README.md 1.03 KB
Newer Older
Vincent Guigue's avatar
Vincent Guigue committed
1
2
3
4
5
6
7
8
9
10
11
# TimeSeries prediction (& analysis) Tutorial

## Environments

We are going to use basic python scientific libraries. For begginers, do not hesitate to rely on integrated environnment like Anaconda:
https://www.anaconda.com

All support are provided in python notebook formats (jupyter).


## Prerequirements
Vincent Guigue's avatar
Vincent Guigue committed
12

Vincent Guigue's avatar
Vincent Guigue committed
13
This tutorial requires some basic knowledge in ***python*** and some practice on ***numpy/matplotlib***. If needed, you can refer to 
Vincent Guigue's avatar
Vincent Guigue committed
14
15
16
17

* http://mapsi.lip6.fr/pmwiki.php?n=Main.TutoPythonJN
(resource in French) 
* or to the original numpy tutorial in English
Vincent Guigue's avatar
Vincent Guigue committed
18
https://numpy.org/doc/stable/user/quickstart.html
Vincent Guigue's avatar
Vincent Guigue committed
19
20


Vincent Guigue's avatar
Vincent Guigue committed
21
## Organization
Vincent Guigue's avatar
Vincent Guigue committed
22
23
24
25
26
27
28
29
30
31
32

This tutorial is divided two talks

* Autoregressive approaches, machine learning formalism, pitfalls
* Deep learning architectures for timeseries analysis

And 3 notebooks

* Autoregressive models
* ML & pitfalls
* Deep learning tutorial
Vincent Guigue's avatar
Vincent Guigue committed
33
34
35
36
37

## More deep learning resources

* to adapt numpy skills to pytorch: https://github.com/wkentaro/pytorch-for-numpy-users
* to become really skilly with pytorch (in French):