Isolation_Forest """isolation _forest.ipynb

Automatically generated by Colaboratory.

Original file is located at https://colab.research.google.com/drive/1W8MVh8cTIAEHR80o1G06QM2NkVYmLSxk """

import numpy as np from tensorflow.keras.preprocessing.image import ImageDataGenerator from sklearn.ensemble import IsolationForest import tensorflow as tf import matplotlib.pyplot as plt

IMAGE_SIZE = 256 BATCH_SIZE = 32 CHANNELS = 3 EPOCHS = 50

train_data_dir = r'/content/drive/MyDrive/dataset'

dataset = tf.keras.preprocessing.image_dataset_from_directory( train_data_dir, seed=123, shuffle=True, image_size=(IMAGE_SIZE, IMAGE_SIZE), batch_size=BATCH_SIZE )

class_names = dataset.class_names

def get_dataset_partitions_tf(ds, train_split=0.8, val_split=0.1, test_split=0.1, shuffle=True, shuffle_size=10000): assert (train_split + test_split + val_split) == 1

ds_size = len(ds)

if shuffle: ds = ds.shuffle(shuffle_size, seed=12)

train_size = int(train_split * ds_size) val_size = int(val_split * ds_size)

train_ds = ds.take(train_size) val_ds = ds.skip(train_size).take(val_size) test_ds = ds.skip(train_size).skip(val_size)

return train_ds, val_ds, test_ds

train_ds, val_ds, test_ds = get_dataset_partitions_tf(dataset)

X_train = [] for image_batch, _ in train_ds: for image in image_batch: X_train.append(image.numpy().flatten()) X_train = np.array(X_train)

X_train = X_train / 255.0

isolation_forest = IsolationForest() isolation_forest.fit(X_train)

def display_predictions(images_batch, labels_batch, predictions): plt.figure(figsize=(15, 15)) for i in range(9): ax = plt.subplot(3, 3, i + 1) plt.imshow(images_batch[i].numpy().astype("uint8"))

predicted_class, confidence = predictions[i]
actual_class = class_names[labels_batch[i]]

plt.title(f"Actual: {actual_class},\n Predicted: {predicted_class}.\n Confidence: {confidence}%" if predicted_class != 'Anomaly' else "Actual: {actual_class},\n Predicted: {predicted_class}.")

plt.axis("off")

anomaly_scores = isolation_forest.decision_function(X_train) predictions = [] for i, score in enumerate(anomaly_scores): if score < -0.5: # Adjust threshold as needed predicted_class = 'Anomaly' else: predicted_class = class_names[np.argmax(score)] confidence = round(100 * np.max(score), 2) if predicted_class != 'Anomaly' else None predictions.append((predicted_class, confidence))

for images, labels in test_ds.take(1): display_predictions(images, labels, predictions) plt.show() 305685331-7c6ea458-5a10-4191-a0cf-377562b70701 Output: