Subject Cross Validation in Human Activity Recognition
This addresses evaluation bias in Human Activity Recognition systems, which is an incremental improvement for researchers and practitioners in the field.
The paper tackled the problem of overestimated performance in Human Activity Recognition due to correlated samples from the same subjects in k-fold cross validation, finding that it artificially increases performance by about 10% (16% with overlapping windows) and that overlapping windows provide no performance gain.
K-fold Cross Validation is commonly used to evaluate classifiers and tune their hyperparameters. However, it assumes that data points are Independent and Identically Distributed (i.i.d.) so that samples used in the training and test sets can be selected randomly and uniformly. In Human Activity Recognition datasets, we note that the samples produced by the same subjects are likely to be correlated due to diverse factors. Hence, k-fold cross validation may overestimate the performance of activity recognizers, in particular when overlapping sliding windows are used. In this paper, we investigate the effect of Subject Cross Validation on the performance of Human Activity Recognition, both with non-overlapping and with overlapping sliding windows. Results show that k-fold cross validation artificially increases the performance of recognizers by about 10%, and even by 16% when overlapping windows are used. In addition, we do not observe any performance gain from the use of overlapping windows. We conclude that Human Activity Recognition systems should be evaluated by Subject Cross Validation, and that overlapping windows are not worth their extra computational cost.