Classification metrics can't handle a mix of binary and continuous targets

This statement is generally true as classification metrics are designed to evaluate models that predict categorical or discrete outcomes. Continuous targets, on the other hand, represent a range of values and are not discrete. Therefore, using classification metrics to evaluate models that predict continuous targets can lead to inaccurate results.

For example, if we have a model that predicts the probability of a customer buying a product (continuous target), using a classification metric like accuracy (which measures the percentage of correctly classified instances) would not be appropriate. Instead, we would need to use regression metrics such as mean squared error (MSE) or mean absolute error (MAE) to evaluate the performance of the model.

It's important to choose the appropriate metric for the type of problem you are trying to solve and the type of target variable you are predicting.