[Pycon] [new paper] "Aileen Nielsen" - Modern time series analysis in Python

[info a pycon.it]

Title: Modern time series analysis in Python
Duration: 240 (includes Q&A)
Q&A Session: 0
Language: en
Type: Training

Abstract: This tutorial is an overview of intermediate topics in time series analysis most likely to be encountered in common scenarios, such as analyzing biosensor data, making business demand predictions, or protecting a website from fraudulent activity. This tutorial assumes an audience familiar with basic time series analysis and is intended as a follow-up to a 2016 PyCon tutorial introducing the fundamentals of time series analysis.

Outline & Timing:

1.     Online analysis of time series data – 30 minutes
a.     Introduction to the concept of online analysis and a bit of history – 10 minutes
b.     Online estimation of statistical properties – 10 minutes
c.     Online anomaly detection – 10 minutes

2.     Bayesian forecasting – 20 minutes
a.     Introduction to Bayesian thinking – 5 minutes
b.     How to do a Bayesian forecast and what it means – 15 minutes

3.     Markov processes – 40 minutes
a.     Introduction to Markov processes – 5 minutes
b.     Simulating Markov processes and how these simulations are used – 15 minutes
c.     Hidden Markov Models – 20 minutes

4.     Break – 10 minutes

5.     Neural networks – 90 minutes
a.     Brief overview of convolutional neural networks – 10 minutes
b.     Applying convolutional neural networks to time series classification – 30 minutes
c.     Brief overview of recursive neural networks – 15 minutes
d.     Applying recursive neural networks to time series prediction – 35 minutes

Tags: [u'#data', u'DataVisualization', u'data-science', u'#Python']