TRAMO stands for "Time series Regression with ARIMA noise, Missing values and Outliers" and SEATS for "Signal Extraction in ARIMA Time Series". These programs (which are commonly used together) were developed by Victor Gómez and Agustín Maravall at the Bank of Spain.
The home page for these programs is at www.bde.es. To quote the description given there, "The programs are fundamentally aimed at monthly or lower frequency time series. Although structured to meet the needs of an expert analyst, they can be reliably used in an entirely automatic manner. The main applications are forecasting, seasonal adjustment, trend-cycle estimation, construction of composite leading indicators, interpolation, detection and correction of outliers, estimation of special effects, and quality control of data."
To install TRAMO/SEATS for use with gretl under GNU/Linux please use the binary rpm tramo-seats-98-1.i386.rpm. (Source code for the programs is not freely available, although the authors were kind enough to give me access to the source to build this version.)
If you wish to use TRAMO/SEATS with gretl on Windows, you should download the self-installer, ts_install.exe and run it. You will be prompted for an installation directory, by default c:\userdata\tramo. This package contains the programs tramo.exe and seats.exe in win32 console form, along with examples and documentation in PDF format. I am grateful to Gianluca Caporello for permission to redistribute tramo and seats in this form.
Support for TRAMO/SEATS is new in version 1.0.3 of gretl and at present it is quite limited, but if there is interest in this feature I will probably aim to develop it further. The following pictures illustrate what you can do as of now.
After opening a suitable data set (i.e. containing seasonal time series), select a variable in the gretl main window and pull down the Variable menu. Select "TRAMO analysis".
You will then see a dialog box which gives you the choice of saving (into the current gretl data set) up to three of the series generated by SEATS, namely the seasonally adjusted version of the input series, the estimated trend/cycle component of the series, and the estimated "irregular" component. At this point you can also choose whether or not to display a graph showing these series.
If you choose to display the graph, you should see something like this:
Besides saving output series and displaying the graph, you will also see the full output from the SEATS analysis. An excerpt from this is shown below.