Artificial Intelligence: Question 8
Syllabus 18.1
A fitness-tracking company wants to predict a user's resting heart rate, in beats per minute, from their age. It trains a regression method on a large set of past records, each already stating a user's age and their measured resting heart rate.
(a) Describe what a regression method does with this training data in order to build a model that can make predictions. [2]
(b) Once trained, the model is given the age of a new user whose resting heart rate has not been measured. Explain how the trained model uses this to produce its predicted resting heart rate for the new user. [2]
(c) A separate team at the company instead wants to predict whether a user should be labelled "active" or "inactive", not a numeric value. Explain why a regression method, on its own, would not be an appropriate choice for this second task. [2]
Show worked solution Hide worked solution
Worked solution
Part (a): What a regression method does with the training data
A regression method looks at the whole set of training records, pairs of a user’s age and their actual, measured resting heart rate, and fits a function (mathematical relationship) between age and resting heart rate.
The function is fitted so that the values it would predict for each age in the training data are, on the whole, as close as possible to the actual resting heart rates recorded in that same training data. That is, the fitting process minimises the overall difference between predicted and actual values across all the training records.
[2 marks]: [1] for describing a function being fitted to the training data, [1] for explaining that fitting minimises the difference between predicted and actual values across the training data.
Part (b): Using the trained model to predict for a new user
To make a prediction for the new user, the model is given only the new user’s age. The input, and applies the function it fitted during training to that age, calculating an output value.
This output value is the model’s predicted resting heart rate for the new user. The model never needs, or is given, the new user’s actual resting heart rate. The whole purpose of training is to be able to estimate it for users whose true value is not yet known.
[2 marks]: [1] for stating that the new user’s age is put into the fitted function, [1] for explaining that the function’s output is the predicted resting heart rate, without needing the new user’s actual value.
Part (c): Why regression is not suitable for the second task
A regression method is designed to predict a continuous numeric value. A number that could, in principle, take a wide range of values, such as a resting heart rate in beats per minute.
Labelling a user as “active” or “inactive” is not a numeric-value prediction: it is a choice between a small, fixed set of named labels. Since a regression method fits a function that outputs a number, it is not designed to produce one of a small set of named labels, so it is not, on its own, an appropriate choice for this second task.
[2 marks]: [1] for stating that regression methods predict a continuous numeric value, [1] for explaining that “active” / “inactive” is a choice between named labels, not a numeric value, so regression does not fit this task.
Final answers
- (a) A regression method fits a function to the training data, minimising the difference between its predicted values and the actual recorded resting heart rates
- (b) The new user’s age is put into the fitted function, which outputs the predicted resting heart rate, without needing the new user’s actual value
- (c) Regression methods predict a continuous numeric value, not a choice between named labels such as “active” / “inactive”, so regression alone is not suitable for this task