It is at very initial stage. Getting this reviewed can speed up things with lesser mistakes.
Diff Detail
- Repository
- R262 LabPlot
- Lint
Lint Skipped - Unit
Unit Tests Skipped
Changed model to data driven model so that, we can tests on multiple sets without writing multiple compare functions.
src/backend/hypothesisTest/HypothesisTest.cpp | ||
---|---|---|
1073 ↗ | (On Diff #60916) | fabs instead of abs. |
tests/stats/ttest/TTestTest.cpp | ||
41 | add comment: "first sample" | |
49 | add comment "second sample" | |
54 | add space before | |
87 | Can you use the DEBUG() macro instead of qDebug()? | |
92 | using an absolute error of 0.01 depends heavily on tValue. Why is the error so big? Can you use relative errors instead? | |
139 | DEBUG()? | |
144 | see above | |
159 | check comment | |
191 | DEBUG()? | |
196 | see above |
tests/stats/ttest/TTestTest.cpp | ||
---|---|---|
54 | Sorry, I didn't get. Add space before? | |
87 | QDEBUG and DEBUG macro are not working in my pc, may be due to some setting. I have to check for it. I will use the macro, once I will fix the setting. | |
92 | I tried using relative error. But, I am not getting what value of delta to use. If I use as small as 0.1 also then the test is not passing for values: actual: 1.51815e-05 which is very more accurate than expected. |
tests/stats/ttest/TTestTest.cpp | ||
---|---|---|
54 | I meant to add an empty line to make it more readable. | |
87 | I thought that you already fixed it for you. If it's not working for you, please add a TODO-comment when using qDebug(). | |
92 | the problem here is that the expected value is not known more precise? An absolute error of 0.01 make not much sense here. |