Assume that you are given a data set and a neural network model trained on the data set. You are asked to build a decision tree model with the sole purpose of understanding/interpreting the built neural network model. In such a scenario, which among the following measures would you concentrate most on optimising?

The correct answer is D. comprehensibility of the decision tree model, measured in terms of the size of the corresponding rule set.

The purpose of building a decision tree model is to understand/interpret the built neural network model. This means that the decision tree model should be easy to understand and interpret, so that we can learn more about how the neural network model works.

Accuracy and F1 measure are measures of the performance of a model, but they do not tell us anything about the model’s interpretability. Fidelity is the fraction of instances on which the neural network and the decision tree give the same output. This is a measure of how well the decision tree model captures the behavior of the neural network model, but it does not tell us anything about how easy the decision tree model is to understand.

Comprehensibility is a measure of how easy a model is to understand. It is often measured in terms of the size of the corresponding rule set. A smaller rule set means that the model is easier to understand, because there are fewer rules to keep track of.

Therefore, the best measure to optimize in this scenario is comprehensibility.