基于双向神经网络的聚类器

文章提出一种基于双向神经网络的不同粒度的聚类器，它具有算法简单、权系数无需计算和软件模拟效果好等优点，是一种较好的聚类算法.

基于信息熵更新权重的数据自适应聚类研究

数据聚类过程不稳定且容易出现概念漂移等问题,降低了聚类的准确性。针对当前聚类方法存在的漂移点筛选能力和抗干扰能力较弱的问题,设计了一种基于信息熵更新权重的数据自适应聚类方法。采用IEWU算法求得信息熵值,将信息熵分为数值型与分类型,分别对不同类型的信息熵进行属性加权。设计基础聚类器,分别更新基础聚类器以及信息熵权重,利用聚类决策完成了数据自适应聚类。实验结果表明,所提方法具有较强的漂移点筛选能力和抗干扰能力,聚类准确率始终高于90%,聚类效果更好。

基于数据集特点的增强聚类集成算法

当前流行的聚类集成算法无法依据不同数据集的不同特点给出恰当的处理方案,为此提出一种新的基于数据集特点的增强聚类集成算法,该算法由基聚类器的生成、基聚类器的选择与共识函数构成。该算法依据数据集的特点,通过启发式方法,选出合适的基聚类器,构建最终的基聚类器集合,并产生最终聚类结果。实验中,对ecoli,leukaemia与Vehicle三个基准数据集进行了聚类,所提出算法的聚类误差分别是0.014,0.489,0.479,同基于Bagging的结构化集成(BSEA)、异构聚类集成(HCE)和基于聚类的集成分类(COEC)算法相比,所提出算法的聚类误差始终最低;而在增加候基聚类器的情况下,所提出算法的标准化互信息(NMI)值始终高于对比算法。实验结果表明,同对比的聚类集成算法相比,所提出算法的聚类精度最高,可伸缩性最强。

基于聚类集成的微博话题发现方法

微博中短文本、用语不规范和大量噪音等特性使得传统话题发现方法不能很好地从中获取新话题。针对微博以上特性和话题动态性提出一种基于聚类集成的微博话题发现方法,该方法考虑微博发布的非线性时间因子,采用改进的K-Means方法分别融合微博的各个特性构造其对应的基聚类器,并评估各基聚类器之间的有效性和差异性,以此设置集成投票权值并最终进行聚类集成。实验对比结果表明,该方法将微博发现话题的准确性提升约9.5%,能够更有效地探测到新话题。

基于深度数据挖掘的高校信息数据整合方法

针对高效信息数据整合方法中未能使用Z-score算法对信息数据进行标准化处理,导致信息数据整合时存在整合精度低、时间长等问题,提出基于深度数据挖掘的高校信息数据整合方法。首先依据Z-score算法标准化信息数据,并依据标准化结果提取信息数据特征;再使用深度数据挖掘方法中的K-means算法计算数据特征生成基聚类器;最后将信息数据放入基聚类器中进行聚类;实现信息数据的整合。实验结果表明,运用该方法进行数据整合时,整合精度高、整合时长较短。

Study on Grading Standard for One-Year-Old Container Seedling Quality of Phoebe zhennan

In order to improve the survival rate of planting seedlings of Phoebe zhen-nan, the grading standard for one-year-old container seedlings of Phoebe zhennan was developed by using cluster analysis. The results showed that the quality of Phoebe zhennan container seedlings could be estimated from seedling height and ground diameter. The Phoebe zhennan container seedlings were divided into 3 grades： Grade 1 （seedling height ≥ 38 cm; ground diameter ≥ 0.65 cm）, Grade 2 （31.7 cm ≤ seedling height 〈 38 cm; 0.56 cm ≤ ground diameter 〈 0.65 cm） and Grade 3 （seedling height 〈 31.7 cm; ground diameter 〈 0.56 cm）.

Electron Injection Enhancement by Diamond-Like Carbon Film in Polymer Electroluminescence Devices

A diamond-like carbon （DLC） film is deposited as an electron injection layer between the polymer light-emitting layer（MEH-PPV） and aluminum （Al） cathode electrode in polymer electroluminescence devices （PLEDs） using a radio frequency plasma deposition system. The source material of the DLC is n-butylamine. The devices consist of indium tin oxide （ITO）/MEH-PPV/DLC/Al. Electron injection properties are investigated through I-V characteristics,and the mechanism of electron injection enhancement due to a thin DLC layer has been studied. It is found that： （1） a DLC layer thinner than 1.0nm leads to a higher turn-on voltage and decreased electroluminescent （EL） efficiency; （2） a 5.0nm DLC layer significantly enhances the electron injection and results in the lowest turn-on voltage and the highest EL efficiency; （3） DLC layer that exceeds 5.0nm results in poor device performance;and（4） EL emission can hardly be detected when the layer exceeds 10.0nm. The properties of ITO/MEH-PPV/DLC/Al and ITO/MEH-PPV/LiF/Al are investigated comparatively.

Quantum Information Splitting of an Arbitrary Three-Qubit State by Using Cluster States and Bell States

A new application of cluster states is investigated for quantum information splitting （QIS） of an arbitrary three-qubit state. In our scheme, a four-qubit cluster state and a Bell state are shared by a sender （Alice）, a controller （Charlie）, and a receiver （Bob）. Both the sender and controller only need to perform Bell-state measurements （BSMs）, the receiver can reconstruct the arbitrary three-qubit state by performing some appropriately unitary transformations on his qubits after he knows the measured results of both the sender and the controller. This QIS scheme is deterministic.

A Cooperative MIMO Transmission Scheme for Cluster-based Wireless Sensor Networks

Wireless sensor network is becoming more and more popular in recent years, but energy- constrained characteristic of sensor nodes is one of the critical issues that we must consider in system design. In this paper, a cluster-based virtual VBLAST transmission scheme is proposed to achieve energy savings for energy-constrained wireless sensor networks. In the proposed scheme, instead of using cluster member as cooperative nodes, multiple cluster heads cooperate to form virtual antenna array so that V-BLAST based virtual MIMO transmission can be implemented. Based on the communication energy consumption model, a way to optimize the parameters for the scheme is given. In addition, detailed simulation is performed to evaluate the performance of the proposed scheme for both densely and sparsely deployed sensor networks. Theoretical analysis and simulation results verify the energy efficiency of the proposed scheme.

Identifying Anomaly Aircraft Trajectories in Terminal Areas Based on Deep Autoencoder and Its Application in Trajectory Clustering

Anomalous trajectory detection and traffic flow classification for complicated airspace are of vital importance to safety and efficiency analysis.Some researchers employed density-based unsupervised machine learning method to exploit these trajectories related to air traffic control(ATC)actions.However,the quality of position data and the tiny density difference between traffic flows in the terminal area make it particularly challenging.To alleviate these two challenges,this paper proposes a novel framework which combines robust deep auto-encoder(RDAE)model and density peak(DP)clustering algorithm.Specifically,the RDAE model is utilized to reconstruct denoising trajectory and identify anomaly trajectories in the terminal area by two different regularizations.Then,the nonlinear components captured by the encoder of RDAE are input in the DP algorithm to classify the global traffic flows.An experiment on a terminal airspace at Guangzhou Baiyun Airport(ZGGG)with anomaly label shows that the proposed combination can automatically capture non-conventional spatiotemporal traffic patterns in the aircraft movement.The superiority of RDAE and combination are also demonstrated by visualizing and quantitatively evaluating the experimental results.

Reliability Analysis of Sensor Networks

To Integrate the capacity of sensing, communication, computing, and actuating, one of the compelling technological advances of these years has been the appearance of distributed wireless sensor network (DSN) for information gathering tasks. In order to save the energy, multi-hop routing between the sensor nodes and the sink node is necessary because of limited resource. In addition, the unpredictable conditional factors make the sensor nodes unreliable. In this paper, the reliability of routing designed for sensor network and some dependability issues of DSN, such as MTTF (mean time to failure) and the probability of connectivity between the sensor nodes and the sink node are analyzed. Unfortunately, we could not obtain the accurate result for the arbitrary network topology, which is #P-hard problem. And the reliability analysis of restricted topologies clustering-based is given. The method proposed in this paper will show us a constructive idea about how to place energy-constrained sensor nodes in the network efficiently from the prospective of reliability.

Automatic Mid-Level Concepts Clusteringfor Violent Movie Scenes Detection

This paper presents a novel system for violent scenes detection, which is based on machine learning to handle visual and audio features. MKL （Multiple Kernel Learning） is applied so that multimodality of videos can be maximized. The largest features of our system is that mid-level concepts clustering is proposed and implemented in order to learn mid-level concepts implicitly. By this algorithm, our system does not need manually tagged annotations. The whole system is trained on the dataset from MediaEval 2013 Affect Task and evaluated by its official metric. The obtained results outperformed its best score.

Free clustering optimal particle probability hypothesis density(PHD) filter

As to the fact that it is difficult to obtain analytical form of optimal sampling density and tracking performance of standard particle probability hypothesis density(P-PHD) filter would decline when clustering algorithm is used to extract target states,a free clustering optimal P-PHD(FCO-P-PHD) filter is proposed.This method can lead to obtainment of analytical form of optimal sampling density of P-PHD filter and realization of optimal P-PHD filter without use of clustering algorithms in extraction target states.Besides,as sate extraction method in FCO-P-PHD filter is coupled with the process of obtaining analytical form for optimal sampling density,through decoupling process,a new single-sensor free clustering state extraction method is proposed.By combining this method with standard P-PHD filter,FC-P-PHD filter can be obtained,which significantly improves the tracking performance of P-PHD filter.In the end,the effectiveness of proposed algorithms and their advantages over other algorithms are validated through several simulation experiments.

Energy Consumption in Wireless Sensor Networks

Wireless sensor networks （WSNs） can be used to collect surrounding data by multi-hop. As sensor networks have the constrained and not rechargeable energy resource, energy efficiency is an important design issue for its topology. In this paper, the energy consumption issue under the different topology is studied. We derive the exact mathematical expression of energy consumption for the fiat and clustering scheme, respectively. Then the energy consumptions of different schemes are compared. By the comparison, multi-level clustering scheme is more energy efficient in large scale networks. Simulation results demonstrate that our analysis is correct from the view of prolonging the large-scale network lifetime and achieving more power reductions.

FLBS: Fuzzy lion Bayes system for intrusion detection in wireless communication network

An important problem in wireless communication networks (WCNs) is that they have a minimum number of resources, which leads to high-security threats. An approach to find and detect the attacks is the intrusion detection system (IDS). In this paper, the fuzzy lion Bayes system (FLBS) is proposed for intrusion detection mechanism. Initially, the data set is grouped into a number of clusters by the fuzzy clustering algorithm. Here, the Naive Bayes classifier is integrated with the lion optimization algorithm and the new lion naive Bayes (LNB) is created for optimally generating the probability measures. Then, the LNB model is applied to each data group, and the aggregated data is generated. After generating the aggregated data, the LNB model is applied to the aggregated data, and the abnormal nodes are identified based on the posterior probability function. The performance of the proposed FLBS system is evaluated using the KDD Cup 99 data and the comparative analysis is performed by the existing methods for the evaluation metrics accuracy and false acceptance rate (FAR). From the experimental results, it can be shown that the proposed system has the maximum performance, which shows the effectiveness of the proposed system in the intrusion detection.

Adaptive triangular mesh coarsening with centroidal Voronoi tessellations

We present a novel algorithm for adaptive triangular mesh coarsening. The algorithm has two stages. First, the input triangular mesh is refined by iteratively applying the adaptive subdivision operator that performs a so-called red-green split. Second, the refined mesh is simplified by a clustering algorithm based on centroidal Voronoi tessellations (CVTs). The accuracy and good quality of the output triangular mesh are achieved by combining adaptive subdivision and the CVTs technique. Test results showed the mesh coarsening scheme to be robust and effective. Examples are shown that validate the method.

C60-based clustering scheme for sensor management in STSS

Clustering-based sensor-management schemes have been widely used for various wireless sensor networks(WSNs), as they are well suited to the distributive and collaborative nature of WSN. In this paper, a C60-based clustering algorithm is proposed for the specific planned network of space tracking and surveillance system(STSS), where all the sensors are partitioned into 12 clusters according to the C60(or football surface) architecture, and then a hierarchical sensor-management scheme is well designed.Finally, the algorithm is applied to a typical STSS constellation,and the simulation results show that the proposed method has better target-tracking performance than the nonclustering scheduling method.

Efficient page layout analysis on small devices

Previously we have designed and implemented new image browsing facilities to support effective offiine image contents on mobile devices with limited capabilities： low bandwidth, small display, and slow processing. In this letter, we fulfill the automatic production of cartoon contents fitting small-screen display, and introduce a clustering method useful for various types of cartoon images as a prerequisite stage for preserving semantic meaning. The usage of neural networks is to properly cut the various forms of pages. Texture information that is useful for grayscale image segmentation gives us a good clue for page layout analysis using the multilayer perceptron （MLP） based x-y recursive algorithm. We also automatically frame the segment MLP using agglomerative segmentation. Our experimental results show that the combined approaches yield good results of segmentation for several cartoons.