HANDS-ON TUTORIAL

Advanced Time Series Anomay Detection

with

LSTM AutoEncoder + CNN

Incl. Autoencoder, Isolation Forest Anomaly Detection & Time Series Analysis & FULL Python Code

Presented by Dr. Maryam Miradi

Online | Pre-recorded | One-hour

RESERVE MY SPOT

Hosted by:

Dr. Maryam Miradi

About Dr. Maryam Miradi

Maryam Miradi is the CEO and Chief AI Scientist of Profound Analytics.

She has over 20 years of experience in AI model development, holds a PhD in AI, has published 24 articles, authored 2 books, received 5 awards, and has experience coaching over 200 data scientists across 12 industries.

She has been awarded as the Best European Researcher in Future Vision.

📊Time Series Analysis Techniques

Includes using StatsModel for visualization, decomposition into trend, seasonality, and residuals, and checking data stationarity with Rolling Statistics and the Augmented Dickey-Fuller test.

🛠️ Deep Learning Autoencoder Anomaly Detection, LSTM Autoencoder

Utilizing Keras Autoencoder, including encoder and decoder structures, building a sequential model with Tensorflow Keras, using the sigmoid function, and calculating the reconstruction error.
Discusses LSTM Autoencoder with relu activation function and adam optimizer, evaluating performance using mean square error loss function, and comparing precision and recall metrics.

📈 CNN Feature Extraction & Model Comparison

Comparing Isolation Forest, AutoEncoder, LSTM Autoencoder, and LSTM Autoencoder with CNN Feature Extraction, highlighting the use of Max Pooling in CNN and deep learning models' comparison using AUC PR due to imbalanced data.

🔑 Key Techniques:

Decomposition into trend, seasonality, and residuals, and checking data stationarity with Rolling Statistics and the Augmented Dickey-Fuller test.
Autoencoder, including encoder and decoder structures, building a sequential model with Tensorflow Keras
LSTM Autoencoder with relu activation function and adam optimizer
Comparing Isolation Forest, AutoEncoder, LSTM Autoencoder, and LSTM Autoencoder with CNN Feature Extraction, highlighting the use of Max Pooling in CNN

Sign Up For Hands-On Tutorial Today

If you're looking for a unique end-to-end data science project to enhance your portfolio and want to have All Machine Learning Models explained, this Tutorial is perfect for you. Join me as I guide you step-by-step through Time Series Analysis & Time Series Anomaly Detection. using Autoencoder, LSTM Autoencoder, LSTM Autoencoder with CNN feature extraction in and Isolation Forest.