| 10 | |
| 11 | |
| 12 | def test_knn_regression(N=15): |
| 13 | np.random.seed(12345) |
| 14 | |
| 15 | i = 0 |
| 16 | while i < N: |
| 17 | N = np.random.randint(2, 100) |
| 18 | M = np.random.randint(2, 100) |
| 19 | k = np.random.randint(1, N) |
| 20 | ls = np.min([np.random.randint(1, 10), N - 1]) |
| 21 | weights = np.random.choice(["uniform", "distance"]) |
| 22 | |
| 23 | X = np.random.rand(N, M) |
| 24 | X_test = np.random.rand(N, M) |
| 25 | y = np.random.rand(N) |
| 26 | |
| 27 | knn = KNN( |
| 28 | k=k, leaf_size=ls, metric=euclidean, classifier=False, weights=weights |
| 29 | ) |
| 30 | knn.fit(X, y) |
| 31 | preds = knn.predict(X_test) |
| 32 | |
| 33 | gold = KNeighborsRegressor( |
| 34 | p=2, |
| 35 | leaf_size=ls, |
| 36 | n_neighbors=k, |
| 37 | weights=weights, |
| 38 | metric="minkowski", |
| 39 | algorithm="ball_tree", |
| 40 | ) |
| 41 | gold.fit(X, y) |
| 42 | gold_preds = gold.predict(X_test) |
| 43 | |
| 44 | for mine, theirs in zip(preds, gold_preds): |
| 45 | np.testing.assert_almost_equal(mine, theirs) |
| 46 | print("PASSED") |
| 47 | i += 1 |
| 48 | |
| 49 | |
| 50 | def test_knn_clf(N=15): |