Class Central is learner-supported. When you buy through links on our site, we may earn an affiliate commission.

Time Series Forecasting with GRUs

Go to class Write review

Overview

This course explores Gated Recurrent Units (GRUs) in TensorFlow for multivariate time series forecasting. We will build, evaluate, and apply advanced GRU techniques like Bidirectional GRUs, Attention Mechanisms, and Hybrid GRU-CNN models to improve forecasting accuracy.

Syllabus

Unit 1: Building & Training GRUs for Multivariate Forecasting

Experimenting with GRU Model Architecture
Adding the Final Output Layer
Compiling GRU Models for Training
Training GRUs

Unit 2: Evaluating GRU Model Performance

Calculating RMSE for Model Evaluation
Expanding Model Evaluation with MAE R2
Calculating Adjusted R² for GRU Models
Visualizing GRU Model Learning Curves
Visualizing Predictions Against Actual Values
Comparing GRU and RNN Performance Metrics

Unit 3: Advanced GRU Techniques for Time Series Forecasting

Bidirectional GRUs for Better Forecasting
Enhancing GRUs with Attention Mechanism
Building a Bidirectional Attention GRU Model

Unit 4: Hybrid GRU Models for Enhanced Time Series Forecasting

Building CNN Feature Extraction for GRUs
Implementing GRU for CNN Feature Processing
Building Complete CNN GRU Hybrid Model

Reviews

Start your review of Time Series Forecasting with GRUs