When computing overall accuracy as done in the table above (ie., accuracy for the model overall rather than conditional accuracy statistics), what is an assumption of this approach when evaluating how well the model is accurately predicting median collegiate ACT score?

Options:

The overall accuracy statistics assume the error is similar for all median ACT scores.

The overall accuracy statistics assume the errors are part of a symmetric distribution.

The overall accuracy statistics assume that there is a lot of data in their computation.

Use the following table for the next two questions: Error Type Mean Median Minimum Maximum Number Training Data 1.60 1.19 O 13.5 967 Testing Data 1.75 1.32 O 15 322 Note: These statistics are the absolute error which is calculated as the absolute value of the difference between the observed median ACT score and the predicted median ACT score. The "Number" column here represents how many data points the statistics were computed for in the training and testing data respectively. For example, 967 means that there were 967 absolute error statistics for the training data