The succession of events and the way they relate to the concepts of present, past and future. Explaining the past and predicting the future is a reason for constant reflection and study from the origins of humanity. In statistics there are some objetive techniques to predicting the future. Some techniques can be classified as descriptive and others as inferences when considering the presence of a probability distribution. All of them, of course, try to minimize errors in them and all rely on some fundamental principles:

1) Have enought information about the past.

2) The information must be quantifiable in some form of associated data at regular time intervals.

3) It is assumed that there is some inertia in the studied phenomena that results in patterns that are repeated at least partially in the future.

About this material.

The material presented has been developed with free software, the code that the notebooks contain is mostly its own and made for academic and teaching purposes.

Software used during the elaboration of the content:

Jupyter notebooks	http://jupyter.org/
Python 3	https://www.python.org/download/releases/3.0/
Docker	https://hub.docker.com/r/jupyter/all-spark-notebook/

Contents

Topic	SubTopic
Time Series Stationary Models	Time Series Smoothing Methods	View notebook
Time Series Stationary Models	Stationary processes	View notebook
Time Series Stationary Models	Autoregresive Time Series	View notebook
Time Series Stationary Models	Moving Average Time Series	View notebook
Time Series Stationary Models	ARMA & ARIMA Time Series	View notebook

Go to github repository

References

Most of the theoretical content is not proper but is a compendium of classic results and examples from of diverse sources.

"Time Series Analysis Forecasting and Control"	Box, G. Jenkins, Holden Day, 1971
"Métodos de Pronósticos"	Makridakis, S.; Wheelwright. Limusa, 2000