Online Trends

Research on the parallel Affective classification algorithm Yu Yonghong to the June Shanglin in the mass data set to perform the emotion classification task, the traditional single Machine affective classification algorithm extensibility becomes the bottleneck of the system. On the cloud computing platform Hadoop, the mapreduce of algorithms such as feature extraction, feature vector weighting and emotion classification are realized in the Affective classification task. On the affective Corpus dataset, the precision of affective classification algorithm and the time cost of each algorithm are compared and analyzed. The experimental results verify the parallel ...

The MapReduce calculation model of brain-computer interface Huang Zhihua The time relationship between the computational tasks of brain-computer interface is studied, the computational process of brain-computer interface is divided into three stages, and the first two phases are mapped to the map and reduce of the mapreduce mechanism. A mapreduce calculation model of brain-computer interface is proposed. Because most of the computational volume of brain-computer interface is concentrated in the first two stages, this model can significantly reduce the computational time of brain-computer interface. MapReduce computational model of brain-computer interface

A parallel clustering model based on MapReduce Gurechun in the process of clustering massive data, the limitation of traditional serial mode is more and more obvious, it is difficult to get satisfactory result in the effective time, this paper proposes a parallel clustering model based on the MapReduce framework under the Hadoop platform. The theoretical and experimental results show that the model has an acceleration ratio of near linear velocity and high efficiency for mass data. A parallel clustering model based on MapReduce

Peer-to-peer traffic recognition in cloud computing environment Tankai Gaozhonger Li Fengban because of the memory limit, the Peer-to-peer traffic recognition method can only deal with the small-scale data set. Moreover, the attribute characteristics of naive Bayesian classification are artificial selection, so the recognition rate is limited and lacks objectivity. Based on the above problem analysis, this paper puts forward naive Bayesian classification algorithm in cloud computing environment and improves the attribute reduction algorithm in cloud computing environment, and combines these two algorithms to realize the fine granularity recognition of the encrypted peer-to-peer traffic. The experimental results show that the method can ...

Research on job scheduling algorithm under Hadoop platform Zhengzhou University Zhaoxiaobing The main content of this paper is the improvement of job scheduling algorithm on Hadoop. An improved algorithm (Btis) for task backup scheduling is proposed for late algorithm to estimate task schedule value and SAMR algorithm does not consider the problem of backup execution node. The Btis algorithm calculates the progress of the task accurately through the history record, finds the real slow task that needs to start the backup, and takes the work node to perform the task for the slow task when selecting the Fast node to start the backup ...

Research and realization of clustering and convex-package algorithm in MapReduce framework Chengdu University of Technology Zhaoju first, this paper makes a research on the generation and value growth of large data, and explains the necessity of improving the execution efficiency of the data mining algorithm, and introduces the technology and tools that support the large-data processing now. Then the paper studies the running mechanism of Hadoop file system, the stored procedure and the programming model of MapReduce framework, and the operation principle. Secondly, in a certain size of Hadoop cluster on the data distributed processing, so as to assess the whole cluster of sex ...

Parallel Feature Selection Based on MapReduce Zhanquan Sun in-monitors, a Parallel Feature Selection method Based on MapReduce Model is proposed. Large-scale dataset is partitioned into Sub-datas ...

On the change of the "time shift" rule of copyright law in the cloud computing environment in the Optus case Xu Yanbing the legality of cloud storage and cloud services has been judicially reviewed by jurisprudence before scholars concerned and legislative adjustments. The case for cloud computing, which took place in the United States, Singapore and Germany, was not the same as the April 2012 decision by the Australian court in the Optus case, which was not the same as in the Cablevision case of the Second Circuit of the United States in 2008, which affected the copyright law in the cloud computing environment. "Time shift" rule, "cloud service provider ..."

Dynamic adjustment of cloud computing virtual cluster based on multi-objective optimization Zhang Hao Shiling Sun Meng Tangshao The virtual cluster provided in the current cloud computing application can not be dynamically adjusted according to the actual network application load, which leads to the network virtual cluster resource utilization and load not adapt, so that more computing resources are idle in the cluster, Generate energy waste. In view of the above problems, a multi-objective optimization method of cloud computing virtual cluster Dynamic adjustment is proposed, and the dynamic adjustment framework of virtual cluster is given. Experimental results show that this method can automatically adjust the virtual cluster size to meet the network application load.

Based on MapReduce GML parallel query Xu Guan Hong The problem of massive spatial data query represented by the application of geo-markup Language (Geography Markup LANGUAGE,GML), A GML parallel Query method based on MapReduce is proposed. The query transformation of GML document query to GML spatial feature set query is realized by extracting GML spatial feature set, and the parallel query of spatial features is realized by MapReduce. Based on MapReduce ...

The improvement of PageRank algorithm based on cloud computing Zhengjing A new method is proposed to solve the problem that the PageRank algorithm needs a lot of iterations in the web structure mining. This method improves the calculation formula of the original PageRank value and reduces the number of iterations. Experiments show that the new method reduces the consumption of network communication and access HDFs in the cloud computing environment, and is superior to the traditional PageRank algorithm in time cost. Improvement of PageRank algorithm based on cloud computing

Research on the recommended algorithm of limited Boltzmann machine based on cloud Zhengzhi Yun Li Buyuan The exponential growth of the blunt data of Li Lun and the complexity of the algorithm itself make the limited Boltzmann machine face the problem of computational efficiency. Based on the detailed analysis of the restricted Boltzmann machine, the proposed algorithm of limited Boltzmann machine based on cloud platform is put forward by combining the parallel computing architecture of the limited Boltzmann machine and the Hadoop platform. The algorithm solves the problem of data relativity by copying mechanism, and decomposes the traditional limited Boltzmann process into several hadoo ...

On-line interactive application based on cloud computing and large data Khan Khudan Coke Dujian based on the characteristics of cloud computing and large data, this paper analyzes the problems faced by online interaction, constructs an online interactive platform model based on large data and cloud computing support, and focuses on the realization of the function of the platform, which enables users to achieve perfect online interactive activities. To achieve a high level of interaction, and then summed up the use of cloud computing and large data online interactive platform features and advantages. Research on online interactive application based on cloud computing and large data

Research and implementation of stochastic forest algorithm based on cloud computing platform Juan Wang Jianhua with the increasing popularity of mass data in the network era, it has become a hot research hotspot to excavate valuable information from it. In this paper, a distributed scalable random forest algorithm based on mass data is proposed for constructing random forest data mining algorithms. Based on MapReduce distributed computing model, this algorithm can efficiently analyze and process massive data, and can accurately classify and forecast the data. A lot of experimental results show that the algorithm presented in this paper has good ...

Real-time cloud infrastructure and virtualization data Graphic design with Intel architecture Harris Zeng Edwin Verplanke--Virtualization of communications applications--the latest advances in Intel's virtualization technology--virtualization support for Intel's data Surface development suite--real-time virtualization and improved platform service quality Real-time cloud infrastructure and virtualization data Graphic design with Intel architecture

Millet Hadoop&hbase Micro Practice Cheliang • Selection Basis Upstream important issue cluster check list some cases analysis • Some micro-improvement points and community feedback Millet Hadoop&hbase Micro Practice

Using MapReduce to build text indexes in Hadoop Shup Hadoop is an open-source, distributed system infrastructure that, with Hadoop, can develop distributed programs without understanding the distributed underlying details. Text indexing is widely used in production and life, and it is necessary to use text indexing from the search engine's inverted index to the operating system's instructions. Building text indexes in a Hadoop environment provides support for search engines and Full-text indexing of documents, while balancing the benefits of distributed systems. Build this ... in a hadoop environment.

The author of this paper, Qi Haijiang, Qingdao Five-Pulse Spring Information Co., Ltd. Technical director, University of Pennsylvania Bioengineering, Ph. D., Nanjing University. For many years engaged in graphic images, 3D vision, neural computing, machine learning algorithms such as research. "Abstract" cloud computing services is essentially a sharing of social intelligence resources, through the cloud of technology packets, reducing the difficulty threshold, so that more users can use "very advanced" technology. China's mobile interconnection new economy is highly prosperous, need to have the corresponding technology high cloud computing service as keel support. Today's computing is the obvious trend: Video audio graphics + ...

As the CIO of Iverson and the provider of IT services, CIO Frank wants to deliver better things to his internal customers that communicate cheaper and faster. "I am very confident that a better technical support can help people in many aspects of the business level," Accenture said, "but it needs to be better and cheaper." If you hit three more children, emphasize. It also began to become very noticeable. "Over the past few years, cloud computing has become very attractive to Essen. Frank's first attempt to enter the cloud was an IaaS enrollment system five years ago. Now, frank two degrees into the cloud application ...

Starting today, we will launch the Windows Azure AppFabric Introductory teaching article, from the beginning to the other, so that we can gradually understand and master Windows Azure AppFabric. Before we begin our study, we recommend that readers read first: AppFabric Chinese and English white papers. We know that APPFABIRC consists of service bus and Access Control service, and in order to use these features and features, as well as subsequent AP ...