Added an Isolation Forest algorithm for anomaly detection

machine_learning/isolation_forest_for_anamoly_prediction.py

@@ -0,0 +1,69 @@
+import numpy as np
+from sklearn.datasets import make_blobs
+from sklearn.ensemble import IsolationForest
+from sklearn.model_selection import train_test_split
+
+
+def isolation_forest(features: np.ndarray, test_features: np.ndarray) -> np.ndarray:
+    """
+    This function trains an Isolation Forest algorithm and predicts anomalies.
+
+    More on Isolation Forest:
+    https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.IsolationForest.html
+    https://en.wikipedia.org/wiki/Isolation_forest
+
+    Parameters:
+    features (np.ndarray): The training data (features) on which the Isolation Forest model will be trained.
+    test_features (np.ndarray): The test data (features) to predict whether they are anomalies or not.
+
+    Returns:
+    np.ndarray: Array of predictions where -1 indicates an anomaly.
+
+    Raises:
+    ValueError: If `features` or `test_features` are not two-dimensional arrays or have mismatched feature sizes.
+
+    Examples:
+    >>> features, _ = make_blobs(n_samples=100, centers=3, cluster_std=0.60, random_state=0)
+    >>> test_features = np.array([[0, 0], [10, 10], [-1, -1]])  # Adjusted to ensure variability
+    >>> predictions = isolation_forest(features, test_features)
+    >>> np.unique(predictions)
+    array([-1,  1])  # Expecting some normal points and some anomalies
+
+    >>> isolation_forest(np.array([[1, 2]]), np.array([[1]]))  # Test for ValueError
+    Traceback (most recent call last):
+        ...
+    ValueError: `features` and `test_features` must have the same number of features.
+    """
+    # Validate input shapes
+    if features.ndim != 2 or test_features.ndim != 2:
+        raise ValueError(
+            "`features` and `test_features` must be two-dimensional arrays."
+        )
+    if features.shape[1] != test_features.shape[1]:
+        raise ValueError(
+            "`features` and `test_features` must have the same number of features."
+        )
+
+    iso_forest = IsolationForest(n_estimators=100, random_state=42)
+    iso_forest.fit(features)
+
+    predictions = iso_forest.predict(test_features)
+    return predictions
+
+
+def main() -> None:
+    """
+    Main function to demonstrate the use of Isolation Forest on a synthetic dataset.
+    """
+    features, _ = make_blobs(n_samples=100, centers=3, cluster_std=0.60, random_state=0)
+    x_train, x_test = train_test_split(features, test_size=0.2, random_state=42)
+
+    test_features = np.array([[1, 1], [5, 5], [10, 10], [6, 6], [-1, -1]])
+    predictions = isolation_forest(x_train, test_features)
+
+
+if __name__ == "__main__":
+    import doctest
+
+    doctest.testmod(verbose=True)
+    main()