Now only t-test is available in RapidMiner. F-test is also commonly-used test, e.g. in Multiple Linear regression.
Now in linear regression, we can use M5 Prime or greedy algorithm to select attributes. But for the result linear coefficients, there is no hypothesis testing on whether the "linear" is suitable model. In Clemetine there is such function, I wish RapidMiner will support it too.
