Building Blocks of Deep Learning for Time Series
While we laid the foundations of deep learning in the previous chapter, it was very general. Deep learning is a vast field with applications in all possible domains, but the focus of this book is time series forecasting.
So, in this chapter, let’s strengthen the foundation by looking at a few building blocks of deep learning that are commonly used in time series forecasting. Even though the global machine learning models perform well in time series problems, some deep learning approaches have also shown good promise. They are a good addition to your toolset due to the flexibility they allow when modeling.
In this chapter, we will cover the following topics:
- Understanding the encoder-decoder paradigm
- Feed-forward networks
- Recurrent neural networks
- Long short-term memory (LSTM) networks
- Gated recurrent unit (GRU)
- Convolution networks
