arrow
返回

Nonparametric filtering, estimation and classification using neural jump ODEs

delete2025-09-01
delete0
PRE
AI
J
Jakob Heiss
F
Florian Krach *
T
Thorsten Schmidt
F
Félix B. Tambe-Ndonfack
DOI:10.1515/strm-2025-0001delete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
神经跳跃ODE模型通过神经ODE对观测值之间的条件期望进行建模,并在新观测值到达时进行跳跃。它们已在非规则和不完整观测值设置中展示了完全数据驱动的在线预测的有效性,且在弱正则性假设下运行。本研究将此框架扩展到输入-输出系统,从而能够在在线滤波和分类中直接应用。我们为该方法建立了理论收敛保证,为L2-最优滤波提供了稳健的解决方案。经验实验强调了该模型在经典参数化方法之上的优越性能,特别是在具有复杂底层分布的场景中。这些结果强调了该方法在金融和健康监测等时效性领域中的潜力,其中实时精度至关重要。
Keyword:
Classification
filtering
input-output systems
neural jump ODEs
optimal estimation

期刊

S
STATISTICS & RISK MODELING
IF:
0.9
论文数:
8
被引数:
0

机构

暂无机构信息
引用论文

引用论文

Learning Internal Representations by Error Propagation
err
IF0
err1985-09-01
err0
PREAI
errDavid E. Rumelhart; Geoffrey E. Hinton; Ronald J. Williams
err分享
err收藏
err分享
err收藏
Investigating a Domain Adaptation Approach for Integrating Different Measurement Instruments in a Longitudinal Clinical Registry
err2025-02-01
err0
PREAI
errHackenberg,Maren; Pfaffenlehner,Michelle; Behrens,Max; Pechmann,Astrid; Kirschner,Janbernd; Binder,Harald
err分享
err收藏
Stochastic Calculus and Applications
err2015-01-01
err0
PREAI
errSamuel N. Cohen; Robert J. Elliott
err分享
err收藏
KalmanNet: Neural Network Aided Kalman Filtering for Partially Known Dynamics
err2022-01-01
err0
errOAAI
errGuy Revach; Nir Shlezinger; Xiaoyong Ni; Adria Lopez Escoriza; Ruud J. G. van Sloun; Yonina C. Eldar
err分享
err收藏
err分享
err收藏
学者 查看更多内容