ECG信号自动诊断中回归建模法特征提取的研究

This article explores the ability of multivariate autoregressive model (MAR) and scalar AR model to extract the features from two-lead electrocardiogram signals in order to classify certain cardiac arrhythmias. The classification performance of four different ECG feature sets based on the model coefficients are shown. The data in the analysis including normal sinus rhythm, atria premature contraction, premature ventricular contraction, ventricular tachycardia, ventricular fibrillation and superventricular tachycardia is obtained from the MIT-BIH database. The classification is performed using a quadratic discriminant function. The results show the MAR coefficients produce the best results among the four ECG representations and the MAR modeling is a useful classification and diagnosis tool.     

Modeling timbre distance with temporal statistics from polyphonic music

Timbre distance and similarity are expressions of the phenomenon that some music appears similar while other songs sound very different to us. The notion of genre is often used to categorize music, but songs from a single genre do not necessarily sound similar and vice versa. In this work, we analyze and compare a large amount of different audio features and psychoacoustic variants thereof for the purpose of modeling timbre distance. The sound of polyphonic music is commonly described by extracting audio features on short time windows during which the sound is assumed to be stationary. The resulting down sampled time series are aggregated to form a high-level feature vector describing the music. We generated high-level features by systematically applying static and temporal statistics for aggregation. The temporal structure of features in particular has previously been largely neglected. A novel supervised feature selection method is applied to the huge set of possible features. The distances of the selected feature correspond to timbre differences in music. The features show few redundancies and have high potential for explaining possible clusters. They outperform seven other previously proposed feature sets on several datasets with respect to the separation of the known groups of timbrally different music.     

非线性时间序列建模的混合自回归滑动平均模型

提出了一类用于非线性时间序列建模的混合自回归滑动平均模型(MARMA).该模型是由K个平稳或非平稳的ARMA分量经过混合得到的.讨论了MARMA模型的平稳性条件和自相关函数.给出了MARMA模型参数估计的期望极大化(expectation maximization)算法.运用贝叶斯信息准则(Bayes information criterion)来选择该模型.MARMA模型分布形式富于变化的特征使得它能够对具有多峰分布以及条件异方差的序列进行建模.通过两个实例验证了该模型,并和其他模型进行比较,结果表明MARMA模型能够更好地描述这些数据的特征.     

基于MAR模型的SAR图像的聚类分割

合成孔径雷达（SAR）图像中的斑点噪声是SAR图像处理困难的主要原因,如何抑制斑点噪声及图像处理一直是SAR图像研究的热点。根据SAR图像的成像机理,采用能够描述不同尺度（分辨率）下固有特性的多尺度自回归（MAR）模型,提出一种有效的多尺度抑制斑点噪声和分割方法。首先对SAR图像多分辨率序列建立MAR模型,然后依据模型对SAR图像抑制斑点噪声,重构,最后用Ward聚类分割方法对SAR图像进行分割、比较。     

Ordinal-based and frequency-based integration of feature selection methods for sentiment analysis

Feature subset selection with the aim of reducing dependency of feature selection techniques and obtaining a high-quality minimal feature subset from a real-world domain is the main task of this research. For this end, firstly, two types of feature representation are presented for feature sets, namely unigram-based and part-of-speech based feature sets. Secondly, five methods of feature ranking are employed for creating feature vectors. Finally, we propose two methods for the integration feature vectors and feature subsets. An ordinal-based integration of different feature vectors (OIFV) is proposed in order to obtain a new feature vector. The new feature vector depends on the order of features in the old vectors. A frequency-based integration of different feature subsets (FIFS) with most effective features, which are obtained from a hybrid filter and wrapper methods in the feature selection task, is then proposed. In addition, four well-known text classification algorithms are employed as classifiers in the wrapper method for the selection of the feature subsets. A wide range of comparative experiments on five widely-used datasets in sentiment analysis were carried out. The experiments demonstrate that proposed methods can effectively improve the performance of sentiment classification. These results also show that proposed part-of-speech patterns are more effective in their classification accuracy compared to unigram-based features.     

基于多尺度随机模型的SAR图像无监督分割

合成孔径雷达（synthetic aperture radar,SAR）是一种基于相干原理的成像系统,在SAR图像中存在严重影响图像质量的斑点噪声,使得SAR图像的可靠分割非常困难。〖BP）〗根据SAR图像的成像机理,利用两种多尺度随机模型,即多尺度自回归（Multiscale Autoregressive,MAR）模型和多尺度自回归滑动平均（Multiscale Aautoregressive Moving Average, MARMA）模型,分别来描述同一场景不同分辨率SAR图像像素间的统计相关性,并构造了相应的多分辨混合算法实现SAR图像的无监督分割。试验结果表明,提出的两种无监督分割方法是可行的,且MARMA模型比MAR模型能够更精确地捕捉SAR图像多尺度序列中不同类型地形的统计信息,使分割质量具有明显的改进。     

基于多模式特征聚合的未来商业预测

准确预测商业销售量未来趋势对于企业开发经营、政府宏观调控等至关重要.传统的数据预测方法计算时间开销大,具有主观性,而现有基于数据驱动的未来商业预测方法没有考虑到数据集中的特征多样.商业销售量数据是一个时序数据,时序数据中包含了丰富的时间窗特征、滞后历史特征和价格变化趋势特征等众多特征,先前的研究往往只注重于其中的某些特征,对于特征的融合和增强探究偏少,现有的未来商业预测方法的预测精度仍然有待提高.为此,本文提出了一种基于多模式特征聚合的未来商业预测方法,该方法首先将商业销售量数据进行预处理;然后基于特征工程提取数据集的5组不同的时间窗特征和其他特征;在机器学习上对于5组时间窗特征采用硬投票机制选择合适的模型训练,同时也采用神经网络的优化模型提取时序特征和预测结果,然后分析销售量数据集和某些特征之间的依赖关系;最后基于软投票模型完整地模型融合实现了商业销售量的高精度预测.一系列实验结果表明,本文提出的方法具有较高预测精度和效率,明显优于现有预测方法.     

多元时序的深度自编码器聚类算法

针对深度聚类算法对多变量时间序列数据（MTS）的特征提取能力不足等问题,提出一种新的深度聚类结构模型（MDTC）。为了提取MTS的关键特征并实现降维,提出一维卷积学习MTS的属性和时序维度的特征表示与循环神经网络等网络层组成的自编码器结构;为了提高模型对时序特征的表示能力,提出了MCBAM时序注意力模块,用于增强MTS序列中不同时间段的表示特征。在九个公开UEA多元时序数据集进行了实验,模型的自编码器结构相较其他自编码器在七个数据集上提升了2%~9%;模型的MCBAM模块相较其他注意力模块在六个数据集上提升了0.3%~2%。实验表明MDTC模型结构和MCBAM模块的有效性,同时模型对比其他聚类算法具有优异的表现。     

基于混合网络模型的多维时间序列预测

为捕捉时间序列中潜在的特征依赖关系并实现高维时序数据的快速模糊预测,构建基于时间卷积网络（TCN）与自注意力机制的两种混合网络模型：TSANet和TSANet-MF。TSANet模型通过全局和局部两个并行卷积分量结构提取特征后,利用自注意力机制增强特征点关联程度,并结合并行的TCN增大卷积的感受野范围,最大程度地捕捉多维时序数据的周期性特征。TSANet-MF模型将TSANet作为矩阵分解算法的正则化项,使高维数据转化为具有更多时序特征的低维数据,减少计算复杂度,实现高维数据的快速模糊预测。在4种不同领域的时间序列数据集上的实验结果表明,TSANet模型在3种数据集上的预测性能均优于基准模型,尤其在高维Traffic数据集上相对平方根误差降低了19.52%~56.37%,TSANet-MF模型在Electricity和Traffic高维数据集上的训练时间相比于基准模型明显减少。上述实验结果验证了两种混合网络模型均具有较好的多维时间序列预测性能。     

Regional classification of Chinese folk songs based on CRF model

Music regional classification, which is an important branch of music automatic classification, aims at classifying folk songs according to different regional style. Chinese folk songs have developed various regional musical styles in the process of its evolution. Regional classification of Chinese folk songs can promote the development of music recommendation systems which recommending proper style of music to users and improve the efficiency of the music retrieval system. However, the accuracy of existing music regional classification systems is not high enough, because most methods do not consider temporal characteristics of music for both features extraction and classification. In this paper, we proposed an approach based on conditional random field (CRF) which can fully take advantage of the temporal characteristics of musical audio features for music regional classification. Considering the continuity, high dimensionality and large size of the audio feature data, we employed two ways to calculate the label sequence of musical audio features in CRF, which are Gaussian Mixture Model (GMM) and Restricted Boltzmann Machine (RBM). The experimental results demonstrated that the proposed method based on CRF-RBM outperforms other existing music regional classifiers with the best accuracy of 84.71% on Chinese folk songs datasets. Besides, when the proposed methods were applied to the Greek folk songs dataset, the CRF-RBM model also performs the best.

     

Large margin mixture of AR models for time series classification

In this paper, we propose the large margin autoregressive (LMAR) model for classification of time series patterns. The parameters of the generative AR models for different classes are estimated using the margin of the boundaries of AR models as the optimization criterion. Models that use a mixture of AR (MAR) models are considered for representing the data that cannot be adequately represented using a single AR model for a class. Based on a mixture model representing each class, we propose the large margin mixture of AR (LMMAR) models. The proposed methods are applied on the simulated time series data, electrocardiogram data, speech data for E-set in English alphabet and electroencephalogram time series data. Performance of the proposed methods is compared with that of support vector machine (SVM) based classifier that uses AR coefficients based features. The proposed methods give a better classification performance compared to the SVM based classifier. Being generative models, the LMAR and LMMAR models provide a generative interpretation that enables utilization of the rejection option in the high risk classification tasks. The proposed methods can also be used for detection of novel time series data.     

基于多尺度跳跃深度长短期记忆网络的短期多变量负荷预测

近年来,以循环神经网络（RNN）为主体构建的预测模型在短期电力负荷预测中取得了优越的性能。然而,由于RNN不能有效捕捉存在于短期电力负荷数据的多尺度时序特征,因而难以进一步提升负荷预测精度。为了捕获短期电力负荷数据中的多尺度时序特征,提出了一种基于多尺度跳跃深度长短期记忆（MSD-LSTM）网络的短期电力负荷预测模型。具体来说,以长短期记忆（LSTM）网络为主体构建预测模型能够较好地捕获长短期时序依赖,从而缓解时序过长时重要信息容易丢失的问题。进一步地,采用多层LSTM架构并且对各层设置不同的跳跃连接数,使得MSD-LSTM的每一层能够捕获不同时间尺度的特征。最后,引入全连接层把各层提取到的多尺度时序特征进行融合,再利用该融合特征进行短期电力负荷预测。实验结果表明,与单层LSTM和多层LSTM相比,MSD-LSTM的均方误差总体下降了10%。可见MSD-LSTM能够更好地提取短期负荷数据中的多尺度时序特征,从而提高短期电力负荷预测的精度。     

A novel classifier for multivariate instance using graph class signatures

Applications like identifying different customers from their unique buying behaviours, determining ratingsof a product given by users based on different sets of features, etc. require classification using class-specific subsets of features. Most of the existing state-of-the-art classifiers for multivariate data use complete feature set for classification regardless of the different class labels. Decision tree classifier can produce class-wise subsets of features. However, none of these classifiers model the relationship between features which may enhance classification accuracy. We call the class-specific subsets of features and the features’ interrelationships as class signatures. In this work, we propose to map the original input space of multivariate data to the feature space characterized by connected graphs as graphs can easily model entities, their attributes, and relationships among attributes. Mostly, entities are modeled using graphs, where graphs occur naturally, for example, chemical compounds. However, graphs do not occur naturally in multivariate data. Thus, extracting class signatures from multivariate data is a challenging task. We propose some feature selection heuristics to obtain class-specific prominent subgraph signatures. We also propose two variants of class signatures based classifier namely: 1) maximum matching signature (gMM), and 2) score and size of matched signatures (gSM). The effectiveness of the proposed approach on real-world and synthetic datasets has been studied and compared with other established classifiers. Experimental results confirm the ascendancy of the proposed class signatures based classifier on most of the datasets.     

An evaluation of Landsat TM spectral data and SAR-derived textural information for lithological discrimination in the Red Sea Hills,Sudan

The effectiveness of spectral and textural information in the identification of surface rock types in an arid region, the Red Sea Hills of Sudan, is evaluated using spectral information from the six Landsat TM optical bands and textural features derived from Shuttle Imaging Radar-C (SIR-C) C-band HH polarization data. An initial classification is derived from Landsat TM data alone using three classification algorithms, Gaussian maximum likelihood, a multi-layer feed-forward neural network and a Kohonen self-organizing feature map (SOM), to generate lithological maps, with classification accuracy being measured using a confusion matrix approach. The feed-forward neural net produced the highest overall classification accuracy of 57 per cent and was, therefore, selected for the second experiment, in which texture measures from SIR-C C-band HH-polarized synthetic aperture radar (SAR) data are added to selected TM spectral features. Four methods of measuring texture are employed, based on the Fourier power spectrum, grey level co-occurrence matrix (GLCM), multi-fractal measures, and the multiplicative autoregressive random field (MAR) model. The use of textural information together with a subset of the TM spectral features leads to an increase in classification accuracy to almost 70 per cent. Both the MAR model and the GLCM matrix approach perform better than Fourier and multi-fractal based methods of texture characterization.     

Semisupervised self-learning granular neural networks for remote sensing image classification

Accuracy of a pattern classification model mostly depends on ample number of training samples, which is the major bottleneck for classifying land cover of remote sensing images. Further, the unbalance scenario typically encountered in hyperspectral remote sensing images, i.e., limited number of training samples with more dimensions, makes the decision-making process cumbersome. Under such inevitable constraints, the article aims to develop an improved classification model using semisupervised self-learning granular neural networks (GNNs) for remote sensing images. The proposed semisupervised method has adopted a new strategy for selecting the potential candidate samples from the unlabeled dataset and used GNN as the base classifier. We have considered GNN because of its transparent architecture that leads to improved performance with less computational complexity compared to the conventional neural networks. Performance of the model is further enhanced with fuzzy granulation of features using class belonging information and selection of granulated features using neighborhood rough sets (NRS). The proposed model thus takes the mutual advantages of GNN architecture, fuzzy granulation with class belonging information, NRS-based feature selection and the most important, improved semisupervised self-learning approach. Performance of the model is compared with other similar methods and verified in terms of different performance measurement indexes, using two multispectral and two hyperspectral remote sensing images.     

基于合成数据的刚体目标姿态追踪网络

现有刚体姿态估计存在数据稀缺、复杂场景下的低鲁棒性及低实时性等问题,为此提出一种基于合成数据的刚体目标位姿追踪网络结构。采用时空间特征融合技术,捕捉时间与空间特征信息,生成具有时空敏感的特征图;利用残差连接学习更为丰富和抽象的优质特征,改善追踪目标的准确性;对稀缺数据进行数据增强,生成符合现实物理特性的复杂合成数据,以此训练深度学习模型,提高模型的泛化性。在YCB-Video数据集中选取7个物体进行实时姿态追踪实验,结果表明,提出的方法相较于同类相关方法,在复杂场景下对刚体姿态估计的更为准确,在实时估计效率上表现最优。     

Recognizing diseases with multivariate physiological signals by a DeepCNN-LSTM network

The usage of multivariate time series to identify diseases plays an important role in the medical field, as it can help medical staff to improve diagnose accuracy and reduce medical costs. Current research shows that deep Convolutional Neural Networks (CNN) can automatically capture features from raw data and Long Short-Term Memory (LSTM) networks can manage and learn temporal dependence between time series data such as physiological signals. In this work, we propose a deep learning framework called DeepCNN-LSTM by combining the CNN and LSTM to leverage their respective advantages for disease recognition, allowing itself to characterize complex temporal varieties with multiple autoencoded features. In particular, we use stationary wavelet transform together with median filter to preprocess low-frequency signal data, and introduce sliding window to segment physiological time series before model training for performance improvement on the training speed as well as the accuracy for recognizing diseases. In addition, we validate our model on a hybrid benchmark dataset collecting from MIMIC and Fantasia databases and set up four kinds of comparative experiments. Empirical evaluations on the benchmark dataset demonstrate that the proposed model outperforms other competitive approaches.