RF: Optimize ICC_rep_anova using a global cache #3407
Conversation
Codecov Report
@@ Coverage Diff @@
## master #3407 +/- ##
==========================================
- Coverage 65.21% 65.21% -0.01%
==========================================
Files 307 307
Lines 40474 40485 +11
Branches 5351 5354 +3
==========================================
+ Hits 26395 26402 +7
- Misses 13004 13006 +2
- Partials 1075 1077 +2
Flags with carried forward coverage won't be shown. Click here to find out more.
Continue to review full report at Codecov.
|
There was a problem hiding this comment.
@bbfrederick Thanks for your patience. I think when I first looked at this, I wanted to try to figure out an alternative to using global, but I don't think that should hold this up. Giving up on that dream and revisiting this, the only suggestion I have is that we can use the @ operator to have more readable matrix multiplication. Happy to merge this in as-is, though.
| x = kron(eye(nb_conditions), ones((nb_subjects, 1))) # sessions | ||
| x0 = tile(eye(nb_subjects), (nb_conditions, 1)) # subjects | ||
| X = hstack([x, x0]) | ||
| centerbit = dot(dot(X, pinv(dot(X.T, X))), X.T) |
There was a problem hiding this comment.
| centerbit = dot(dot(X, pinv(dot(X.T, X))), X.T) | |
| centerbit = X @ pinv(X.T @ X) @ X.T |
|
|
||
| # Sum Square Error | ||
| predicted_Y = dot(dot(dot(X, pinv(dot(X.T, X))), X.T), Y.flatten("F")) | ||
| predicted_Y = dot(centerbit, Y.flatten("F")) |
There was a problem hiding this comment.
| predicted_Y = dot(centerbit, Y.flatten("F")) | |
| predicted_Y = centerbit @ Y.flatten("F") |
|
Actually, yes, let's revert the |
effigies
changed the title
|
@bbfrederick Looking to do another release next week, if you want to try to get this one finished off? |
|
Consider #3454 as an alternative. |
Summary
Fixes #3406 .
List of changes proposed in this PR (pull-request)
Acknowledgment