Robust Dataset Classification Approach Based on Neighbor Searching and Kernel Fuzzy C-Means

Dataset classification is an essential fundament of computational intelligence in cyber-physical systems (CPS). Due to the complexity of CPS dataset classification and the uncertainty of clustering number, this paper focuses on clarifying the dynamic behavior of acceleration dataset which is achieved from micro electro mechanical systems (MEMS) and complex image segmentation. To reduce the impact of parameters uncertainties with dataset classification, a novel robust dataset classification approach is proposed based on neighbor searching and kernel fuzzy c-means (NSKFCM) methods. Some optimized strategies, including neighbor searching, controlling clustering shape and adaptive distance kernel function, are employed to solve the issues of number of clusters, the stability and consistency of classification, respectively. Numerical experiments finally demonstrate the feasibility and robustness of the proposed method. © 2014 Chinese Association of Automation.

NEW SHADOWED C-MEANS CLUSTERING WITH FEATURE WEIGHTS

Partition-based clustering with weighted feature is developed in the framework of shadowed sets. The objects in the core and boundary regions, generated by shadowed sets-based clustering, have different impact on the prototype of each cluster. By integrating feature weights, a formula for weight calculation is introduced to the clustering algorithm. The selection of weight exponent is crucial for good result and the weights are updated iteratively with each partition of clusters. The convergence of the weighted algorithms is given, and the feasible cluster validity indices of data mining application are utilized. Experimental results on both synthetic and real-life numerical data with different feature weights demonstrate that the weighted algorithm is better than the other unweighted algorithms.

Fuzzy c-means clustering based on spatial neighborhood information for image segmentation

Fuzzy c-means （FCM） algorithm is one of the most popular methods for image segmentation. However, the standard FCM algorithm is sensitive to noise because of not taking into account the spatial information in the image. An improved FCM algorithm is proposed to improve the antinoise performance of FCM algorithm. The new algorithm is formulated by incorporating the spatial neighborhood information into the membership function for clustering. The distribution statistics of the neighborhood pixels and the prior probability are used to form a new membership func- tion. It is not only effective to remove the noise spots but also can reduce the misclassified pixels. Experimental results indicate that the proposed algorithm is more accurate and robust to noise than the standard FCM algorithm.

Residual-driven Fuzzy C-Means Clustering for Image Segmentation

In this paper,we elaborate on residual-driven Fuzzy C-Means(FCM)for image segmentation,which is the first approach that realizes accurate residual(noise/outliers)estimation and enables noise-free image to participate in clustering.We propose a residual-driven FCM framework by integrating into FCM a residual-related regularization term derived from the distribution characteristic of different types of noise.Built on this framework,a weighted?2-norm regularization term is presented by weighting mixed noise distribution,thus resulting in a universal residual-driven FCM algorithm in presence of mixed or unknown noise.Besides,with the constraint of spatial information,the residual estimation becomes more reliable than that only considering an observed image itself.Supporting experiments on synthetic,medical,and real-world images are conducted.The results demonstrate the superior effectiveness and efficiency of the proposed algorithm over its peers.

Fingerprint image segmentation using modified fuzzy c-means algorithm

Fingerprint segmentation is a crucial step in fingerprint recognition system, and determines the results of fingerprint analysis and recognition. This paper proposes an efficient approach for fingerprint segmentation based on modified fuzzy c-means (FCM). The proposed method is realized by modifying the objective function in the Szilagyi’s algorithm via introducing histogram-based weight. Experimental results show that the proposed approach has an efficient performance while segmenting both original fingerprint image and fingerprint images corrupted by different type of noises.

Intelligent diagnosis of the solder bumps defects using fuzzy C-means algorithm with the weighted coefficients

Solder bump technology has been widely used in electronic packaging. With the development of solder bumps towards higher density and finer pitch, it is more difficult to inspect the defects of solder bumps as they are hidden in the package. A nondestructive method using the transient active thermography has been proposed to inspect the defects of a solder bump, and we aim at developing an intelligent diagnosis system to eliminate the influence of emissivity unevenness and non-uniform heating on defects recognition in active infrared testing. An improved fuzzy c-means(FCM) algorithm based on the entropy weights is investigated in this paper. The captured thermograms are preprocessed to enhance the thermal contrast between the defective and good bumps. Hot spots corresponding to 16 solder bumps are segmented from the thermal images. The statistical features are calculated and selected appropriately to characterize the status of solder bumps in FCM clustering. The missing bump is identified in the FCM result, which is also validated by the principle component analysis. The intelligent diagnosis system using FCM algorithm with the entropy weights is effective for defects recognition in electronic packages.

A weighted fuzzy C-means clustering method for hardness prediction

The hardness prediction model was established by support vector regression(SVR).In order to avoid exaggerating the contribution of very tiny alloying elements,a weighted fuzzy C-means(WFCM)algorithm was proposed for data clustering using improved Mahalanobis distance based on random forest importance values,which could play a full role of important features and avoid clustering center overlap.The samples were divided into two classes.The top 10 features of each class were selected to form two feature subsets for better performance of the model.The dimension and dispersion of features decreased in such feature subsets.Comparing four machine learning algorithms,SVR had the best performance and was chosen to modeling.The hyper-parameters of the SVR model were optimized by particle swarm optimization.The samples in validation set were classified according to minimum distance of sample to clustering centers,and then the SVR model trained by feature subset of corresponding class was used for prediction.Compared with the feature subset of original data set,the predicted values of model trained by feature subsets of classified samples by WFCM had higher correlation coefficient and lower root mean square error.It indicated that WFCM was an effective method to reduce the dispersion of features and improve the accuracy of model.

Research on Wind Power Prediction Modeling Based on Adaptive Feature Entropy Fuzzy Clustering

Wind farm power prediction is proposed based on adaptive feature weight entropy fuzzy clustering algorithm.According to the fuzzy clustering method,a large number of historical data of a wind farm in Inner Mongolia are analyzed and classified.Model of adaptive entropy weight for clustering is built.Wind power prediction model based on adaptive entropy fuzzy clustering feature weights is built.Simulation results show that the proposed method could distinguish the abnormal data and forecast more accurately and compute fastly.

A Novel Filtering-Based Detection Method for Small Targets in Infrared Images

Infrared small target detection technology plays a pivotal role in critical military applications,including early warning systems and precision guidance for missiles and other defense mechanisms.Nevertheless,existing traditional methods face several significant challenges,including low background suppression ability,low detection rates,and high false alarm rates when identifying infrared small targets in complex environments.This paper proposes a novel infrared small target detection method based on a transformed Gaussian filter kernel and clustering approach.The method provides improved background suppression and detection accuracy compared to traditional techniques while maintaining simplicity and lower computational costs.In the first step,the infrared image is filtered by a new filter kernel and the results of filtering are normalized.In the second step,an adaptive thresholding method is utilized to determine the pixels in small targets.In the final step,a fuzzy C-mean clustering algorithm is employed to group pixels in the same target,thus yielding the detection results.The results obtained from various real infrared image datasets demonstrate the superiority of the proposed method over traditional approaches.Compared with the traditional method of state of the arts detection method,the detection accuracy of the four sequences is increased by 2.06%,0.95%,1.03%,and 1.01%,respectively,and the false alarm rate is reduced,thus providing a more effective and robust solution.

基于FCM的快速模糊聚类算法研究

为改善FCM算法的运算性能、获得和原FCM算法等价的分类结果,本文提出了基于加权样本的fFCM(fast FCM)算法。此算法首先构造原待聚类集合的权集,并在权集上应用改进的FCM算法——WFCM(weighted FCM)算法快速获得和原FCM算法近似的分割结果;然后,将得到的分割结果作为FCM算法的初值再次利用FCM算法以获得最终的分割结果。理论证明和相关实验表明,fFCM不仅能获得和原FCM算法等价的分类结果,还具有良好的运算性能,具有广泛的适用性。

具有特征排序功能的鲁棒性模糊聚类方法

提出了一种加权模糊聚类算法,其优势在于能在实现有效聚类的同时,对样本噪音进行识别和按样本特征对聚类的贡献程度进行排序.因此,本文所提出的方法具有鲁棒性,并可对所得的特征排序进行特征选择,实验结果表明了该方法具有上述优势.

球磨机负荷加权模糊控制算法设计与仿真

针对目前球磨机料位单冲量控制系统响应速度慢、控制特性差等缺点,提出三冲量球磨机负荷加权模糊控制算法,即通过球磨机料位、球磨机进/出口温度和进出口压差三冲量模糊控制对球磨机负荷进行调节。在分析了球磨机的系统特性和控制目标后,结合电厂手动调节的特点,研究了三冲量模糊控制方案、负荷加权模糊控制器的结构以及模糊控制器的实现。利用Matlab/Simulink工具箱进行仿真研究.结果表明所提出的控制方法具有较好的动态特性和鲁棒性。

特征加权的模糊C有序均值聚类算法

Fuzzy C-ordered-means clustering(FCOM)算法基于排序进行模糊聚类,虽然其鲁棒性得到提高,但是耗时的排序操作降低了算法的效率。本文基于FCOM算法,将排序加权模式进行改进,提出一种特征加权的模糊C有序均值聚类算法(feature weighted fuzzy C-ordered-means clustering,FWFCOM)。为了验证算法的有效性,选取6个UCI数据集进行试验。结果表明,FWFCOM算法不仅在聚类准确率和鲁棒性方面有较好的表现,而且运行效率也得到有效提升。

一种新的带模糊权的粗糙聚类算法

针对粗糙聚类算法缺乏对数据比例变换的鲁棒性的问题,在粗糙聚类的框架下融合模糊聚类的思想,将临界区域中对象的模糊隶属度作为它们对于聚类中心调整的作用权值,得到一种带有模糊权的粗糙聚类算法(fuzzy we igh ing rough c lustering algorithm,FWRCA).实验表明,该算法不仅对于数据的比例变化具有鲁棒性,且在一定程度上克服了粗糙C均值聚类算法对划分阈值ε的敏感性,在性能上优于传统粗糙C均值聚类算法(如RCMCA),可应用于水电工程科学等以原型模型为研究手段并有大量需做比例变换的观测数据的领域.

模糊抗差自适应粒子滤波及其在组合导航中的应用

在研究模糊控制理论的基础上,吸收了粒子滤波、自适应滤波和抗差估计的优点,提出一种新的模糊抗差自适应粒子滤波算法。文中根据量测向量中的粗差对状态向量滤波的影响,建立抗差自适应粒子滤波模型,获得自适应因子,然后对滤波处理后的数据残差基于模糊理论构造等价权函数,利用等价权函数和自适应因子合理地分配信息,因而可以达到一般滤波方法无法达到的滤波精度,且能有效地控制粗差对导航解的影响。最后,将该算法应用到组合导航系统中,并进行仿真验证。仿真结果证明,文中提出的模糊抗差自适应粒子滤波算法的滤波精度相对于扩展Kalman滤波和粒子滤波提高了3至5倍,明显提高了导航定位精度,且计算简单,便于实时性计算。

鲁棒最小二乘支持向量机及其在软测量中的应用

针对最小二乘支持向量机在利用产生于工业现场的非理想数据集进行建模预测时,稀疏化模型鲁棒性差的问题,提出了一种基于模糊C均值聚类和密度加权的稀疏化方法.首先通过模糊C均值聚类将训练样本划分为若干个子类;然后计算每个子类中各样本的可能贡献度,依次从每个子类中选取具有最大可能贡献度的样本作为支持向量;最后更新每个样本的可能贡献度,继续从各个子集中增选支持向量,直至稀疏化后的模型性能满足要求.仿真结果和磨机负荷实际应用表明,该方法能够兼顾模型在整体样本集和各工况子集上的性能,在实现模型稀疏化的同时,能够显著改善最小二乘支持向量机模型的鲁棒性.

基于模糊隶属度的电力系统抗差估计

提出了一种电力系统模糊自适应抗差估计(fuzzy adaptive robust estimation,FARE)方法。计及了量测权重的不确定性,以连续的模糊隶属度评价测点的优劣,很好地解决了测点非优即劣的问题,以最小化测点劣质性的加权模糊隶属度之和为优化目标,采用原对偶内点法(Primal-Dual Interior Point Method,PDIPM)求解,并且实现了对量测粗差的自适应。多个IEEE标准算例以及波兰系统的仿真测试结果表明,该方法具有良好的抗差性能。

模糊树鲁棒回归算法的研究及其应用

针对实际工程中噪声难以避免和预测的问题,提出了鲁棒性较强的加权模糊树(W-FT)算法,采用基于局部异常因子(LOF)的加权最小二乘法代替最小二乘法学习模糊规则的后件参数,通过2个典型的非线性例子验证了该算法的有效性.应用W-FT算法建立了电站锅炉NOx排放特性模型,并与其他建模方法所建模型进行了对比.结果表明:所提出的W-FT算法能够有效地辨识噪声和异常值,具有较强的鲁棒性,所建立的模型预测精度较高,泛化能力较强.

不确定时滞模糊系统的时滞相关鲁棒H_∞控制

研究了一类带有时变时滞的不确定模糊系统时滞相关鲁棒H∞控制问题。基于模糊Lyapunov-Krasovskii泛函(LKF),引入多个模糊时滞自由权值矩阵,提出并证明了闭环系统新的时滞相关鲁棒H∞渐近稳定的充分条件。根据并行分布补偿算法(PDC)设计了反馈控制器,控制器可由线性矩阵不等式(LMI)求解得到。数例仿真验证了所提方法的有效性。

模糊残差算法对离群点数据的优化研究

离群点数据在数据分析中会影响回归模型的拟合精度,特别是高维不确定性离群点数据,通过模糊回归方法,可减少离群点对多维数据模型拟合的影响.本文提出模糊残差算法,首先在模糊域中构建模糊回归模型,其次迭代计算观测值和估计值之间的残差确定权重,通过最小化加权目标函数估计模型参数,得到基于加权优化的鲁棒模糊回归模型,最终算法优化的模型更精确拟合目标数据.仿真结果表明,在离群点影响情况下,与RTS-L1算法和FLAR算法相比,模糊残差算法在估计精度和鲁棒性方面有更好的提升.