Predict Your Temporal Data

Look into some specific characteristics of time series data and predict future observations based on past dynamics.

Learning objectives

At the end of this session you should be able to

• explain some basic characteristics of time series information,
• evaluate common requirements for predicting time series out of themselves, and
• use ARIMA models for short-term predictions of time series data.

Time series data

Time series data describe how a variable changes over time.

An important characteristic of time series is the autocorrelation between successive observations. This autocorrelation forms the basis for some kinds of prediction or gap filling models but it also has to be considered in time series analysis.

The above graphic shows the auto-correlation between the individual mean monthly air temperatures shown in the first graph.

The above graphic shows the partial auto-correlation between the individual mean monthly air temperatures shown in the first graph.