{"id":"arxiv/2609.24153","url":"https://arxiv.org/abs/2609.24153","tags":["stat.ME","math.ST"],"score":0,"title":"Quantifying Long-Range Dependence in Object-Valued Time Series","domain":"stat.ME","source":"arxiv","authors":["Won-Ki Seo","Jiazhen Xu","Han Lin Shang"],"reviews":0,"subject":"stat.ME","summary":"Object-valued time series, including distributions, covariance matrices, networks, and compositions, lack the linear structure required by conventional autocovariance-based definitions of long memory. We develop an intrinsic framework for defining and estimating long memory in metric spaces of negative type. An isometric embedding yields a centered Hilbert-valued process, and the trace of its lag-covariance operator provides a signed, additive measure of temporal dependence. Crucially, this trace equals the difference between the marginal mean pairwise distance and expected lagged distance,…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-09-21","reactions":{}}