This question was migrated from the Slack channel #questions.
Original Question:
Hello everyone. I am using FLOWER to develop a federated model on tabular datasets. It is a multiclassification problem, and I am using ANN as the base model. I want to generate a confusion matrix for each client at the end of each round. How should I do that? Has anyone tried that before?
Answer:
Hi! In order to compute a confusion matrix for each client at the end of each round, the easiest way to do it is to compute your confusion matrix inside the evaluate function of your clients. To compute the confusion matrix you can, for instance, use the following scikit-learn function: scikit-learn.org/stable/modules/generated/sklearn.metrics.confusion_matrix.html.