Book Image

Applied Deep Learning with Python

By : Alex Galea, Luis Capelo
Book Image

Applied Deep Learning with Python

By: Alex Galea, Luis Capelo

Overview of this book

Taking an approach that uses the latest developments in the Python ecosystem, you’ll first be guided through the Jupyter ecosystem, key visualization libraries and powerful data sanitization techniques before you train your first predictive model. You’ll then explore a variety of approaches to classification such as support vector networks, random decision forests and k-nearest neighbors to build on your knowledge before moving on to advanced topics. After covering classification, you’ll go on to discover ethical web scraping and interactive visualizations, which will help you professionally gather and present your analysis. Next, you’ll start building your keystone deep learning application, one that aims to predict the future price of Bitcoin based on historical public data. You’ll then be guided through a trained neural network, which will help you explore common deep learning network architectures (convolutional, recurrent, and generative adversarial networks) and deep reinforcement learning. Later, you’ll delve into model optimization and evaluation. You’ll do all this while working on a production-ready web application that combines TensorFlow and Keras to produce meaningful user-friendly results. By the end of this book, you’ll be equipped with the skills you need to tackle and develop your own real-world deep learning projects confidently and effectively.
Table of Contents (9 chapters)

Summary

In this chapter, we learned how to evaluate our model using the metrics mean squared error (MSE), squared mean squared error (RMSE), and mean averaged percentage error (MAPE). We computed the latter two metrics in a series of 19 week predictions made by our first neural network model. We then learned that it was performing well.

We also learned how to optimize a model. We looked at optimization techniques typically used to increase the performance of neural networks. Also, we implemented a number of these techniques and created a few more models to predict Bitcoin prices with different error rates.

In the next chapter, we will be turning our model into a web application that does two things: re-trains our model periodically with new data, and is able to make predictions using an HTTP API interface.