Data Miner has a test result that is called a Performance Matrix. It contains what is also generally called a confusion matrix. This tracks where the model has false positives, false negatives etc.
You can manually reduce the false rates by adding weights to the model build or score process. But this assumes that you have some real costs that would make sense to apply to the predictions you are making.
If you simply want to make the model more accurate, then you will need to look at the methodology you are using.