minLSE is not being computed when data contains missing values

kdahlquist / GRNmap

Gene Regulatory Network modeling and parameter estimation

BSD 3-Clause "New" or "Revised" License

4 stars 3 forks source link

minLSE is not being computed when data contains missing values #353

Closed kdahlquist closed 6 years ago

kdahlquist commented 6 years ago

minLSE is not being computed when data contains missing values. Need to fix.

azinge commented 6 years ago

Leaving comment here due to there being no central thread for this week, both this and #352 have been fixed in beta, with tests passing. We are currently making tests for the new function, "computeStatistics".

kdahlquist commented 6 years ago

Track down the undeclared qq variable.

kdahlquist commented 6 years ago

We need to document the actual minLSE computation since it seems to keep coming up as a white board exercise. Will @cazinge and @jtorre39 write up the equation (in an equation editor) and post? Create a page for it on the wiki; later it will be moved to more formal documentation.

I understand that tests and function need to be written for this issue this week.

im-deepfriedwater commented 6 years ago

I have a picture of the equation but am not quite sure how we are handling posting images to the wiki. I will be able to post it on the wiki after the next meeting however.

azinge commented 6 years ago

We found some bugs in computeStatistics which we fixed and have also finished the test for it. @jtorre39 will run the full test suite on his laptop before the next meeting to make sure everything is else good.

im-deepfriedwater commented 6 years ago

Ran the tests on the current beta branch and all tests passing!

dondi commented 6 years ago

Code review results:

Include a comment in the computeStatisticsTest explaining when the "zero threshold" is 1E-15 vs. 1E-4
Remove TODO comment
Add test cases with negative numbers and zeroes

im-deepfriedwater commented 6 years ago

Added test cases and removed the TODO comment.

im-deepfriedwater commented 6 years ago

Added the diagram of minLSE to the wiki and also added the comment in our test specifying our use of lesser precision.

kdahlquist commented 6 years ago

Two final clean-up items and we can close this:

Add a trivial all-zero case to the calculation tests, just so we have that corner case
Add a description of the variables in the minLSE wiki page

im-deepfriedwater commented 6 years ago

Added the trivial all-zero case and added the description of the variables in the minLSE page.

kdahlquist commented 6 years ago

Looks good; closing.