mb706

"Human in the Loop MBO" with multipoint proposal method "moimbo" proposes non-integer values for integer parameters

2

```r ps = makeParamSet( makeNumericParam("q", lower = -1, upper = 2), makeIntegerParam("v", lower = -2, upper = 3) ) des = generateDesign(n = 7, par.set = ps) des$y = c(1.20,...

"Human in the Loop MBO" breaks with AdaCB infill crit

1

Example, adapted from the [tutorial](https://mlrmbo.mlr-org.com/articles/supplementary/human_in_the_loop_MBO.html): ```r ps = makeParamSet( makeNumericParam("q", lower = -1, upper = 2), makeIntegerParam("v", lower = -2, upper = 3) ) des = generateDesign(n = 7, par.set...

FocusSearch doesn't handle discrete vector parameters

4

In particular, I reliably get an error [here](https://github.com/mlr-org/mlrMBO/blob/7ca9ad456e18024f8d1d1cacfe8f1a3ad3a67360/R/infillOptFocus.R#L58) when I have a discrete vector parameter, since `val` is a list and `setdiff(val.names, val)` can make `val.names` empty even when it...

fix.factors.prediction should not be set for the surrogate model learner

2

... in `checkLearner`. With `fix.factors.prediction` set to `TRUE`, factors that have not been seen before are treated the same as factors that have their requirement not fulfilled. Instead, `convertOptPathToDf` should...

RegrKm SE prediction for epistemic uncertainty

1

RegrKm SE prediction with nugget currently predicts the epistemic uncertainty (uncertainty of the mean prediction) + aleatoric uncertainty (nugget SE, uncertainty that the model sees as random error). E.g. sampling...

give glmnet the $importance slot

2

because then it could be used in combination with FilterEmbedded in mlr3featsel for feature selection in order of L1 inclusion. Importance could be the (approximate) lambda value at which a...

Cloning of classes with objects that reference each other

4

I already [commented on this](https://github.com/r-lib/R6/issues/110#issuecomment-461114660) in #110, but I'd like to get an "official" opinion on this (and I could make a PR if this is something desirable). Currently it...

Implement all.equal.R6 to cope with handling of ABs in recent R versions

2

R-devel has changed their handling of `as.list.environment()` regarding active bindings. Instead of turning the ABs into functions, it now gives the *values* that the ABs would have had, had they...

e.g. `po("nop_1")` to create a `po("nop")` with ID `nop_1`.

e.g. postfix `_1` or something