Worst-Case Analysis of Selective Sampling for Linear-Threshold Algorithms

We provide a worst-case analysis of selective sampling algorithms for learning linear threshold functions. The algorithms considered in this paper are Perceptron-like algorithms, i.e., algorithms which can be efficiently run in any reproducing kernel Hilbert space. Our algorithms exploit a simple margin-based randomized rule to decide whether to query the current label. We obtain selective sampling algorithms achieving on average the same bounds as those proven for their deterministic counterparts, but using much fewer labels. We complement our theoretical findings with an empirical comparison on two text categorization tasks. The outcome of these experiments is largely predicted by our theoretical results: Our selective sampling algorithms tend to perform as good as the algorithms receiving the true label after each classification, while observing in practice substantially fewer labels.

• Type: Conference
• Year: 2004
• Where: NIPS
• Authors: Nicolò Cesa-Bianchi, Claudio Gentile, Luca Zaniboni
• Keywords: Fewer Labels, NIPS 2004, NIPS 2007, Perceptron-like Algorithms, Selective Sampling Algorithms