Optimal amortized regret in every interval

TL;DR

This paper introduces a randomized algorithm that guarantees near-optimal regret bounds in any interval for sequence prediction, with practical evaluation on stock data.

Contribution

It presents a novel randomized algorithm achieving amortized $O(\,\sqrt{x})$ regret in any interval, improving interval-specific performance guarantees.

Findings

Achieves $O(\,\sqrt{x})$ regret in any interval

Constant factor in regret bound estimated around 2.1 for sequences up to length 2000

Effective in high-frequency stock data prediction

Abstract

Consider the classical problem of predicting the next bit in a sequence of bits. A standard performance measure is {\em regret} (loss in payoff) with respect to a set of experts. For example if we measure performance with respect to two constant experts one that always predicts 0's and another that always predicts 1's it is well known that one can get regret $O(\sqrt T)$ with respect to the best expert by using, say, the weighted majority algorithm. But this algorithm does not provide performance guarantee in any interval. There are other algorithms that ensure regret $O(\sqrt {x \log T})$ in any interval of length $x$. In this paper we show a randomized algorithm that in an amortized sense gets a regret of $O(\sqrt x)$ for any interval when the sequence is partitioned into intervals arbitrarily. We empirically estimated the constant in the $O()$ for $T$ upto 2000 and found it to be small -- around 2.1. We also experimentally evaluate the efficacy of this algorithm in predicting high frequency stock data.