Curve matching is a new big data technique to predict the future growth of individual growth and development. The key idea is to find a small number of children in the existing data who are 'similar' to the child for which we want prediction. The realized growth patterns of the matched children suggest how the target child might evolve in the future.
An appealing feature of curve matching is that each matched growth trajectory represents real growth of real children. The spread between the matched curves provides a natural indication of the uncertainty of the prediction.
The key scientific issues revolve around the exact definition of 'optimal matches', and the statistical properties of the resulting inferences. In this lecture, I will outline the principle, discuss limitations and extensions, and demonstrate how curve matching can be used in practice.
|