Twin neural network improved k-nearest neighbor regression

Twin neural network regression is trained to predict the difference between the regression targets of two data points rather than the individual targets themselves. In order to improve k-nearest neighbor regression, this difference prediction can be applied to adjust the targets of the k-nearest neighbors of an unknown data point before averaging. The proposed algorithm is shown to outperform both neural networks and k-nearest neighbor regression on most of our small- to medium-sized datasets.