Hyperparameters for ARFClassifier

[jpfeil]

I want to parameterize the ARFClassifier to weigh false positives more than false negatives. I was wondering if anyone has recommended hyperparameters to focus the model on avoiding false positives. I tried using MacroFBeta with a small beta, but I'm not sure if this is doing what I think it is doing. I'll run a few tests but wanted to see if anyone has insight on this problem.

imblearn.RandomSampler(
            forest.ARFClassifier(
                n_models=200,
                metric=metrics.MacroFBeta(beta=0.1),
                memory_estimate_period=50000,
                remove_poor_attrs=True,
            ),
            desired_dist={True: 0.3, False: 0.7},