Normalize cost-benefit matrix values

[angela97lin]

Closes #1098

[codecov]

Codecov Report

Merging #1099 into main will not change coverage.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main    #1099   +/-   ##
=======================================
  Coverage   99.91%   99.91%           
=======================================
  Files         192      192           
  Lines       10719    10719           
=======================================
  Hits        10710    10710           
  Misses          9        9           

Impacted Files	Coverage Δ
evalml/objectives/cost_benefit_matrix.py	100.00% <100.00%> (ø)
.../tests/objective_tests/test_cost_benefit_matrix.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 554848c...817cae4. Read the comment docs.

[jeremyliweishih]

LGTM 🦺

[dsherry]

Nice, good call on this! I'd expect no change in ML performance from this, but its a good clarification. The benefit: if a user runs the same cost-benefit on two different training sets, where one has more rows than the other, the cost-benefit scores will have the same bounds.

I left one request for a math clarification.

[kmax12]

i think we need to update the docs after this change

[angela97lin]

@kmax12 Thanks for reminding me, updated docs here: https://evalml.alteryx.com/en/1098_normalize/demos/cost_benefit_matrix.html