FIGSHARE.pdf (600.8 kB)


Download (600.8 kB)
posted on 2017-08-11, 00:44 authored by Cyrus CousinsCyrus Cousins, Eli Upfal
Preprint of paper "The k -Nearest Representatives Classifier:
A Distance-Based Classifier with Strong Generalization Bounds" to appear in DSAA 2017. Abstract follows:
We define the k-Nearest Representatives (k-NR) classifier, a distance-based classifier similar to the k-nearest neighbors classifier with comparable accuracy in practice, and stronger generalization bounds. Uniform convergence is shown through Rademacher complexity, and generalizability is controlled through regularization. Finite-sample risk bound are also given. Compared to the k-NN, the k-NR requires less memory to store and classification queries may be made more efficiently. Training is also efficient, being polynomial in all parameters, and is accomplished via a simple empirical risk minimization process.


NSF award IIS-1247581, DARPA/Army award W911NF-16-1-0553, and DARPA/AFRL award FA8750-17-2-0102