Time-dependent series variance learning with recurrent mixture density networks

Nikolay Nikolaev; Peter Tino; Evgueni Smirnov

首页> 外文期刊>Neurocomputing >Time-dependent series variance learning with recurrent mixture density networks

【24h】

Time-dependent series variance learning with recurrent mixture density networks

机译：基于时间的序列方差学习与递归混合密度网络

获取原文

获取原文并翻译 | 示例

掌桥外文数据库（机构版） >>

开具论文收录证明 >>

文献代查 >>

页面导航

摘要
著录项
相似文献
相关主题

摘要

This paper presents an improved nonlinear mixture density approach to modeling the time-dependent variance in time series. First, we elaborate a recurrent mixture density network for explicit modeling of the time conditional mixing coefficients, as well as the means and variances of its Gaussian mixture components. Second, we derive training equations with which all the network weights are inferred in the maximum likelihood framework. Crucially, we calculate temporal derivatives through time for dynamic estimation of the variance network parameters. Experimental results show that, when compared with a traditional linear heteroskedastic model, as well as with the nonlinear mixture density network trained with static derivatives, our dynamic recurrent network converges to more accurate results with better statistical characteristics and economic performance.

机译：本文提出了一种改进的非线性混合密度方法，用于对时间序列中与时间有关的方差建模。首先，我们精心设计了一个循环混合密度网络，用于对时间条件混合系数及其高斯混合分量的均值和方差进行显式建模。其次，我们导出训练方程，利用该方程可以在最大似然框架中推断所有网络权重。至关重要的是，我们通过时间计算时间导数，以动态估算方差网络参数。实验结果表明，与传统的线性异方差模型以及采用静态导数训练的非线性混合密度网络相比，我们的动态递归网络收敛于更准确的结果，具有更好的统计特性和经济表现。

著录项

来源
《Neurocomputing》 |2013年第25期|501-512|共12页
作者
Nikolay Nikolaev; Peter Tino; Evgueni Smirnov;
展开▼
作者单位

Department of Computing, Goldsmiths College, University of London. London SE14 6NW, UK;

School of Computer Science, The University of Birmingham, Birmingham B15 2TT, UK;

Department of Knowledge Engineering, Faculty of Humanities and Science, Maastricht University, Maastricht 6200, MD, The Netherlands;

展开▼
收录信息美国《科学引文索引》(SCI);美国《工程索引》(EI);
原文格式 PDF
正文语种 eng
中图分类
关键词
Mixture density neural networks; GARCH models; Real-time recurrent learning algorithm;

机译：混合物密度神经网络GARCH模型;实时递归学习算法;

相似文献

外文文献
中文文献
专利

1. Learning continuous trajectories in recurrent neural networks with time-dependent weights [J] . Galicki M., Leistritz L. IEEE Transactions on Neural Networks . 1999,第4期

机译：学习时间依赖权重的递归神经网络中的连续轨迹
2. Machine learning for pore-water pressure time-series prediction:Application of recurrent neural networks [J] . Xin Wei, Lulu Zhang, Hao-Qing Yang, 地学前缘(英文版) . 2021,第001期

机译：孔隙水压时间序列预测的机器学习：经常性神经网络的应用
3. Uncertainty Quantification in Machine Learning Modeling for Multi-Step Time Series Forecasting: Example of Recurrent Neural Networks in Discharge Simulations [J] . International journal of applied mechanics . 2020,第3期

机译：用于多步时间序列预测的机器学习建模中的不确定性量化：排出模拟中经常性神经网络的示例
4. Time-Dependent Series Variance Estimation via Recurrent Neural Networks [C] . Nikolay Nikolaev, Peter Tino, Evgueni Smiruov International conference on artificial neural networks;ICANN 2011 . 2011

机译：递归神经网络的时间相关序列方差估计
5. Using Recurrent and Mixture Density Network Architectures to Model National Basketball Association In-Game Win Probabilities [D] . Poole, Henry Cleveland. 2020

机译：利用反复化和混合密度网络架构模拟全国篮球协会的游戏中的赢概率
6. Time-Dependent Recovery of Outcome Parameters in Ulnar Shortening for Positive Ulnar Variance: A Prospective Case Series [O] . Gernot Schmidle, Tobias Kastenberger, Rohit Arora 2018

机译：尺骨正方差缩短结果参数随时间的恢复：一个预期病例系列
7. Uncertainty-Aware Learning from Demonstration using Mixture Density Networks with Sampling-Free Variance Modeling [O] . Choi, Sungjoon, Lee, Kyungjae, Lim, Sungbin, 2017

机译：使用混合密度从示范中学习不确定性具有无采样方差建模的网络

Time-dependent series variance learning with recurrent mixture density networks

摘要

著录项

相似文献

相关主题

期刊订阅