After collecting the data set, the system is trained and achieves a high accuracy on the test data set. It is then deployed in the business and performs extremely poorly. In this case the system has probably learned the dataset off by heart instead of learning the correct trends. Select all of the statements below that are true, noting that negative marking is applied to wrong answers:

Group of answer choices

Day-to-day changes in appearance are confusing the system

The original labels definitely have errors

The training data does not have enough variety

The model generalises well to the data

The model has over-fitted the training data

The testing data is not representative of the real world deployment