Sometimes you need for your linear regression to be forced through the origin or any value you want. This is a very old and standard feature available in Excel, which is important as an option when running linear regressions, that not necessarily need to go through a free fitting. I am sure lots of people will find this feature useful.
Please note that Spotfire has two distinct, non-overlapping ways to ask for a linear regression, and it is not clear whether this Idea is about
(a) Scatter Plot Properties > Lines & Curves > Linear Curve Fit
or
(b) Tools > Linear Regression Modeling...
.
The following public knowledge base article shows a way to specify a y-axis intercept of zero in case (b):
https://support.tibco.com/s/article/How-to-specify-a-Spotfire-regression-model-with-no-intercept
.
The following Idea (merged with this one, among others) specifically asks for implementation of a solution for use in case (a):
---------------------
https://ideas.tibco.com/ideas/TS-I-7355
Straight line fit through origin (0,0) or with zero intercept [Merged]
Line & curves function allows to add a straight line fit. However, there is no option to force this fit line through (0,0) origin or with zero intercept. Like the way we have such option in MS Excel or SAS JMP, this isnt available in spofire.
Having such an option helps in comparing line fits/line slopes in cases where the lines are expected to start from origin. It gives the power to prove that two or more categories in a data set are have same or different trends and make a decision accordingly.
Lack of this is hindering the progress in many of our analytics projects and we had to rely on other softwares to do this as an intermittent approach.