We're using the Data Relationships Tool to perform pairwise linear regressions on a large number of variables. But at the end we're not only interested in p-values, FStat and RSq but we need to know the slopes and off-sets of all the straight line fits.
While you can write an IronPython script iterating over hundreds of scatter plots to retreive the straight line fit parameters this is too complicated and time consuming for the average user and I belive the values are actually available within the linear regression computation already.
So why not simply add slope and off-set to the output table? I believe the Data Relationships Tool runs in Spotfire's memory engine so it would be significantly faster than re-inventing the wheel by writing a TERR or Python data function for linear regression.
Implemented in | 12.0 |
Starting with Spotfire 12.0 LTS, the data relationships tool can now also include fit parameters (Slope, Intercept) when using linear regression as the comparison method.
Learn more about this new feature and everything new in this release from the what's new in Spotfire 12.0 LTS page.
Great! While this might seem a minor thing compare to all the ML stuff out there it is a very useful feature implemented very quickly, so thanks!