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

, so how can we evaluate the performance of a model that predicts both known and unknown targets?

There are a few ways to approach this problem. One option is to split the dataset into known and unknown targets and evaluate the model's performance on each subset separately. For the unknown subset, you can use metrics such as precision, recall, and F1 score to evaluate the performance of the model's predictions. Another option is to use outlier detection techniques to identify the unknown targets and evaluate the model's performance on the remaining known targets. Additionally, you can use clustering algorithms to group targets together based on their similarity and evaluate the model's performance on each cluster separately. Ultimately, the best approach depends on the specific task and dataset, as well as the goals of the evaluation.