add autoencoders for dimenstionality reduction & feature extraction tutorial with tensorflow and keras

Author: x4nth055

README.md

@@ -95,6 +95,7 @@ This is a repository of all the tutorials of [The Python Code](https://www.thepy
 -[Dimensionality Reduction: Feature Extraction using Scikit-learn in Python](https://www.thepythoncode.com/article/dimensionality-reduction-using-feature-extraction-sklearn). ([code](machine-learning/dimensionality-reduction-feature-extraction))
 -[Dimensionality Reduction: Using Feature Selection in Python](https://www.thepythoncode.com/article/dimensionality-reduction-feature-selection). ([code](machine-learning/dimensionality-reduction-feature-selection))
 -[A Guide to Explainable AI Using Python](https://www.thepythoncode.com/article/explainable-ai-model-python). ([code](machine-learning/explainable-ai))
+-[Autoencoders for Dimensionality Reduction using TensorFlow in Python](https://www.thepythoncode.com/article/feature-extraction-dimensionality-reduction-autoencoders-python-keras). ([code](machine-learning/feature-extraction-autoencoders))
 
 -### [General Python Topics](https://www.thepythoncode.com/topic/general-python-topics)
 -[How to Make Facebook Messenger bot in Python](https://www.thepythoncode.com/article/make-bot-fbchat-python). ([code](general/messenger-bot))

machine-learning/feature-extraction-autoencoders/README.md

@@ -0,0 +1,4 @@
+# [Autoencoders for Dimensionality Reduction using TensorFlow in Python](https://www.thepythoncode.com/article/feature-extraction-dimensionality-reduction-autoencoders-python-keras)
+To run this:
+-`pip install -r requirements.txt`
+- Download the dataset (`gafgyt_danmini_doorbell_train.csv` and `gafgyt_danmini_doorbell_test.csv`) from [here](https://www.kaggle.com/datasets/saurabhshahane/anomaly-detection-using-deep-learning?resource=download)

machine-learning/feature-extraction-autoencoders/dimensionality_reduction_autoencoders.py

@@ -0,0 +1,47 @@
+importpandasaspd
+fromsklearn.preprocessingimportMinMaxScaler
+fromtensorflow.keras.modelsimportModel
+fromtensorflow.keras.layersimportInput, Dense, LeakyReLU
+
+
+# Reading Data
+df=pd.read_csv("gafgyt_danmini_doorbell_train.csv")
+df_test=pd.read_csv("gafgyt_danmini_doorbell_test.csv")
+# Keeping only features columns for the train set
+df_features=df.loc[:, df.columns!="target"]
+print(f"Shape of the train set: {df_features.shape}")
+y_train=df.target
+# Keeping only features for the test set
+df_features_test=df_test.loc[:, df_test.columns!="target"]
+y_test=df_test.target
+# Applying the normalization on the train then test set
+scaler=MinMaxScaler()
+df_features=scaler.fit_transform(df_features)
+df_features_test=scaler.transform(df_features_test)
+
+# Implementation of the Autoencoder Model
+# input from df_features, dense64, leakyrelu, dense32, leakyrelu, dense16, tanh 
+input=Input(shape=df_features.shape[1:])
+enc=Dense(64)(input)
+enc=LeakyReLU()(enc)
+enc=Dense(32)(enc)
+enc=LeakyReLU()(enc)
+# latent space with tanh
+latent_space=Dense(16, activation="tanh")(enc)
+
+dec=Dense(32)(latent_space)
+dec=LeakyReLU()(dec)
+dec=Dense(64)(dec)
+dec=LeakyReLU()(dec)
+
+dec=Dense(units=df_features.shape[1], activation="sigmoid")(dec)
+# init model
+autoencoder=Model(input, dec)
+# compile model
+autoencoder.compile(optimizer="adam",metrics=["mse"],loss="mse")
+# train model
+autoencoder.fit(df_features, df_features, epochs=50, batch_size=32, validation_split=0.25)
+encoder=Model(input, latent_space)
+# predict on test set
+test_au_features=encoder.predict(df_features_test)
+print(test_au_features.shape)

machine-learning/feature-extraction-autoencoders/requirements.txt

@@ -0,0 +1,3 @@
+tensorflow
+pandas
+sklearn