An Improved Kernel K-Mean Cluster Method and Its Application in Fault Diagnosis of Roller Bearing

For the kernel K-mean cluster method is run in an implicit feature space, the initial and iterative cluster centers cannot be defined explicitly. Against the deficiency of the initial cluster centers selected in the original space discretionarily in the existing methods, this paper proposes a new method for ensuring the clustering center that virtual clustering centers are defined in the feature space by the original classification as the initial cluster centers and the iteration clustering centers are ensured by the further virtual classification. The improved method is used for fault diagnosis of roller bearing that achieves a good cluster and diagnosis result, which demonstrates the effectiveness of the proposed method.

改进全局K-Means聚类算法的汽车行驶工况研究

为有效促进汽车节能减排和新技术发展,文中提出了一种改进全局K Means聚类算法的汽车行驶工况构建方法,通过采集城市道路行驶工况的数据并对数据进行预处理,利用主成分分析法和改进的K Means聚类算法分别对运动学片段中实验数据的12个特征参数进行降维和聚类,拟合出某城市汽车行驶工况。分析结果表明:拟合曲线的汽车运动特性能更好代表所采集数据源的相应特性,两者的误差小,时耗低,行驶工况拟合度高,能综合反映实际车辆运行的状况。

Global Warming in Japanese Cities from 1960 to 2019 Using Machine Learning

In this study, we investigated the variations in warming between Japanese cities for 1960-1989, and 1990-2019 using principal component analysis (PCA) and k-means clustering. The precipitation and sunshine hours exhibited opposite tendencies in the PCA results. It was found that 1960M and 1990M had a correlation (r = 0.51). The 1960M and 1990M are the mean temperature anomalies in Japanese cities for 1960-1989 and 1990-2019, respectively. There was a strong correlation between temperature and precipitation (r = 0.62). There was an inverse correlation between 1960M and sunshine hours (r = −0.25), but a correlation between 1990M and sunshine hours (r = 0.11). Sunshine hours had less effect on the 1960M but more impact on the 1990M. The k-means clustering for 1960M and 1990M can be classified into four types: high 1960M and high 1990M, which indicates that global warming is progressing rapidly (Sapporo, Tokyo, Kyoto, Osaka, Fukuoka, Nagasaki), low 1960M and low 1990M, global warming is progressing slowly (Nemuro, Ishinomaki, Yamagata, Niigata, Fushiki, Nagano, Karuizawa, Mito, Suwa, Iida, Hamada, Miyazaki, Naha), low 1960M and high 1990M, global warming has accelerated since 1990 (Utsunomiya, Kofu, Okayama, Hiroshima), and normal 1960M and normal 1990M, the rate of warming is normal among the 38 cities (Asahikawa, Aomori, Akita, Kanazawa, Maebashi, Matsumoto, Yokohama, Gifu, Nagoya, Hamamatsu, Kochi, Kagoshima). Higher annual temperatures were correlated with higher annual precipitation according to the k-means clustering of temperature and precipitation. Two of the four categories consisted of places with high annual temperatures and high precipitation (Fushiki, Kanazawa, Kochi, Miyazaki, Kagoshima, Naha, Ishigakijima), and places with low annual temperatures and low precipitation (Asahikawa, Nemuro, Sapporo, Karuizawa).

A K-Means Clustering-Based Multiple Importance Sampling Algorithm for Integral Global Optimization

In this paper, we propose a K-means clustering-based integral level-value estimation algorithm to solve a kind of box-constrained global optimization problem. For this purpose, we introduce the generalized variance function associated with the level-value of the objective function to be minimized. The variance function has a good property when Newton’s method is used to solve a variance equation resulting by setting the variance function to zero. We prove that the largest root of the variance equation is equal to the global minimum value of the corresponding optimization problem. Based on the K-means clustering algorithm, the multiple importance sampling technique is proposed in the implementable algorithm. The main idea of the cross-entropy method is used to update the parameters of sampling density function. The asymptotic convergence of the algorithm is proved, and the validity of the algorithm is verified by numerical experiments.

A privacy-preserving vehicle trajectory clustering framework

As one of the essential tools for spatio‒temporal traffic data mining,vehicle trajectory clustering is widely used to mine the behavior patterns of vehicles.However,uploading original vehicle trajectory data to the server and clustering carry the risk of privacy leakage.Therefore,one of the current challenges is determining how to perform vehicle trajectory clustering while protecting user privacy.We propose a privacy-preserving vehicle trajectory clustering framework and construct a vehicle trajectory clustering model(IKV)based on the variational autoencoder(VAE)and an improved K-means algorithm.In the framework,the client calculates the hidden variables of the vehicle trajectory and uploads the variables to the server;the server uses the hidden variables for clustering analysis and delivers the analysis results to the client.The IKV’workflow is as follows:first,we train the VAE with historical vehicle trajectory data(when VAE’s decoder can approximate the original data,the encoder is deployed to the edge computing device);second,the edge device transmits the hidden variables to the server;finally,clustering is performed using improved K-means,which prevents the leakage of the vehicle trajectory.IKV is compared to numerous clustering methods on three datasets.In the nine performance comparison experiments,IKV achieves optimal or sub-optimal performance in six of the experiments.Furthermore,in the nine sensitivity analysis experiments,IKV not only demonstrates significant stability in seven experiments but also shows good robustness to hyperparameter variations.These results validate that the framework proposed in this paper is not only suitable for privacy-conscious production environments,such as carpooling tasks,but also adapts to clustering tasks of different magnitudes due to the low sensitivity to the number of cluster centers.

Improved algorithm of cluster-based routing protocols for agricultural wireless multimedia sensor networks

Low Energy Adaptive Clustering Hierarchy(LEACH)is a routing algorithm in agricultural wireless multimedia sensor networks(WMSNs)that includes two kinds of improved protocol,LEACH_D and LEACH_E.In this study,obstacles were overcome in widely used protocols.An improved algorithm was proposed to solve existing problems,such as energy source restriction,communication distance,and energy of the nodes.The optimal number of clusters was calculated by the first-order radio model of the improved algorithm to determine the percentage of the cluster heads in the network.High energy and the near sink nodes were chosen as cluster heads based on the residual energy of the nodes and the distance between the nodes to the sink node.At the same time,the K-means clustering analysis method was used for equally assigning the nodes to several clusters in the network.Both simulation and the verification results showed that the survival number of the proposed algorithm LEACH-ED increased by 66%.Moreover,the network load was high and network lifetime was longer.The mathematical model between the average voltage of nodes(y)and the running time(x)was concluded in the equation y=−0.0643x+4.3694,and the correlation coefficient was R2=0.9977.The research results can provide a foundation and method for the design and simulation of the routing algorithm in agricultural WMSNs.

基于改进引力搜索的混合K-调和均值聚类算法研究

为了解决聚类算法容易陷入局部最优的问题,以及增强聚类算法的全局搜索能力,基于KHM算法以及改进的引力搜索算法,提出一种混合K-调和均值聚类算法(G-KHM)。G-KHM算法具有KHM算法收敛速度快的优点,但同时针对KHM算法容易陷入局部最优解的问题,在初始化后数据开始搜索聚类中心时采用了一种基于对象多样性及收敛性增强的引力搜索算法,该方法改进了引力搜索算法容易失去种群多样性的缺点,并同时具有引力搜索算法较强的全局搜索能力,可以使算法收敛到全局最优解。仿真结果表明,G-KHM算法能有效地避免陷入局部极值,具有较强的全局搜索能力以及稳定性,并且相比KHM算法、K-means聚类算法、C均值聚类算法以及粒子群算法,在分类精度和运行时间上表现出了更好的效果。

融合改进PSO和K-调和均值的混合聚类算法

为了提高聚类算法的全局搜索能力,提出了一种融合改进的粒子群算法(IPSO)和K-调和均值聚类(KHM)的混合聚类算法(IPSO-KHM)。该算法为了改进PSO算法容易陷入局部最优的缺点,提出了一种粒子突变策略,根据粒子分布密集程度及粒子在当前最优值附近的相对分布位置,通过移动低效粒子使之远离当前局部最优值,从而提高粒子全局搜索的效率,避免陷入局部最优。实证分析结果表明:IPSO-KHM算法的聚类效果、收敛速度、分类精度等性能优于其他算法。

基于改进主成分和全局k均值聚类的汽车行驶工况构建

为解决传统聚类算法构建工况初始中心易陷入局部最优、执行时耗长的问题,提出了一种改进全局k均值聚类(improved global k-means clustering,IGKM)算法,以缩小作为候选下一簇的初始中心点集,减少算法的迭代次数;采用小波分层阈值降噪和小波分解域量化压缩对原始数据进行预处理,结合改进主成分分析(improved principal component analysis,IPCA)对片段进行降维和分类;最后,合成汽车行驶工况。实验结果表明,所提方法构建行驶工况的速度-加速度联合分布差异值仅为0.87%,聚类平均耗时仅为83.35 s,行驶工况拟合度较高,更能综合反映实际车辆的运行状况。

融合云模型优化萤火虫的K-mediods聚类算法

文章针对K-中心点聚类算法(K-mediods)易陷入局部最优及运行代价过大的问题,提出一种融合云模型优化萤火虫的K-mediods聚类算法。首先,将基于优秀萤火虫的云模型优化策略与基于普通萤火虫的云动态调整因子策略以及自主随机搜索相结合,对基本的萤火虫优化算法(GSO)进行改进;其次,从全局收敛性的角度对改进的GSO进行分析;最后,将改进的GSO与K-mediods算法融合成一种新的K-mediods算法。实验结果表明,该算法不仅在4种测试函数的求解精度上效果更优,而且对5个数据集的聚类结果均有改善,有效地抑制了K-mediods算法易陷入局部最优的问题,并且减少了算法的运行时间。

Fault Diagnosis in a Hydraulic Position Servo System Using RBF Neural Network

Considering the nonlinea r, time-varying and ripple coupling properties in the hydraulic servo system, a two-stage Radial Basis Function （RBF） neural network model is proposed to realize the failure detection and fault localization. The first-stage RBF neural network is adopted as a failure observer to realize the failure detection. The trained RBF observer, working concurrently with the actual system, accepts the input voltage signal to the servo valve and the measurements of the ram displacements, rebuilds the system states, and estimates accurately the output of the system. By comparing the estimated outputs with the actual measurements, the residual signal is generated and then analyzed to report the occurrence of faults. The second-stage RBF neural network can locate the fault occurring through the residual and net parameters of the first-stage RBF observer. Considering the slow convergence speed of the K-means clustering algorithm, an improved K-means clustering algorithm and a self-adaptive adjustment algorithm of learning rate arc presented, which obtain the optimum learning rate by adjusting self-adaptive factor to guarantee the stability of the process and to quicken the convergence. The experimental results demonstrate that the two-stage RBF neural network model is effective in detecting and localizing the failure of the hydraulic position servo system.

基于改进层次全局模糊熵和MCFS的滚动轴承损伤识别

针对传统的多尺度特征提取方法无法捕捉振动信号高频故障信息的问题,提出了一种基于改进层次全局模糊熵(IHGFE)全局全频段特征提取、多聚类特征选择(MCFS)特征降维和支持向量机分类的滚动轴承故障诊断方法。首先,提出了能够捕捉振动信号低频到高频的全局特征的IHGFE非线性动力学方法,并将其用于滚动轴承的故障特征提取;然后,利用MCFS对初始特征向量进行了维数约简和优化,构建了低维且对故障敏感的故障特征向量;最后,建立了基于支持向量机的多故障分类器,实现了滚动轴承损伤的智能化识别,并通过两个滚动轴承实验进行了对比分析。研究结果表明:IHGFE的分类准确率和识别稳定性均优于对比方法,证明了其在特征提取中能够在一定程度上解决现有方法无法同时考虑信号的高频特征和全局特征的问题,可为进一步扩展模糊熵方法在滚动轴承损伤识别中的应用提供参考。

一种基于全局K-均值聚类的改进算法

全局K-均值聚类算法需要随机选取初始的聚类中心,本文基于K中心点算法的思想,将其作为全局K-均值聚类算法的初始聚类中心,并对全局K-均值聚类算法进行改进。依托人工模拟数据和学习库中的数据分析,对比两种算法的性能,得出改进算法聚类时间短,鲁棒性强的结论。