Whoever thought of the idea is a genius.
I can’t see anything to be gained by analysing the OP’s example using the Reed-Kellogg system. It has never been used in the UK, and it is just about redundant elsewhere, so why waste time with it?
Like your tree diagram (and my own tree diagram in this thread, in case you missed it), the Reed-Kellogg diagrams make claims about syntactic structure. Two different sets of authors, who wrote long after Reed and Kellogg, have proposed two different Reed-Kellogg structures for this type of syntactic entity, whatever name we give it. Both of those Reed-Kellogg diagrams have been given in this thread.
All the diagrams given in this thread, including yours, represent
different ways of conceiving of the nature of the construction, and clearly depict those ways of conceiving it (provided one understands the diagramming system). If the great Otto Jespersen had drawn some sort of diagram of the sentence, whether as a tree or a cobbling together of bubbles, I'm sure there would have been no objection to looking at it as an example of how the construction was historically conceived of.
The crucial thing about this example is that consists of a 'fused relative construction', a common-enough way of analysing such expressions nowadays.
The customary term for such constructions in generative grammar in the United States has for over forty years been
free relative clauses.
The fusion that exists in the word “whoever” can best be shown using a conventional tree diagram:
View attachment 4060
Here, the single word “whoever” can clearly be seen functioning not just as head of the NP but also as the relativised element functioning as subject of the relative clause. The meaning is thus comparable to the non-fused “the person who thought of that”.
In mainstream generative grammar in the United States, it is by no means conventional to represent a word as simultaneously realizing two syntactic positions. Such a representation, whatever style of diagram is used to make it, makes a claim about syntactic structure that would raise the eyebrows of many a syntactician.
I'll reinsert the tree diagram I gave earlier, in case the Reed-Kellogg diagrams, which I drew in addition to the tree, caused it to be snubbed by anyone.