Added Long Short_term Memory

app.py

@@ -4,6 +4,7 @@ import matplotlib.pyplot as plt
 import json
 
 from isolation_forest import apply_isolation_forest
+from lstm import apply_lstm
 from oc_svm import apply_oc_svm
 
 # Title for Streamlit app
@@ -23,8 +24,10 @@ num_dimensions = st.selectbox('Select number of dimensions:', [1, 2, 3],
                               index=2)
 
 # Select Algorithm
-algorithm = st.selectbox('Select algorithm:', ["Isolation Forest", "One-Class Support Vector Machine"], 
-                         index=0)
+options = ["Isolation Forest", "One-Class Support Vector Machine", "Long Short-Term Memory"]
+algorithm_box = st.selectbox('Select algorithm:', 
+                             options, 
+                             index=0)
 
 with col1:
     with st.container(border=True):
@@ -65,11 +68,15 @@ if st.session_state.json_content:
     # Generating synthetic data
     rng = np.random.RandomState(42)
 
-    if algorithm == 'Isolation Forest':
+    print(algorithm_box)
+    selected_algorithm_index = options.index(algorithm_box)
+    if selected_algorithm_index == 0:
         plotted_result = apply_isolation_forest(rng, 
                                                 X)
-    else:
+    elif selected_algorithm_index == 1:
         plotted_result = apply_oc_svm(X)
+    else:
+        plotted_result = apply_lstm(X)
 
     # Create a figure
     fig, ax = plt.subplots(figsize=(10, 7), subplot_kw={'projection': '3d'} if num_dimensions == 3 else {})
@@ -89,7 +96,7 @@ if st.session_state.json_content:
         ax.set_ylabel("Heartrate" if num_dimensions > 1 else "")
         
     # Set common properties and show plot
-    ax.set_title(algorithm)
+    ax.set_title(algorithm_box)
     ax.grid(True)
     st.pyplot(fig)
 

lstm.py

@@ -0,0 +1,40 @@
+# Long Short-Term Memory system
+
+from sklearn import svm
+import numpy as np
+from sklearn.preprocessing import StandardScaler
+import tensorflow as tf
+
+def apply_lstm(all_data: np.ndarray) -> np.ndarray:
+    scaled_features = StandardScaler().fit_transform(all_data)
+
+    time_steps = 1
+    samples = scaled_features.shape[0]
+    features = scaled_features.shape[1]
+    lstm_input = scaled_features.reshape(samples, time_steps, features)
+
+    # Define the model
+    model = tf.keras.Sequential([
+        tf.keras.layers.LSTM(50, input_shape=(time_steps, features)),
+        tf.keras.layers.Dropout(0.2),
+        tf.keras.layers.Dense(features)
+    ])
+
+    # Compile the model
+    model.compile(optimizer='adam', loss='mae')
+
+    # Fit model
+    history = model.fit(lstm_input, scaled_features, epochs=50, batch_size=1, verbose=1)
+
+    # Prediction and error calculation
+    predictions = model.predict(lstm_input)
+    mse = np.mean(np.power(scaled_features - predictions, 2), axis=1)
+
+    # Define a threshold for what you consider an outlier
+    threshold = np.quantile(mse, 0.8)
+
+    # Outliers where mse is greater than the threshold
+    outliers = mse > threshold
+
+    mapped_array = np.where(outliers, -1, 1)
+    return mapped_array

oc_svm.py

@@ -11,4 +11,5 @@ def apply_oc_svm(all_data: np.ndarray) -> np.ndarray:
 
     # Initialize One-Class SVM
     oc_svm = svm.OneClassSVM(kernel='rbf', gamma='auto', nu=0.2)
+    print(oc_svm.fit_predict(X_scaled))
     return oc_svm.fit_predict(X_scaled)

requirements.txt

@@ -1,4 +1,5 @@
 streamlit
 matplotlib
 scikit-learn
-numpy
+numpy
+tensorflow