Faq why PLS Toolbox have a boxplot function that conflicts with Stats Toolbox

From Eigenvector Research Documentation Wiki
Revision as of 14:11, 8 January 2019 by imported>Lyle
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

Issue:

Why does PLS_Toolbox have a boxplot function that conflicts with the Mathworks Statistics Toolbox function of the same name?

Possible Solutions:

There has been a long-standing conflict between the Eigenvector Research (EVRI) DataSet object and an identically-named object supplied in the Mathworks (MW) Statistics Toolbox. Read more on that conflict here.

However, in most versions of the MW Statistics Toolbox, boxplot is the only function which uses the MW dataset object (unlike the 65-75% of PLS_Toolbox functions which use EVRI's DataSet object.) This means that moving the MW dataset object to the top of the path would break MOST of EVRI's code, but moving EVRI's to the top of the path would only break that one function (and, unfortunately, any user code which used their dataset.)

We strive to make the two DataSet objects as compatible as possible, but there are some features of the MW dataset object which we either did not think were prudent to include or that were too invasive to the workings of our existing object. The result is that the two objects are NOT compatible.

With no other recourse, EVRI decided to make our own boxplot function to reproduce and overload the Statistics Toolbox function. Our boxplot covers a reasonable portion of the functionality provided by the Mathwork's boxplot, but not all of it. If there are any key behaviors you notice it missing, let us know and we'll add them ( helpdesk@eigenvector.com ).

As long as PLS_Toolbox is on the top of the Matlab path, our boxplot function will be used.

To Use The Mathworks' Boxplot Function:

If you wish to use the Mathworks' boxplot function instead of Eigenvector's boxplot function, you must first move PLS_Toolbox to the bottom of the path (or remove it completely), then type the command:

>> clear classes

Note that this command will clear all variables from memory, but this is the only way to switch between the Eigenvector and Mathworks' DataSet objects.


Still having problems? Please contact our helpdesk at helpdesk@eigenvector.com