一种用于神经网络样本划分的自聚类算法

建立神经网络模型时 ,能否合理地划分训练样本和检验样本直接关系到建模的效率 .在很多实际应用中 ,检验样本是随机抽取的 .本文提出了一种基于欧氏距离的自聚类算法 ,根据样本的空间分布情况对其自动分类 ,然后确定检验样本 .算例研究表明 ,应用此算法能够改善检验效果 ,从而提高建模效率 .

K-means改进算法在电力用户聚类辨识中的应用

科学、准确的用户用电特征分析对掌握负荷发展变化规律,提高电力需求预测的准确性,保障系统规划和经济运行具有重要意义。文中在对K-means算法深入研究的基础上,结合电力负荷数据海量、多维等特点,通过归一化处理,异常数据剔除,改进的二分K-means算法进行自聚类,对各优化算法进行分析,克服了传统K-means算法对异常数据敏感和初始聚类中心的随机性问题。实验结果表明,优化的自聚类算法能够提高分类的准确性,提高收敛效率,实现用户数据特征自动辨识分类。

Recognition of Spontaneous Combustion in Coal Mines Based on Genetic Clustering

Spontaneous combustion is one of the greatest disasters in coal mines. Early recognition is important because it may be a potential inducement for other coalmine accidents. However, early recognition is difficult because of the complexity of different coal mines. Fuzzy clustering has been proposed to incorporate the uncertainty of spontaneous combustion in coal mines and it can give a clear degree of classification of combustion. Because FCM clustering tends to become trapped in local minima, a new approach of fuzzy c-means clustering based on a genetic algorithm is there- fore proposed. Genetic algorithm is capable of locating optimal or near optimal solutions to difficult problems. It can be applied in many fields without first obtaining detailed knowledge about correlation. It is helpful in improving the effec- tiveness of fuzzy clustering in detecting spontaneous combustion. The effectiveness of the method is demonstrated by means of an experiment.

Behavior Clustering for Anomaly Detection

We presented a novel framework for automatic behavior clustering and unsupervised anomaly detection in a large video set. The framework consisted of the following key components： 1 ） Drawing from natural language processing, we introduced a compact and effective behavior representation method as a stochastic sequence of spatiotemporal events, where we analyzed the global structural information of behaviors using their local action statistics. 2） The natural grouping of behavior patterns was discovered through a novel clustering algorithm. 3 ） A run-time accumulative anomaly measure was introduced to detect abnormal behavior, whereas normal behavior patterns were recognized when sufficient visual evidence had become available based on an online Likelihood Ratio Test （LRT） method. This ensured robust and reliable anomaly detection and normal behavior recognition at the shortest possible time. Experimental results demonstrated the effectiveness and robustness of our approach using noisy and sparse data sets collected from a real surveillance scenario.

Turnout fault diagnosis based on DBSCAN/PSO-SOM

In order to diagnose the common faults of railway switch control circuit,a fault diagnosis method based on density-based spatial clustering of applications with noise(DBSCAN)and self-organizing feature map(SOM)is proposed.Firstly,the three-phase current curve of the switch machine recorded by the micro-computer monitoring system is dealt with segmentally and then the feature parameters of the three-phase current are calculated according to the action principle of the switch machine.Due to the high dimension of initial features,the DBSCAN algorithm is used to separate the sensitive features of fault diagnosis and construct the diagnostic sensitive feature set.Then,the particle swarm optimization(PSO)algorithm is used to adjust the weight of SOM network to modify the rules to avoid“dead neurons”.Finally,the PSO-SOM network fault classifier is designed to complete the classification and diagnosis of the samples to be tested.The experimental results show that this method can judge the fault mode of switch control circuit with less training samples,and the accuracy of fault diagnosis is higher than that of traditional SOM network.

KFL： a clustering algorithm for image database

It is a fairly challenging issue to make image repositories easy to be searched and browsed. This depends on a technique--image clustering. Kernel-based clustering algorithm has been one of the most promising clustering methods in the last few years, beeanse it can handle data with high dimensional complex structure. In this paper, a kernel fuzzy learning （KFL） algorithm is proposed, which takes advantages of the distance kernel trick and the gradient-based fuzzy clustering method to execute the image clustering automatically. Experimental results show that KFL is a more efficient method for image clustering in comparison with recent renorted alternative methods.

Interactive Protein Data Clustering

In this paper, the authors present three different algorithms for data clustering. These are Self-Organizing Map （SOM）, Neural Gas （NG） and Fuzzy C-Means （FCM） algorithms. SOM and NG algorithms are based on competitive leaming. An important property of these algorithms is that they preserve the topological structure of data. This means that data that is close in input distribution is mapped to nearby locations in the network. The FCM algorithm is an algorithm based on soft clustering which means that the different clusters are not necessarily distinct, but may overlap. This clustering method may be very useful in many biological problems, for instance in genetics, where a gene may belong to different clusters. The different algorithms are compared in terms of their visualization of the clustering of proteomic data.

Research on natural language recognition algorithm based on sample entropy

Sample entropy can reflect the change of level of new information in signal sequence as well as the size of the new information. Based on the sample entropy as the features of speech classification, the paper firstly extract the sample entropy of mixed signal, mean and variance to calculate each signal sample entropy, finally uses the K mean clustering to recognize. The simulation results show that： the recognition rate can be increased to 89.2% based on sample entropy.

Adaptive Clustering Algorithm by Ants＇ Optimization

Inspired by the swarm intelligence in self-organizing behavior of real ant colonies, various ant-based algorithms were proposed recently for many research fields in data mining such as clustering. Compared with the previous clustering approaches such as K-means, the main advantage of ant-based clustering algorithms is that no additional information is needed, such as the initial partitioning of the data or the number of clusters. In this paper, we present an adaptive ant clustering algorithm ACAD. The algorithm uses a digraph where the vertexes represent the data to be clustered. The weighted edges represent the acceptance rate between the two data it connected. The pheromone on the edges is adaptively updated by the ants passing it. Some edges with less pheromone are progressively removed under a threshold in the process. Strong connected components of the final digraph are extracted as clusters. Experimental results on several real datasets and benchmarks indicate that ACAD is conceptually simpler, more efficient and more robust than previous research such as the classical K-means clustering algorithm and LF algorithm which.is also based on ACO