Waterlogging risk assessment based on self-organizing map(SOM)artificial neural networks:a case study of an urban storm in Beijing

Due to rapid urbanization, waterlogging induced by torrential rainfall has become a global concern and a potential risk affecting urban habitant's safety. Widespread waterlogging disasters haveoccurred almost annuallyinthe urban area of Beijing, the capital of China. Based on a selforganizing map(SOM) artificial neural network(ANN), a graded waterlogging risk assessment was conducted on 56 low-lying points in Beijing, China. Social risk factors, such as Gross domestic product(GDP), population density, and traffic congestion, were utilized as input datasets in this study. The results indicate that SOM-ANNis suitable for automatically and quantitatively assessing risks associated with waterlogging. The greatest advantage of SOM-ANN in the assessment of waterlogging risk is that a priori knowledge about classification categories and assessment indicator weights is not needed. As a result, SOM-ANN can effectively overcome interference from subjective factors,producing classification results that are more objective and accurate. In this paper, the risk level of waterlogging in Beijing was divided into five grades. The points that were assigned risk grades of IV or Vwere located mainly in the districts of Chaoyang, Haidian, Xicheng, and Dongcheng.

Enhanced Self-Organizing Map Neural Network for DNA Sequence Classification

The artificial neural networks (ANNs), among different soft computing methodologies are widely used to meet the challenges thrown by the main objectives of data mining classification techniques, due to their robust, powerful, distributed, fault tolerant computing and capability to learn in a data-rich environment. ANNs has been used in several fields, showing high performance as classifiers. The problem of dealing with non numerical data is one major obstacle prevents using them with various data sets and several domains. Another problem is their complex structure and how hands to interprets. Self-Organizing Map (SOM) is type of neural systems that can be easily interpreted, but still can’t be used with non numerical data directly. This paper presents an enhanced SOM structure to cope with non numerical data. It used DNA sequences as the training dataset. Results show very good performance compared to other classifiers. For better evaluation both micro-array structure and their sequential representation as proteins were targeted as dataset accuracy is measured accordingly.

A MULTILAYER FEEDFORWARD NEURAL NETWORK MODEL FOR VISUAL MOTION PERCEPTION

The local visual motion detection mechanism used in the visual systems of primatescan only sense the motion component oriented perpendicularly to the contrast gradient of thebrightness pattern.But the visual system of higher animals can adaptively determine the actualdirection of motion through a learning process.In this paper a multilayered feedforward neuralnetwork model for perception of visual motion is presented.This model employs W.Reichardt’selementary motion detectors array and T.Kohonen’s self-organizing feature map.We explored theself-organizing principles for perception of visual motion.The computer simulations show thatthis neural network is able to recognize the true direction of motion through an unsupervisedlearning process.In addition,the neurons with the same or similar motion direction selectivitytend to appear in“functional columns”which seem to be qualitatively similar to the corticalmotion columns observed by electrophysiological and cytohistochemical studies in certain higherareas such as MT.It proves that motion-detection by spatio-temporal coherences,mapping,co-operation,competition,and Hebb rule may be the basic principles for the self-organization ofvisual motion perception networks.

A New Image Coding Algorithm Based on Self-Organizing Neural Network

The paper deals with a new VQ+DPCM+DCT algorithm based on Self-Organizing Feature Maps(SOFM) algorithm for image coding. In addition. a Frequency sensitive SOFM (FSOFM) has been also devel-oped. Simulation results show that a very good visual quality of the coded image at 0.252 bits/pixel is obtained.

Self-organizing feature map neural network classification of the ASTER data based on wavelet fusion

Most methods for classification of remote sensing data are based on the statistical parameter evaluation with the assumption that the samples obey the normal distribution. How-ever, more accurate classification results can be obtained with the neural network method through getting knowledge from environments and adjusting the parameter (or weight) step by step by a specific measurement. This paper focuses on the double-layer structured Kohonen self-organizing feature map (SOFM), for which all neurons within the two layers are linked one another and those of the competition layers are linked as well along the sides. Therefore, the self-adapting learning ability is improved due to the effective competition and suppression in this method. The SOFM has become a hot topic in the research area of remote sensing data classi-fication. The Advanced Spaceborne Thermal Emission and Reflectance Radiometer (ASTER) is a new satellite-borne remote sensing instrument with three 15-m resolution bands and three 30-m resolution bands at the near infrared. The ASTER data of Dagang district, Tianjin Munici-pality is used as the test data in this study. At first, the wavelet fusion is carried out to make the spatial resolutions of the ASTER data identical; then, the SOFM method is applied to classifying the land cover types. The classification results are compared with those of the maximum likeli-hood method (MLH). As a consequence, the classification accuracy of SOFM increases about by 7% in general and, in particular, it is almost as twice as that of the MLH method in the town.

Morphological self-organizing feature map neural network with applications to automatic target recognition

The rotation invariant feature of the target is obtained using the multi-direction feature extraction property of the steerable filter. Combining the morphological operation top-hat transform with the self-organizing feature map neural network, the adaptive topological region is selected. Using the erosion operation, the topological region shrinkage is achieved. The steerable filter based morphological self-organizing feature map neural network is applied to automatic target recognition of binary standard patterns and real world infrared sequence images. Compared with Hamming network and morphological shared-weight networks respectively, the higher recognition correct rate, robust adaptability, quick training, and better generalization of the proposed method are achieved.

Neural network-based matrix effect correction in EDXRF analysis

In this paper we discuss neural network-based matrix effect correction in energy dispersive X-ray fluorescence (EDXRF) analysis,with detailed algorithm to classify the samples.The method can correct the matrix effect effectively through classifying the samples automatically,and influence of X-ray absorption and enhancement by major elements of the samples is reduced.Experiments for the complex matrix effect correction in EDXRF analysis of samples in Pangang showed improved accuracy of the elemental analysis result.

System partitioning on MCM using a new neural network model

A new self-organizing neural network model is presented, which can get rid of some fatal defects facing the Kohonen self-organizing neural network, known as the slow training speed, difficulty in designing neighboring zone, and disability to deal with area constraints directly. Based on the new neural network, a new approach for performance-driven system partitioning on MCM is presented. In the algorithm, the total routing cost between the chips and the circle time are both minimized, while satisfying area and timing constraints. The neural network has a reasonable structure and its training speed is high. The algorithm is able to deal with the large scale circuit partitioning, and has total optimization effect. The algorithm is programmed with Visual C + + language, and experimental result shows that it is an effective method.

基于混合双层自组织径向基函数神经网络的优化学习算法

针对传统方法采用先训练后测试两阶段学习机制极易导致的过拟合或欠拟合问题,提出一种基于混合双层自组织径向基函数神经网络的优化学习(hybrid bilevel self-organizing radial basis function neural network optimization learning,Hb-SRBFNN-OL)算法。首先,将训练过程和测试过程集成到一个统一的框架中,规避过拟合或欠拟合问题。其次,基于进化学习机制,提出上下2层的交互式优化学习算法,上层基于网络复杂度和测试误差自组织调整网络结构,下层采用列文伯格-马夸尔特(Levenberg Marquardt,LM)算法作为优化器对自组织径向基函数神经网络(self-organizing radial basis function neural network,SO-RBFNN)的连接权值进行优化。最后,利用来自多个子网络的综合信息生成模型的最终输出,加速网络全局收敛。为验证所提方法的可行性,分别在多个分类和预测任务中进行了测试实验。结果表明,在与传统神经网络结构相似甚至更好的测试和分类精度下,该方法不仅能实现更快的训练收敛,而且能进化成更精简紧凑的径向基函数神经网络(radial basis function neural network,RBFNN)模型。尤其在污水处理过程中总磷的质量浓度预测实验中,测试集中均方根误差(root mean squared error,RMSE)最高可降低48.90%,实际场景实验结果验证了所提算法的精确性更佳且泛化能力更强。

基于主成分自组织神经网络法的测井曲线分层技术

在砂岩型铀矿找矿工作中,提高测井岩性分层效率和精度至关重要。为提高砂岩型铀矿岩性分层效果,本文采用主成分分析法对多个测井曲线进行降维处理,将主成分分析法的第一主成分、第二主成分、第三主成分作为自组织神经网络的样本数据,进行自组织神经网络训练,将训练好的网络模型用于砂岩型铀矿岩性的自动化分层。实验结果显示:主成分自组织神经网络法岩性分层精度可达到85%以上,高于传统自组织神经网络算法78%的分层精度,具有更好的测井岩性分层效果。因此,主成分自组织神经网算法的岩性分层方法有效减少了输入样本的种类,简化了自组织神经网络结构,其自动化分层效果要优于传统的自组织神经网络算法。本文的研究结果表明,主成分自组织神经网算法在砂岩型铀矿领域岩性识别工作中具有较好的应用效果。

基于SOM-BP的全自动口罩机传动系统故障检测

针对口罩机在多工序生产中故障特征难以诊断的问题,提出了一种基于自组织映射(SOM)和误差反向传播网络(BP)的故障检测模型。首先针对4种减速机故障类型搭建SOM-BP复合型神经网络模型并完成检测分类,其次通过提取原振动信号的20组时域和频域参数作为SOM网络的输入样本进行初步聚类,并根据仿真结果确定最佳竞争层结构,最后将聚类后结果输入BP网络进行预测并完成分类,实现故障检测。研究结果表明,7×7竞争层结构下的SOM-BP复合型神经网络对于减速机的8种时域和频域参数的检测效果最优,分类准确率可达93.5%,173次迭代即可收敛,数据拟合度最高达0.99876,达到实际检测要求,验证了该方案的有效性和可行性。

压差式流量计误差自动化修正算法研究

针对在多干扰源扰动下压差式流量计测量结果面临输出不稳、误差较大的问题,提出多源扰动下的压差式流量计误差自动化修正算法。考虑全补偿气体可膨胀性系数、压缩系数、密度系数和流出系数等因素,研究压差式流量计误差自动化修正算法。利用均值滤波滤除信号中的高斯噪声,结合一阶滞后滤波优化卡尔曼滤波算法,修正多源扰动误差。引入自组织算法和Volterra神经网络进一步改进卡尔曼滤波算法,并优化卡尔曼滤波算法的先验模型参数,以实现多源扰动误差的自动化修正。试验结果表明,经该算法控制后:当参考流量为900 m^(3)/h时,示值误差绝对值为0.203%;当参考流量为700 m^(3)/h时,流量计重复性为0.06%。该研究可以有效识别并修正由于多源扰动造成的流量异常值,且流量测量精度较高。

基于模糊神经网络的数字媒体数据自动化采集系统设计

为了改善数字媒体数据传输效果,确保系统安全、稳定运行,设计基于模糊神经网络的数字媒体数据自动化采集系统。构建数字媒体数据自动化采集系统框架,数据采集层利用数据采集卡获取不同渠道的数字媒体数据,由数据处理层的数据融合模块调用改进模糊神经网络算法完成数字媒体数据的融合处理后,通过数据传输层的分层自组织无线网络将其传输至存储应用层,实现数字媒体数据的存储、查询、显示与输出。实验结果表明:该系统采集的音频信号波形规律、曲线平滑、功率均值波动误差在(0,0.14),满足允许误差范围;数据传输效果优于采用DSR协议的单点传输方式,当数字媒体数据源为3时,平均峰值信噪比指标最高;具有数字媒体数据查询功能。

基于有效性分析的自组织模糊神经网络建模方法

提出了一种基于有效性分析的自组织模糊神经网络(self-organizingfuzzyneural network based on effectiveness analysis, SOEFNN)建模方法。首先,提出了一种针对模糊规则的有效性评价指标,利用样本与规则层输出之间的映射关系进行网络模型的有效性分析,通过累积触发的方式实现相应模糊规则的增加或删减,使网络模型在能够处理复杂非线性问题的同时降低其冗余性,使模型更为紧凑。采用梯度下降算法对网络模型进行训练。然后,对所提出的SOEFNN模型进行非线性系统仿真实验和污水处理过程中的出水生化需氧量预测建模,并与其他自组织模糊神经网络模型进行对比。仿真结果表明,所提出的SOEFNN模型能够很好地实现结构和参数的自适应调整,并且具有较好的逼近能力。

基于时间序列和神经网络的电力设备状态异常检测方法

为进一步提高电力设备异常检测方法对设备信息的利用率,发现更多潜在的设备故障,结合大数据分析技术和设备评估技术,提出了一种基于时间序列和神经网络的状态数据异常检测方法。通过时间序列自回归模型和自组织映射神经网络将连续的电力设备数据离散为单个序列,计算状态变量在时间轴上的转移概率,通过状态转移概率和聚类算法快速检测数据异常。通过实验对该方法的有效性进行验证。结果表明,该方法可以快速、有效地检测电力设备异常状态。

基于混合自组织映射神经网络的云南省山洪灾害危险性区划

开展云南省山洪灾害危险性区划工作,以自组织映射神经网络为基础,混合Ward、PAM、CLARA、K-means和HK-means的5种方法进行二阶聚类,应用戴维森堡丁指数(Davies-Bouldin index,DBI)、轮廓系数(silhouette coefficient,SC)、聚类模型评估指数(Calinski-Harabaz index,CH)确定最佳聚类方案,之后以变异系数和变异系数一阶拆分确定最佳区划数量.结果显示:①SOM(self organizing map)+CLARA(clustering LARge applications)方法通过聚类有效性检验效果最好,其DBI值为1.0、SC值为0.9、CH值为0.3334,基于该方法得到云南省山洪灾害危险性最佳聚类数为5类,呈现类别空间分离,灾害属性相似的特征;②通过变异系数(coefficient of variation,CV)值变化及变异系数一阶差分(first-order difference,FOD)最低取值确定云南省山洪灾害危险性最佳区划单元为16个,具有形状上与地貌单元相近、数量上与行政单元相同,内部灾害发生机理相似的特征;③通过山洪灾害点、降水量、高程地貌的可视化比较,地理探测器定量分析,表明区划结果有较高的区内一致性和区间异质性.

基于mRMR-SOM的异步电机轴承故障诊断研究

针对异步电机轴承故障诊断问题,提出了一种融合最大相关最小冗余特征选择算法(mRMR)和自组织映射神经网络(SOM)的故障诊断方法,并将其应用于轴承故障诊断的不同阶段。首先,在实验室环境下搭建了异步电机故障诊断试验平台,在不同电机状态下分别采集振动、电流和电压信号,利用统计学方法获取了高维混合特征集;然后,以互信息为背景,利用mRMR根据特征与状态标签间的相关性和特征间的冗余性,筛选了具备强区分能力的特征,以避免计算冗余和后验诊断性能下降;最后,采用SOM对异步电机健康和轴承故障状态进行了分类识别,验证了SOM对异步电机轴承故障诊断的有效性,以及mRMR对故障诊断结果的影响。研究结果表明:基于mRMR-SOM的异步电机轴承故障诊断方法能够准确地区分健康和故障状态,测试集分类准确率达到89%;使用mRMR特征筛选能够将154维特征降低至17维,缩短23.5%的网络收敛时间,并将分类准确率由89%提升至98%;试验结果验证了基于mRMR-SOM的异步电机轴承故障诊断方法对于异步电机轴承故障诊断问题的有效性,且证实其具备良好的诊断效果。

自样本特征构造的1DCNN-BiLSTM网侧光伏功率预测

为解决电网难以获取NWP数据和无法建立光伏功率预测模型的问题,提出一种自样本特征构造的一维卷积双向长短期记忆神经网络光伏发电功率预测方法。通过K均值聚类和功率骤减事件检测的特征工程获取细粒度的天气状态标签,实现基于自身样本的特征构造,以解决样本特征缺少问题;采用卷积和长短期记忆网络结合的模型结构,解决局部特征提取和长期依赖的问题。算例验证结果表明,所提方法改善整体的预测性能,降低多特征数据存在的数据匮乏和数据稳定性风险,为模型输入特征较少的网侧光伏功率短期预测提供一种有效途径。

面向电力营销的多源日志安全数据挖掘方法

针对当前电力营销业务系统内部电力营销数据分散、缺乏对电力营销数据统一管理,在多源电力营销数据库中应用了Apache Lucene的Elasticsearch分布式搜索引擎。通过采用主控芯片型号为XC7Z035FFGH676-2的Cortex-A9处理器,提高了电力营销多源电力营销安全数据信息的挖掘和计算能力;通过自组织映射神经网络与模糊聚类算法的聚类分析方法,提高了电力营销数据异常检测能力;利用自组织映射神经网络与模糊聚类算法减少能源数据消耗,提高了数据挖掘能力。所提方法的聚类分析时间最短为104 s,为下一步研究奠定了基础。