基于改进的Random Subspace的客户投诉分类方法

电信业的客户投诉不断增多而又亟待高效处理。针对电信客户投诉数据的特点，提出了一种面向高维数据的改进的集成学习分类方法。该方法综合考虑客户投诉中的文本信息及客户通讯状态信息，基于Random Subspace方法，以支持向量机（Support Vector Machine，SVM）为基分类器，采用证据推理（Evidential Reasoning，ER）规则为一种新的集成策略，构造分类模型对电信客户投诉进行分类。所提模型和方法在某电信公司客户投诉数据上进行了验证，实验结果显示该方法能够显著提高客户投诉分类的准确率和投诉处理效率。     

Common Mode Voltage Reduction Using 3D-SVPWM for 3-level CI-NPC Inverter with Hybrid Energy System

This paper elucidates Common Mode Voltage (CMV) reduction in transformerless 3-phase 3-level Coupled Inductor Neutral Point (CI-NPC) Clamped Inverter with Hybrid Energy System. The three dimensional Space Vector Pulse Width Modulation (3D-SVPWM) with Nearest State Vector (NSV) is implemented to reduce the CMV by proper selection of medium, large and small vectors in 3D cubic space region. This NSV scheme in addition to CMV reduction, reduces the capacitor voltage balancing issues and minimizes switching losses. The proposed control provides full utilization of dc link voltage with reduced harmonics. This 3-level CI-NPC inverter is energized by hybrid energy source which includes photovoltaic system and wind energy system. The results obtained for the proposed scheme through simulation and experimental setup is compared with the conventional 2D-SVPWM and 3D-SVPWM scheme. From the compared results it is evident that the proposed scheme reduces CMV to a larger extent than 2D and 3D-SVPWM control. The simulation and experimental results are verified using matlab-simulink and FPGA-Spartan-6 controller respectively.     

Stacked autoencoder based deep random vector functional link neural network for classification

Extreme learning machine (ELM), which can be viewed as a variant of Random Vector Functional Link (RVFL) network without the input–output direct connections, has been extensively used to create multi-layer (deep) neural networks. Such networks employ randomization based autoencoders (AE) for unsupervised feature extraction followed by an ELM classifier for final decision making. Each randomization based AE acts as an independent feature extractor and a deep network is obtained by stacking several such AEs. Inspired by the better performance of RVFL over ELM, in this paper, we propose several deep RVFL variants by utilizing the framework of stacked autoencoders. Specifically, we introduce direct connections (feature reuse) from preceding layers to the fore layers of the network as in the original RVFL network. Such connections help to regularize the randomization and also reduce the model complexity. Furthermore, we also introduce denoising criterion, recovering clean inputs from their corrupted versions, in the autoencoders to achieve better higher level representations than the ordinary autoencoders. Extensive experiments on several classification datasets show that our proposed deep networks achieve overall better and faster generalization than the other relevant state-of-the-art deep neural networks.     

How do normalization schemes affect net spillovers? A replication of the Diebold and Yilmaz (2012) study

This paper replicates the Diebold and Yilmaz (2012) study on the connectedness of the commodity market and three other financial markets: the stock market, the bond market, and the FX market, based on the Generalized Forecast Error Variance Decomposition, GEFVD. We show that the net spillover indices (of directional connectedness), used to assess the net contribution of one market to overall risk in the system, are sensitive to the normalization scheme applied to the GEFVD. We show that, considering data generating processes characterized by different degrees of persistence and covariance, a scalar-based normalization of the Generalized Forecast Error Variance Decomposition is preferable to the row normalization suggested by Diebold and Yilmaz since it yields net spillovers free of sign and ranking errors.     

栈式降噪自编码器在波形单元识别中的应用

为提高多功能雷达(Multi-Function Radar,MFR)波形单元的识别准确率和鲁棒性,提出一种栈式降噪自编码器(Stacked Denoising Autoencoders,SDAE)与支持向量机(Support Vector Machine,SVM)相结合的波形单元识别方法.首先摒弃传统依赖脉冲序列分析技术对MFR信号进行处理的方法,通过分析波形单元结构并借助参数间的联合变化特征,提出一种MFR波形单元分段识别模型,将传统对脉冲序列的识别转化为对MFR波形单元的识别;然后在该模型的基础上引入SDAE,对训练样本数据、SDAE隐含层神经元节点进行加噪处理,并利用这些加噪后的样本数据训练优化SDAE网络模型,提取出样本数据的深层稳健特征;最后引入SVM算法,借助SDAE挖掘出的样本深层特征,实现SVM模型的优化,得到最终的波形单元识别模型(SDAE-SVM).仿真实验表明:提出的波形单元识别方法在相同样本数目和测试误差的条件下,与SVM算法相比,能够取得较高的识别准确率,具备更优越的识别效果.证实MFR波形单元识别模型是有效的,且通过SDAE网络的引入,使得SDAE-SVM方法能够自主地挖掘原始信号的深层特征,提高波形单元识别的鲁棒性和准确率.     

Modeling and parameter optimization for the design of vibrating screens

In this paper, we simulated the complex particle flow-behavior and screening efficiency on a linear vibrating screen using the Discrete Element Method (DEM). The simulations were validated with data from an adjustable experimental prototype screen. Then the novel application of non-linear regression modeling based on Support Vector Machines (SVMs) is used for mapping the sample space of operating parameters and vibrating screen configuration. Lastly, parameter optimization is implemented using Particle Swarm Optimization (PSO) algorithm. The primary findings proved that the SVM-based nonparametric model is not only feasible, but also highly adaptive to the parameter optimization that requires large-scale iterative computation. The non-parametric model established using the integration of DEM and SVM, combined with PSO algorithm in subsequent parameter optimization offered insights to the design and manufacture of vibrating screens.     

基于改进PSO-SVM的钻井液侵入储层深度预测

开展钻井液侵入储层深度预测,对于测井评价以及提高油井产能具有一定的现实意义。在分析钻井液侵入储层的机理和特征的基础上,提出了钻井液侵入储层的影响因素指标体系,建立了改进PSO-SVM的钻井液侵入储层深度预测模型,以塔里木塔中35口井为例进行了实证分析,并与传统BP神经网络和SVM模型预测结果进行了对比。研究结果表明:侵入深度与泥饼的渗透率、钻井液与储层压差以及侵入时间正相关,与储层孔隙度和钻井液粘度负相关,改进的PSO-SVM模型预测结果误差小,准确率高,能够用于钻井液侵入储层深度预测,具有广泛的应用前景。     

基于多颜色分量CLBP提取的浮选泡沫状态识别

针对不同状态的浮选泡沫图像之间纹理结构相似、颜色差异不明显的问题，提出一种基于色调、饱和度和亮度（HSV）颜色空间的完全局部二进制模式（CLBP）纹理提取的浮选泡沫状态识别方法。首先使用双域去噪在保留纹理细节的同时滤除图像噪声，然后转换为HSV图像，在H、S和V颜色分量上分别提取三个尺度的CLBP纹理特征。将提取的纹理特征归一化后线性排列，建立高维度的纹理分类模型。最后通过一对一模式的支持向量机分类器对四类泡沫状态的样本集进行纹理提取后的分类训练与测试。结果表明，该方法对不同浮选泡沫状态的分类正确率较高，优于其他纹理描述方法，适用于浮选泡沫状态的识别。     

A novel approach for efficient stance detection in online social networks with metaheuristic optimization

In the 19th and 20th centuries, social networks have been an important topic in a wide range of fields from sociology to education. However, with the advances in computer technology in the 21st century, significant changes have been observed in social networks, and conventional networks have evolved into online social networks. The size of these networks, along with the large amount of data they generate, has introduced new social networking problems and solutions. Social network analysis methods are used to understand social network data. Today, several methods are implemented to solve various social network analysis problems, albeit with limited success in certain problems. Thus, the researchers develop new methods or recommend solutions to improve the performance of the existing methods. In the present paper, a novel optimization method that aimed to classify social network analysis problems was proposed. The problem of stance detection, an online social network analysis problem, was first tackled as an optimization problem. Furthermore, a new hybrid metaheuristic optimization algorithm was proposed for the first time in the current study, and the algorithm was compared with various methods. The analysis of the findings obtained with accuracy, precision, recall, and F-measure classification metrics demonstrated that our method performed better than other methods.     

基于能量回馈与最小励磁电流的电梯节能研究

提出了基于能量回馈与最小励磁电流相结合的电梯节能控制方案,对电梯运行工况及节能方案进行了探讨。设计了基于dsPIC单片机的能量回馈控制系统,给出了相关电路的工作原理及设计方法并进行了实验,结果表明能量回馈系统设计合理。提出了基于最小励磁电流原理的电梯矢量节能控制策略,并进行了仿真实验研究。该方案可以实现电梯四个象限节能控制。