As a general rule who + BE and which + BE can be omitted from so-called restrictive relative clauses.
- Those who are happy with the situation ...
- Those happy with the situation ...
- Those who are thinking of leaving ....
- Those thinking of leaving ...
- Those who are regularly forgotten ...
- Those regularly for gotten ...
- These who have been dropped ...
- Those dropped ...
In the examples above we see the relative pronoun who omitted from sentences with BE: as the only verb; as part of the present continuous; as part of a passive construction, and as part of a present perfect passive construction.
Now whether the examples without a relative pronoun are actually relative clauses is a matter of some debate. Some people would regard them as post-modifying adjective/participle clauses. However, this nonetheless effectively boils down to the fact that you can always omit who/which and BE from the comparable relative clauses.
Those, when used on its own like this, is classified as a pronoun. Many modern grammars treat pronouns as a subset of the Noun category. So in this sense it is a noun. Those with a following noun, however, is usually classified as a determiner. So in the following it would be classified as a determiner:
- Those people who are satisfied with the situation ...
This of course is a rather silly situation where we have to treat the same word with ostensibly the same meaning and the same grammar as a different word just depending on what word follows it. Some forward thinking modern grammarians see those as the same kind of word in both instances.
I have occasionally seen it written that pronouns cannot serve as antecedents for relative clauses. However, this only ever really applies to accusative case pronouns: me, him, her, us, them. Even then this is more a case of awkwardness or style, perhaps, rather than ungrammaticality:
- # May heaven forgive him who steals for his children. (debatable grammaticality)
Notice, though, that this does not apply to those - which does not inflect for case (and could be argued, therefore, to not be accusative).