how did the data used to train the neural network affect the model? Think about the collection method and the responses in the example. How accurate is the model? Does the observed accuracy suffice for implementation of a rewards programme based on movement activity, keeping in mind that there would likely be a financial incentive tied to activity performance? What can be done to improve the accuracy of the model? Think about the various components of the learning problem, such as the data, the model, etc. What other modes of motion could be useful to predict when designing an activity rewards programme