Variance of entropy for testing time-varying regimes with an application to meme stocks
创建于 更新于
摘要
The paper derives an unbiased, high-order (up to O(n^{-4})) analytical approximation for the sample Shannon-entropy variance, proposes a z-test for detecting statistically significant entropy changes between adjacent intervals, and introduces a self-consistent criterion to select the optimal rolling window length; the method is validated in simulations and applied to US meme (GME, BBBY, AMC) and IT stocks showing that low-entropy regimes are associated with increased price predictability and higher returns from a simple empirical-frequency trading rule [page::0][page::6][page::21].
