Sorry for bringing this up again.
I created a dummy-variable-trap tag yesterday. As I wrote, I was doubtful that synonymizing it to categorical-encoding is helpful. By now I am not doubtful any more: I believe it is not helpful. Unfortunately, the synonym has already been established.
Take a look at questions about the dummy variable trap. I have already retagged some of them dummy-variable-trap. Mouse over the tag at one of those questions. Picture a user who is hazy enough about the trap to ask a question here. Will the tag wiki excerpt for dummy-variable-trap (which now is the same as for the synonym categorical-encoding) be helpful for them? Here it is:
Representing categorical variables as sets of numerical variables. Necessary in many types of analysis for them to process categorical data. A common example is using a categorical predictor in regression/ANOVA via dummy coding, effect coding, Helmert coding, user-defined contrasts, etc.
I don't think this will help the user in any way. In particular, the links to questions that actually explained what the trap is, why it occurs, what the consequences are, and how to avoid it, which I included in the dummy-variable-trap tag wiki are of course now gone that it has been synonymized.
Yes, the dummy-variable-trap occurs only in categorical-encoding. The latter is a necessary condition for the former. Does that mean that the two are synonyms? No. Does that mean that questions with the latter tag are usually helpful for someone asking about the former? I don't think so.
I propose de-synonymizing the two tags again. I can reconstruct the tag wiki and excerpt for dummy-variable-trap.
categorical-encoding
are useful for someone asking about the dummy variable trap. Simply looking through the questions currently tagged ascategorical-encoding
will show that this is the case. $\endgroup$