I need help highlighting cells using a Text Similarity Calculation Formula

Hello there. I need help with excel. To say it short (I'm also uploading photos), I have 3 columns of synonymous phrases, side by side, looking like this:

Column A Column B Column C

in America in the us in the usa
make sure you be sure you ensure you
odds are it’s likely that it’s likely
So how Just how So just how
this means that which means that which means
that you like that you want you want
to have a to truly have a to really have a
will be the would be the could be the
You're not You are not You aren't

Example1.png

My task is to calculate the two highest similar phrases between the phrases in columns A,B and C, comparing A with B, A with C and B with C. In fact, I already have done the comparison and its calculations, as shown on the photo, using Excel Powerup tools and the following formulas:

=pwrSIMILARITY(B:B,C:C,FALSE)
=pwrSIMILARITY(A:A,B:B,FALSE)
=pwrSIMILARITY(A:A,C:C,FALSE)

What I need however is, based on the result, the highest value, most similar two phrases, be it in Column A and B, A and C, or B and C, to be highlighted, and the one which is the least similar cell containing the third phrase - removed or just left non-highlighted.

Maybe it sounds a bit complicated at first read, but I think a picture's worth a thousand words and if you have questions, you may ask me, I will clarify it for you.

Example2.png

Most of the phrases inside it, whether on position (cell) 1, 2 or 3 (cells A,B,C), are highly similar, but often a phrase, which is not identical stays between these 2 similar phrases, sometimes in Cell A, sometimes in Cell B and sometimes in Cell C. Ascending and Descending order for arrangement does not work in here, I have already tried it. I need similarity calculation, which I have luckily already accomplished!

I need a formula, a method, a tool(s). Something, anything, which based on the highest similarity score, highlights the two most similar Cells in a row, be it A and B, A and C, or B and C.

Example3.png
Example4.png

Thank You Very Much. Looking forward to your replies.
Anything unclear? Ask me and I will reply immediately!

Try this out.
Columns O,P,Q place a 1 in the column to represent whether A,B,C are in the top 2 choices.

You could most likely put those formulas into the conditional formatting for each column instead.
I didn't try because this way you can tweak the formulas easier if necessary.

OK, I can't help you with the statistics, but I can tell you what my formula does.
And I can't check any other rows than are what are shown in the file because the cells all turn to errors if I try to unhide them.
Also the COMPARESTRING function doesn't exist in Excel 2010 either.
As for my formula, it's quite simple really and the "1" means nothing, it's just a marker.

Looking at the formula in Cell O1:- IF(OR(I1=MAX($F1,$I1,$L1),L1=MAX($F1,$I1,$L1)),1,"")

The parts outside the MAX bits - IF(OR(…….),1,"") - say that IF either of the 2 MAX statements are true then place a 1 in cell O1 or leave it blank (instead of 1 it could say "MAX" or "Highest Result" for instance.
The 1st MAX statement looks to see if I1=MAX(F1,I1,L1) and the 2nd looks to see if L1=MAX(F1,I1,L1)
So if either column I or column L are the MAX of the three then it will place a 1 in cell O1

Cell P1 does the same except it checks to see whether Column F or I have the MAX result
Cell Q1 does the same except it checks to see whether Column F or L have the MAX result

Then the conditional formatting simply colours columns A,B,C if columns O,P,Q have a 1 in their cells.
e.g. Since O34 and P34 have a 1 in them, A34 and B34 get coloured.

I'm not certain this was the best way to do it, but this at least explains what my formula is looking for and what it does.

I hope it helps you.

Thank You Very Much! Your formula works and I believe I understood what it does, at least partially. As for the statistics, I use parts of speech for extra orientation, but still it's not perfect, although better. I would really appreciate if I can somehow get better statistics, based on semantic similarity from someone ... Your formula works! It does!

Thank You