2020 IEEE 23rd International Conference on Information Fusion (FUSION)最新文献_第3页

ND-SMPF: A Noisy Deep Neural Network Fusion Framework for Stock Price Movement Prediction ND-SMPF:一种用于股价运动预测的噪声深度神经网络融合框架

2020 IEEE 23rd International Conference on Information Fusion (FUSION) Pub Date : 2020-07-01 DOI: 10.23919/FUSION45008.2020.9190453

Farnoush Ronaghi, Mohammad Salimibeni, F. Naderkhani, Arash Mohammadi

{"title":"ND-SMPF: A Noisy Deep Neural Network Fusion Framework for Stock Price Movement Prediction","authors":"Farnoush Ronaghi, Mohammad Salimibeni, F. Naderkhani, Arash Mohammadi","doi":"10.23919/FUSION45008.2020.9190453","DOIUrl":"https://doi.org/10.23919/FUSION45008.2020.9190453","url":null,"abstract":"There has been a recent surge of interest on development of news-oriented Deep Neural Network (DNN) architectures to predict stock trend movements. Limited focus is, however, devoted to reliability fusing different available information resources. In this regard, this paper proposes a Noisy Deep Stock Movement Prediction Fusion framework (ND-SMPF) for stock price movement prediction. The proposed ND-SMPF predictive framework uses information fusion to combine twitter data with extended horizon market historical prices to boost the accuracy of the stock movement prediction task. More specifically, Noisy Bi-directional Gated Recurrent Unit (NBGRU) is utilized coupled with a Hybrid Attention Network (HAN) to extract news level temporal information. A two level attention layer is used to identify relevant words with highest correlation and effects on the stock trends, which are then fused with historical price data to perform the prediction task. A real dataset is incorporated to evaluate performance of the proposed ND-SMPF framework, which illustrates superior performance in comparison to its recently developed counterparts.","PeriodicalId":419881,"journal":{"name":"2020 IEEE 23rd International Conference on Information Fusion (FUSION)","volume":"95 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124869832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 4

Multiple Basic Proposal Distributions Model Based Sampling Particle Filter 基于多基本提议分布模型的采样粒子滤波

2020 IEEE 23rd International Conference on Information Fusion (FUSION) Pub Date : 2020-07-01 DOI: 10.23919/FUSION45008.2020.9190385

Lihong Shi, Feng Yang, Litao Zheng, Xiaoxu Wang, Liang Chen

引用次数: 0

A Sequential $L_{p}$-norm Filter for Robust Estimation 一种用于鲁棒估计的序列$L_{p}$范数滤波器

2020 IEEE 23rd International Conference on Information Fusion (FUSION) Pub Date : 2020-07-01 DOI: 10.23919/FUSION45008.2020.9190602

Yang Yang

引用次数: 0

Sound Source Localization and Reconstruction Using a Wearable Microphone Array and Inertial Sensors 基于可穿戴式传声器阵列和惯性传感器的声源定位与重构

2020 IEEE 23rd International Conference on Information Fusion (FUSION) Pub Date : 2020-07-01 DOI: 10.23919/FUSION45008.2020.9190480

Clas Veibäck, Martin A. Skoglund, F. Gustafsson, Gustaf Hendeby

{"title":"Sound Source Localization and Reconstruction Using a Wearable Microphone Array and Inertial Sensors","authors":"Clas Veibäck, Martin A. Skoglund, F. Gustafsson, Gustaf Hendeby","doi":"10.23919/FUSION45008.2020.9190480","DOIUrl":"https://doi.org/10.23919/FUSION45008.2020.9190480","url":null,"abstract":"A wearable microphone array platform is used to localize stationary sound sources and amplify the sound in the desired directions using several beamforming methods. The platform is equipped with inertial sensors and a magnetometer allowing predictions of source locations during orientation changes and compensation for the displacement in the array configuration. The platform is modular, open and 3D printed to allow for easy reconfiguration of the array and for reuse in other applications, e.g., mobile robotics. The software components are based on open source. A new method for source localization and signal reconstruction using Taylor expansion of the signals is proposed. This and various standard and non-standard Direction of Arrival (DOA) methods are evaluated in simulation and experiments with the platform to track and reconstruct multiple and single sources. Results show that sound sources can be localized and tracked robustly and accurately while rotating the platform and that the proposed method outperforms standard methods at reconstructing the signals.","PeriodicalId":419881,"journal":{"name":"2020 IEEE 23rd International Conference on Information Fusion (FUSION)","volume":"474 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132584838","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 3

Worldwide Ground Target State Propagation 全球地面目标状态传播

2020 IEEE 23rd International Conference on Information Fusion (FUSION) Pub Date : 2020-07-01 DOI: 10.23919/FUSION45008.2020.9190210

D. Crouse

引用次数: 0

Inverse Covariance Intersection Fusion of Multiple Estimates 多重估计的逆协方差交集融合

2020 IEEE 23rd International Conference on Information Fusion (FUSION) Pub Date : 2020-07-01 DOI: 10.23919/FUSION45008.2020.9190614

Jiří Ajgl, O. Straka

引用次数: 3

Adaptive IMM-CFusion for a Remote IMM Track and Local Measurements 远程IMM跟踪和本地测量的自适应IMM融合

2020 IEEE 23rd International Conference on Information Fusion (FUSION) Pub Date : 2020-07-01 DOI: 10.23919/FUSION45008.2020.9190354

Rong Yang, Y. Bar-Shalom

{"title":"Adaptive IMM-CFusion for a Remote IMM Track and Local Measurements","authors":"Rong Yang, Y. Bar-Shalom","doi":"10.23919/FUSION45008.2020.9190354","DOIUrl":"https://doi.org/10.23919/FUSION45008.2020.9190354","url":null,"abstract":"The problem addressed in this paper is the tracking a maneuvering target in a distributed sensor network using a real-world-motivated fusion configuration. The local node $mathrm{S}_{mathrm{L}}$ receives a track from the remote node $mathrm{S}_{mathrm{R}}$ generated by its Interacting Multiple Model (IMM) estimator, and the “inside information” such as motion models, mode probabilities and mode-conditioned estimates of the remote track is unknown. The local node $mathrm{S}_{mathrm{L}}$ has its own measurements which need to be fused with the remote IMM track. This problem can be solved using an existing technique with the following steps: 1) generate a $mathrm{S}_{mathrm{L}}$ track based on its own measurements; 2) perform Track-to-Track Fusion (T2TF) on $mathrm{S}_{mathrm{R}}$ and $mathrm{S}_{L}$ tracks. This paper will develop an alternative approach to fuse the remote IMM track and the local measurements directly. We called it the IMM Cumulated information Fusion (IMM-CFusion). The IMM-CFusion estimates the cumulated information of the local SLmeasurements with multiple models, and then fuses this cumulated information with the remote IMM track state. The IMM-CFusion shows better performance than the T2TF approach in a test case.","PeriodicalId":419881,"journal":{"name":"2020 IEEE 23rd International Conference on Information Fusion (FUSION)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130309793","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

Camera Localization Based on Belief Clustering 基于信念聚类的摄像机定位

2020 IEEE 23rd International Conference on Information Fusion (FUSION) Pub Date : 2020-07-01 DOI: 10.23919/FUSION45008.2020.9190358

Huiqin Chen, Emanuel Aldea, S. L. Hégarat-Mascle

引用次数: 0

Track-Before-Detect for Bistatic Radar Based on Velocity Filtering 基于速度滤波的双基地雷达探测前跟踪

2020 IEEE 23rd International Conference on Information Fusion (FUSION) Pub Date : 2020-07-01 DOI: 10.23919/FUSION45008.2020.9190333

Tao Han, Liangliang Wang, Gongjian Zhou

{"title":"Track-Before-Detect for Bistatic Radar Based on Velocity Filtering","authors":"Tao Han, Liangliang Wang, Gongjian Zhou","doi":"10.23919/FUSION45008.2020.9190333","DOIUrl":"https://doi.org/10.23919/FUSION45008.2020.9190333","url":null,"abstract":"Conventional track-before-detect methods for bistatic radar usually detect weak target according to an approximate target motion model. The inaccuracy of this approximation motion model tends to cause degradation of detection performance. To remedy this issue, a weak target detection algorithm for bistatic radar with velocity filtering based on the accurate target motion model is addressed in this paper. First, the evolutions of bistatic range and incident angle are obtained by the constant velocity motion model in Cartesian coordinate system (CCS). Then, the cells in the bistatic polar coordinate system (PCS) are transformed into the CCS and their positions in the CCS are predicted with the help of the supposed velocity. Further, the predicted positions are transformed inversely into the bistatic PCS and the measurement of each cell is added to that of the cell nearest its predicted position to realize multiframe accumulation. The process of energy integration and the output envelope of bistatic PCS are deduced elaborately. Finally, the effectiveness and superiority of the proposed method is demonstrated by several simulation results.","PeriodicalId":419881,"journal":{"name":"2020 IEEE 23rd International Conference on Information Fusion (FUSION)","volume":"248 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123028002","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Uncertainty based active learning with deep neural networks for inertial gait analysis 基于不确定性主动学习的深度神经网络惯性步态分析

2020 IEEE 23rd International Conference on Information Fusion (FUSION) Pub Date : 2020-07-01 DOI: 10.23919/FUSION45008.2020.9190449

Alexander Vaith, B. Taetz, G. Bleser

{"title":"Uncertainty based active learning with deep neural networks for inertial gait analysis","authors":"Alexander Vaith, B. Taetz, G. Bleser","doi":"10.23919/FUSION45008.2020.9190449","DOIUrl":"https://doi.org/10.23919/FUSION45008.2020.9190449","url":null,"abstract":"Inertial measurement units (IMUs) enable the capture of human motion in-field. This can be used in various analysis applications ranging from the medical domain over sports to daily activities. Manual data labelling for classification or regression tasks is often time-consuming and cumbersome, in particular when it comes to larger datasets. Active learning algorithms try to reduce the labelling cost, for instance via suggesting samples with high prediction uncertainty that should be explicitly labelled. In this work, we apply two probabilistic deep learning approaches on different state-of-the-art deep neural network structures for timeseries data, with uncertainty based measures to actively query sample labels. This is applied to gait phase classification using IMU data as inputs. We demonstrate the performance on a newly captured dataset, where we obtained high accuracy (up to 96%) with up to 43% fewer samples as compared to random sampling in an online setting. In an offline setting, we could extract heel strike and toe-off foot events with an accuracy of 99.9% using active learning strategies with up to 58% fewer samples.","PeriodicalId":419881,"journal":{"name":"2020 IEEE 23rd International Conference on Information Fusion (FUSION)","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130272576","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 7