基于模糊理论的不确定轨迹模式挖掘

轨迹模式挖掘是轨迹数据知识发现的一个重要课题.由于定位设备的局限性,采集到的轨迹数据具有不确定性.着眼于不确定轨迹数据模式挖掘,通过引入模糊集方法,提出不确定轨迹数据模式挖掘方法.首先用均匀网格划分对象的运动平面,基于对象实际位置的概率分布计算轨迹点相对于每个网格的隶属度,通过不确定轨迹兴趣区域发现算法,得到所有的不确定兴趣区域,然后在此基础上进行轨迹模式挖掘.实验展示了所提出的不确定轨迹数据模式挖掘方法进行模式挖掘的效果.     

区间值Vague决策系统及其规则提取方法

区间值(i-v)模糊集合和Vague集是不精确知识表达的两种新理论.它们已被广泛地应用于决策系统中对于不确定决策数据的描述.本文将两者有机地结合,给出了i-v Vague集的有关概念,并详细地讨论了其性质.最后,结合一个决策系统实例,利用文中所提出的包含与相交因子,分析了规则生成策略.经过与传统Rough集方法的对比,发现两者所得结论是一致的,但该方法对于不确定性问题的处理却表现得更为灵活.     

A Novel Approach for Mining High‐Utility Sequential Patterns in Sequence Databases

Mining sequential patterns is an important research issue in data mining and knowledge discovery with broad applications. However, the existing sequential pattern mining approaches consider only binary frequency values of items in sequences and equal importance/significance values of distinct items. Therefore, they are not applicable to actually represent many real‐world scenarios. In this paper, we propose a novel framework for mining high‐utility sequential patterns for more real‐life applicable information extraction from sequence databases with non‐binary frequency values of items in sequences and different importance/significance values for distinct items. Moreover, for mining high‐utility sequential patterns, we propose two new algorithms: UtilityLevel is a high‐utility sequential pattern mining with a level‐wise candidate generation approach, and UtilitySpan is a high‐utility sequential pattern mining with a pattern growth approach. Extensive performance analyses show that our algorithms are very efficient and scalable for mining high‐utility sequential patterns.     

Probe generation for active probing

Active probing is a widely adopted approach for developing effective solutions for network monitoring and diagnosing. However, the use of probing techniques incurs costs in terms of additional network traffic. Furthermore, probing stations are required to be configured and maintained in the network for sending out probes. The set of probes used for fault detection and/or diagnosis (called the target probe set) is selected by a probe selection algorithm from a larger set called the candidate probe set. Most of the existing techniques for selecting the target probe set assume that the candidate probe set will preexist and the set is determined by the configured routing model in the network. In this paper, we address the problem of generating an expanded candidate probe set, which results in the selection of a more efficient target probe set. We propose the use of heuristics and network partitioning strategies for generating the candidate probe set. For evaluating our approach, we perform experiments to generate candidate probe sets for the networks of several types and sizes. The candidate probe sets are used by the existing probe selection algorithms for selecting target probe sets for fault detection and localization. Our results demonstrate that the target probe set selected from the candidate probe set generated using our approach has a reduced cost of monitoring the network.     

New method of mining incomplete data

The data used in the process of knowledge discovery often includes noise and incomplete information. The boundaries of different classes of these data are blur and unobvious. When these data are clustered or classified, we often get the coverings instead of the partitions, and it usually makes our information system insecure. In this paper, optimal partitioning of incomplete data is researched. Firstly, the relationship of set cover and set partition is discussed, and the distance between set cover and set partition is defined. Secondly, the optimal partitioning of given cover is researched by the combing and parting method, acquiring the optimal partition from three different partitions set family is discussed. Finally, the corresponding optimal algorithm is given. The real wireless signals offten contain a lot of noise, and there are many errors in boundaries when these data is clustered based on the tradional method. In our experimant, the proposed method improves correct rate greatly, and the experimental results demonstrate the method’s validity.     

Estimating the Position of Mobiles by Multi‐Criteria Decision Making

In this study, we propose a novel mobile tracking method based on Multi‐Criteria Decision Making (MCDM), in which uncertain parameters—the received signal strength, the distance between the mobile and the base station, the moving direction, and the previous location—are used in the decision process using the aggregation function in the fuzzy set theory. Through numerical results, we show that our proposed mobile tracking method provides a better performance than the conventional method using the received signal strength.     

IMIS中基于Rough集的知识推理

粗糙集理论作为智能信息处理技术的一个新成果,是对不完整数据进行分析的新方法。毫无疑问,它是除模糊集理论外的又一种最具挑战性的领域,也是现今计算机应用中一个新的,非常重要且发展迅速的研究和应用领域。粗糙集是知识发现、数据约简、决策支持、分类、模式识别和控制等领域中新的,有效数学方法。首先对粗糙集基本概念和理论进行了介绍,给出了基于粗糙集的知识推理方法。然后以一个实际例子讨论如何利用粗集理论从现有学生综合素质数据库中进行知识约简,获取新决策规则,并使之用于对学生的操行评定。     

数字图像模糊滤波盲取证算法

数字图像模糊滤波操作常用于美化润饰“伪造”图像.针对常用的均值滤波、空域高斯低通滤波与中值滤波,提出了一种能同时检测上述3种操作的盲取证算法.首先将高频残差作为特征提取域,然后分别基于二值局部模式LBP和自回归模型提取特征,最后使用支持向量机构造模糊滤波检测器.实验结果表明,所提算法能有效地检测模糊滤波操作,对抗JPEG压缩的鲁棒性能较好.     

Exploiting data preparation to enhance mining and knowledgediscovery

One of the major obstacles to using organizational data for mining and knowledge discovery is that, in most cases, it is not amenable for mining in its natural form. Using a data set from a large tertiary-care hospital, we provide strong empirical evidence that data enhancement by the introduction of new attributes, along with judicious aggregation of existing attributes, results in higher-quality knowledge discovery. Interestingly, we also found that there is a differential impact of data set enhancements on the performance of different data mining algorithms. We define and use several measures, including entropy, rule complexity and resonance, to evaluate the quality and usefulness of the knowledge discovered     

Design of a fuzzy traffic controller for ATM networks

This paper presents the design of a fuzzy traffic controller that simultaneously manages congestion control and call admission control for asynchronous transfer mode (ATM) networks. The fuzzy traffic controller is a fuzzy implementation of the two-threshold congestion control method and the equivalent capacity admission control method extensively studied in the literature. It is an improved, intelligent implementation that not only utilizes the mathematical formulation of classical control but also mimics the expert knowledge of traffic control. We appropriately choose input linguistic variables of the fuzzy traffic controller so that the controller is a closed-loop system with stable and robust operation. We extract knowledge of conventional control methods from numerous analytical data using a clustering technique and then use this knowledge to set parameters of the membership functions and fuzzy control rules via fuzzy set manipulation (linguistically stated but mathematically treated) with the aid of an optimization technique named genetic algorithm (GA). Simulation results show that the proposed fuzzy admission control improves system utilization by a significant 11%, while maintaining the quality of service (QoS) contract comparable with that of the conventional equivalent capacity method. The performance of the proposed fuzzy congestion control method is also 4% better than that of the conventional two-threshold congestion control method     

图像分割中的交叉熵和模糊散度算法

本文将交叉熵和模糊散度应用于图像分割中,提出了中最优灰度值选取算法,其一是基于均匀分布假设的最小交叉熵算法,其二是利用后难概率的最大类间交叉熵算法,其三是类间最大模糊散度的改进算法,其四是最小模糊散度算法,针对图像阈什化分割的要求,在后两种算法中构造一种新的模糊录改度函数,本文采用均匀测试和开头测试比较各算法的性能,利用多种类型测试 是到的分割结果,显示了所筛算法的有效性和通用性。     

Knowledge bounded least squares method for the identification of fuzzy systems

This paper presents the knowledge bounded least squares method that uses both linguistic information (i.e., human knowledge and experience) and numerical data to identify fuzzy models. Based on the concept of fuzzy interval systems, the basic idea of this method is: first, to utilize all the available linguistic information to obtain a fuzzy interval system and then use the obtained fuzzy interval system to give the admissible model set (i.e., the set of all fuzzy models which are acceptable and reasonable from the point of view of linguistic information); second, to find a fuzzy model in the admissible fuzzy model set which best fits the available numerical data. It is shown that such a fuzzy model can be obtained by a quadratic programming approach. By comparing this method with the least squares method, it is proved that the fuzzy model obtained by the proposed method fits the real model better than the fuzzy model obtained by the least squares method.     

Incorporating domain knowledge into the fuzzy connectedness framework: application to brain lesion volume estimation in multiple sclerosis

A method for incorporating prior knowledge into the fuzzy connectedness image segmentation framework is presented. This prior knowledge is in the form of probabilistic feature distribution and feature size maps, in a standard anatomical space, and "intensity hints" selected by the user that allow for a skewed distribution of the feature intensity characteristics. The fuzzy affinity between pixels is modified to encapsulate this domain knowledge. The method was tested by using it to segment brain lesions in patients with multiple sclerosis, and the results compared to an established method for lesion outlining based on edge detection and contour following. With the fuzzy connections (FC) method, the user is required to identify each lesion with a mouse click, to provide a set of seed pixels. The algorithm then grows the features from the seeds to define the lesions as a set of objects with fuzzy connectedness above a preset threshold. The FC method gave improved interobserver reproducibility of lesion volumes, and the set of pixels determined to be lesion was more consistent compared to the contouring method. The operator interaction time required to evaluate one subject was reduced from an average of 111 min with contouring to 16 min with the FC method.     

用隶属函数法建立多光谱导航模型

从图像的光谱特征和利用模糊数学进行图像的阈值选取两个角度来分析和探讨多光谱图像导航方法,引入模式识别、模糊集合等数学理论,建立了一种多光谱导航新方法一一隶属函数法,实现对飞行器的光谱导航,通过小车的实验对图像信息的处理作了初步的尝试,并给出实验结果和分析．     

A Novel Color Image Segmentation Approach Based on Neutrosophic Set and Modified Fuzzy c-Means

Color image segmentation is an important technique in image processing, pattern recognition and computer vision. Many segmentation algorithms have been proposed. However, it is still a complex task especially when there are noises in the images, which have not been studied in much detail. Neutrosophic set (NS) studies the origin, nature, and scope of neutralities. In this paper, we apply NS in the color image and define some new concepts. A directional α-mean operation is proposed to reduce the set indeterminacy. The fuzzy c-means clustering method is improved by integrating with NS and employed for the color image segmentation. The computation of membership and the clustering termination criterion are redefined accordingly. Moreover, a validity criterion is employed to determine the optimal clustering number. Numerical experiments serve to illustrate the effectiveness and reliability of the proposed approach. Experimental results demonstrate that our approach can segment color images automatically and effectively, produce good results as favorably compared to some existing algorithms. The optimal clustering number is determined automatically and no prior knowledge is required. Especially, it can segment both images with the simple and distinct objects and the images with complex and noisy objects, which is the most difficult task for color image segmentation.     

Model predictive controller design for boost DC–DC converter using T–S fuzzy cost function

This paper proposes a Takagi–Sugeno (T–S) fuzzy method to select cost function weights of finite control set model predictive DC–DC converter control algorithms. The proposed method updates the cost function weights at every sample time by using T–S type fuzzy rules derived from the common optimal control engineering knowledge that a state or input variable with an excessively large magnitude can be penalised by increasing the weight corresponding to the variable. The best control input is determined via the online optimisation of the T–S fuzzy cost function for all the possible control input sequences. This paper implements the proposed model predictive control algorithm in real time on a Texas Instruments TMS320F28335 floating-point Digital Signal Processor (DSP). Some experimental results are given to illuminate the practicality and effectiveness of the proposed control system under several operating conditions. The results verify that our method can yield not only good transient and steady-state responses (fast recovery time, small overshoot, zero steady-state error, etc.) but also insensitiveness to abrupt load or input voltage parameter variations.     

A modular approach to facial feature segmentation on real sequences

In this paper a modular approach of gradual confidence for facial feature extraction over real video frames is presented. The problem is being dealt under general imaging conditions and soft presumptions. The proposed methodology copes with large variations in the appearance of diverse subjects, as well as of the same subject in various instances within real video sequences. Areas of the face that statistically seem to be outstanding form an initial set of regions that are likely to include information about the features of interest. Enhancement of these regions produces closed objects, which reveal—through the use of a fuzzy system—a dominant angle, i.e. the facial rotation angle. The object set is restricted using the dominant angle. An exhaustive search is performed among all candidate objects, matching a pattern that models the relative position of the eyes and the mouth. Labeling of the winner features can be used to evaluate the features extracted and provide feedback in an iterative framework. A subset of the MPEG-4 facial definition or facial animation parameter set can be obtained. This gradual feature revelation is performed under optimization for each step, producing a posteriori knowledge about the face and leading to a step-by-step visualization of the features in search.     

卷积神经网络（CNN）训练中卷积核初始化方法研究

本文提出了一种基于样本图像局部模式聚类的卷积核初始化方法,该方法可用于卷积神经网络（Convolutional neural network, CNN）训练中卷积核的初始化。在卷积神经网络中,卷积核的主要作用可看成是利用匹配滤波提取图像中的局部模式,并将其作为后续图像目标识别的特征。为此本文在图像训练集中选取一部分典型的样本图像,在这些图像中抽取与卷积核相同大小的子图作为图像局部模式矢量集合。首先对局部模式子图集合应用拓扑特性进行粗分类,然后对粗分类后的每一子类采用势函数聚类的方法获取样本图像中的典型局部模式子图,构成候选子图模式集,用它们作为CNN的初始卷积核进行训练。实验结果表明,本文方法可以明显加速CNN网络训练初期的收敛速度,同时对最终训练后的网络识别精度也有一定程度的提高。      

Solder joints inspection using a neural network and fuzzyrule-based classification method

In this paper, we described an approach in automation, the visual inspection of solder joint defects of surface mounted components on a printed circuit board, using a neural network with fuzzy rule-based classification method. Inherently, the solder joints have a curved, tiny, and specular reflective surface. This presents the difficulty in taking good images of the solder joints. Furthermore, the shapes of the solder joints tend to greatly vary with their soldering conditions, and are not identical with each other, even though some of the solder joints belong to a set of the same soldering quality. This problem makes it difficult to classify the solder joints according to their properties. To solve this intricate problem, a new classification method is here proposed which consists of two modules: one based upon an unsupervised neural network, and the other based upon a fuzzy set theory. The novel idea of this approach is that a fuzzy rule table reflecting the knowledge of criteria of a human inspector, is utilized in order to correct any possible misclassification made by the neural network module. The performance of the proposed approach was tested on numerous samples of printed circuit boards in commercially available computers, and then compared with that of a human inspector. Experimental results reveal that the proposed method is superior to the neural network classification method alone, in terms of its accuracy of classification     

一种肿瘤基因表达数据的知识提取方法

本文以多发性骨髓瘤的基因表达数据为例,利用数据挖掘技术,提出了一种针对基因表达数据进行知识发现的方法.该方法通过计算基因的信息增益,结合神经网络,找出了特征基因集合,最后利用决策树进行特征规则的提取,给出了基于多发性骨髓瘤数据样本的产生式规则,为生物医学研究提供了一种分析和研究基因表达数据的参考方法.实验结果表明了该方法的有效性.