#time series
9 resultsMoments of Maximum: Segment of AR(1)
Steven Finch
The paper studies the expected maximum and its variance for five consecutive observations from a stationary AR(1) time series, investigating how these quantities depend on the lag‑…
Sales Demand Forecast in E-commerce using a Long Short-Term Memory Neural Network Methodology
Kasun Bandara, Peibei Shi, Christoph Bergmeir +3
The paper proposes a globally trained LSTM model that leverages cross‑product time series information from an e‑commerce product hierarchy to improve sales demand forecasts, and ev…
Stochastic modeling of the time variability of ALMA calibrators
A. E. Guzmán, C. Verdugo, H. Nagai +5
The paper models the flux and spectral index variability of 39 ALMA calibrator quasars using continuous-time stochastic processes, showing that mixtures of Ornstein-Uhlenbeck proce…
An Independence Test Based on Recurrence Rates
Juan Kalemkerian, Diego Fernández
The paper proposes a new statistical test for independence between random elements, using a Cramér‑von Mises functional applied to a U‑process derived from recurrence rates, and pr…
Multitask learning and benchmarking with clinical time series data
Hrayr Harutyunyan, Hrant Khachatrian, David C. Kale +2
The paper introduces four benchmark prediction tasks derived from the MIMIC-III intensive care database—mortality risk, length of stay, physiologic decline detection, and phenotype…
Change-point detection in dynamic networks via graphon estimation
Zifeng Zhao, Li Chen, Lizhen Lin
The paper introduces a model‑free method that first estimates the underlying graphon of a dynamic network using a modified neighborhood smoothing algorithm, then applies a screenin…
Residual Entropy
Barnaby Rowe
The paper proposes augmenting the mean squared error loss with an entropy-based prior on residuals to detect and mitigate overfitting, especially for ordered data sequences, and de…
General proof of a limit related to AR(k) model of Statistics
Jan Vrbik
The paper proves a general formula for a limit that appears when computing moments of parameter estimators in autoregressive (AR(k)) models, extending earlier results that were lim…
Causal Discovery and Forecasting in Nonstationary Environments with State-Space Models
Biwei Huang, Kun Zhang, Mingming Gong +1
The paper introduces a nonlinear state-space modeling approach that leverages nonstationarity to identify causal relationships and improve forecasting in time series, using Bayesia…