Presented By O’Reilly and Cloudera
Make Data Work
September 11, 2018: Training & Tutorials
September 12–13, 2018: Keynotes & Sessions
New York, NY

Recurrent neural networks for time series analysis

Bruno Goncalves (Data For Science)
1:30pm–5:00pm Tuesday, 09/11/2018
Data science and machine learning
 Location: 1A 12/14 Level: Intermediate
Secondary topics:  Deep Learning, Temporal data and time-series analytics
Average rating: ***..
(3.14, 7 ratings)
Download slides (PDF)

Who is this presentation for?

  • Data scientists, data analysts, and anyone with an interest in RNN and time series

Prerequisite knowledge

  • A working knowledge of Python and Keras

Materials or downloads needed in advance

  • A laptop with Python 3, Keras, and TensorFlow installed

What you'll learn

  • Learn how to use recurrent neural networks, gated recursive units, and long short-term memory for time series analysis

Description

From the closing price of the stock market to the number of clicks per second on a web page or the sequence of venues visited by a tourist exploring a new city, time series and temporal sequences of discrete events are everywhere around us. Understanding them requires taking into account the sequence of values seen in previous steps and even long-term temporal correlations.

Join Bruno Gonçalves to learn how to use recurrent neural networks, a technique originally developed for natural language processing, to model and forecast time series. You’ll also discover the advantages and disadvantages of recurrent neural networks with respect to more traditional approaches.

Outline:

Recurrent neural networks

  • Review of feed-forward networks
  • Introduction to recursion
  • Types of recurrent neural networks
  • Your first recurrent network

Gated recurrent units

  • Advantages of recursion
  • Controlling information flow
  • Gates and internal logic

Long short-term memory

  • Remembering the past
  • Avoiding vanishing gradients
  • Memory cells

All code and slides presented during the tutorial will be made available in the course GitHub repository.

Photo of Bruno Goncalves

Bruno Goncalves

Data For Science

Bruno Gonçalves is a chief data scientist at Data For Science, working at the intersection of data science and finance. Previously, he was a data science fellow at NYU’s Center for Data Science while on leave from a tenured faculty position at Aix-Marseille Université. Since completing his PhD in the physics of complex systems in 2008, he’s been pursuing the use of data science and machine learning to study human behavior. Using large datasets from Twitter, Wikipedia, web access logs, and Yahoo! Meme, he studied how we can observe both large scale and individual human behavior in an obtrusive and widespread manner. The main applications have been to the study of computational linguistics, information diffusion, behavioral change and epidemic spreading. In 2015, he was awarded the Complex Systems Society’s 2015 Junior Scientific Award for “outstanding contributions in complex systems science” and in 2018 was named a science fellow of the Institute for Scientific Interchange in Turin, Italy.

Comments on this page are now closed.

Comments

Dave Vennergrund | DIRECTOR, DATA AND ANALYTICS
09/11/2018 12:10pm EDT

Time Series forecasting is a fascinating topic – great to see LSTM and GRU strengths and weaknesses. Nice survey – thanks.

Picture of Bruno Goncalves
Bruno Goncalves | CHIEF DATA SCIENTIST
08/28/2018 12:59pm EDT

Yes, if you want to follow along you’ll need a laptop to run the code.

Jan Haggstrom | CEO
08/28/2018 11:55am EDT

Do I need to bring a laptop to the tutorial?