print(tensor.value(tied.predict(input: [[1.0]]), row: 0, column: 0))
Make it your experiment.
Use Aner 0.0.5 or newer. Save this program as examples/knn_reference.aner inside a folder for your experiment. Open a terminal in that folder, then check and run the program.
aner check examples/knn_reference.aner
aner run examples/knn_reference.anerThe aner command must be on your PATH. Follow the installation guide if your terminal cannot find it.
Requires Aner 0.0.5 or newer. Uses complete synthetic tensors and the native classical ML runtime. No dataset download, credentials or external service is required. The guide documents exact expected results and all input constraints.
aner check examples/knn_reference.aner
aner run examples/knn_reference.anerBundled datasets retain their own source attribution and license notices. Example outcomes are not comparative benchmarks or evidence of clinical validity.
Dataset sources & attribution