Time Series Analysis

Time series is an ordered sequence of data points spread over a period of time. Here, time is generally an independent variable while the other variable/s keep changing values. The time series data is monitored over constant temporal intervals. This data can be in any measurable and quantifiable parameter related to the field of business, science, finance, etc.


TS Contributor
If you can do anything groundbreaking in time series you are an extremely smart person. Most statisticians don't understand time series as far as I can determine and that is scary. Ignoring all the practical problems.
Has much been done since Box and Jenkins?


Fortran must die
Has much been done since Box and Jenkins?
A great deal. Box-Jenkins (ARIMA) is not well designed for multivariate approaches. In theory you can do it (ARIMAX) but the need to prewhiten each series is painful. Vector Autoregressive models (VAR) and their cousin vector error correction models (VECM) (which requires cointegrantion) are important in economics. However, they are not at all easy to interpret and have huge number of parameters so only a few variables can be used. Auto Regressive Distributed Lag models are probably the easiest to use, some forms do not require the variables to be integrated of the same order (a major issue). There is also regression with ARIMA error - which I know only in theory. It appears simple to use, but I have not read it extensively.

There are also ARCH/GARCH which are used with stocks a lot I think, again I know relatively very little about them. State Space models (exponential smoothing is one early form) are also used and physicists have their own very different models of which I know just names. And there are neural networks which is big, but which I know nothing... :p

I suspect VAR and VECM are used the most by academics because I think it is economists who use time series the most. My guess is box jenkins is not used much in academics although its impossible to know of course (I base this on my readings in various fields).
Last edited: