基于GPU的多数据流相关系数并行计算方法研究

为了满足多数据流处理的实时性需求,提出一种跨PCIE总线的四层滑动窗口模型和基于图形处理器的多数据流并行处理框架模型,在此框架模型下可以并行维护数量巨大的滑动实时多数据流统计信息,同时采用精确方法并行计算多数据流间任意两条的相关系数。通过对比在同样的实验环境下只使用CPU的计算处理方法,验证了新方法的实时计算性能具有显著的提高。

面向不同数据分布的多维直方图算法COCA-Hist

基于代价的RDBMS优化器需要对含有范围查询的合取谓词的结果集基数进行准确的估计,多维直方图对多维数据分布进行模拟,避免在估计结果集基数时采用数据独立性假设,造成估计误差过大,进而导致选择非优化的查询执行计划.在不同的数据分布情况下,传统的多维直方图(如MHist-2)效果有很大不同.数据相关系数和值域密度、值域参数是准确刻画多维数据分布的有效指标,文中提出了根据不同的指标采用不同的动态优化的多维直方图算法COCA-Hist,可以大大改善传统多维直方图在平均情况下的准确性.通过分析传统的多维直方图的最坏情况,COCA-Hist的改进算法可以改进传统的多维直方图在最坏情况下的准确性.实验比较了COCA-Hist和传统的多维直方图MHist-2以及GENHist和STHoles的准确性和时间效率.实验显示无论在平均情况下还是在最坏情况下COCA-Hist的改进算法均优于传统的MHist-2直方图,并且COCA-Hist的准确性和创建时间均比GENHist有极大的改善,在准确性方面COCA-Hist较优于STHoles,而在空间预算有限时STHoles的创建时间比COCA-Hist高两个数量级.

Traffic condition estimation with pre-selection space time model

A pre-selection space time model was proposed to estimate the traffic condition at poor-data-detector,especially non-detector locations.The space time model is better to integrate the spatial and temporal information comprehensibly.Firstly,the influencing factors of the "cause nodes" were studied,and then the pre-selection "cause nodes" procedure which utilizes the Pearson correlation coefficient to evaluate the relevancy of the traffic data was introduced.Finally,only the most relevant data were collected to compose the space time model.The experimental results with the actual data demonstrate that the model performs better than other three models.

Is This Development and Freedom？ Managing Sex Risk Behaviour in the Era of HIV and AIDS at a University in South Africa

This study explored HIV risk perception of university students. A descriptive study design amongst a convenience sample of registered students at the Nelson Mandela Metropolitan University （NMMU）, who gave voluntary informed consent to participate, and with access to the student portal was employed. Frequencies and percentages were used to describe categorical data. The Pearson correlation co-efficient （r） and Spearman＇s rank co-efficient were used to measure the strength or degree of the relationship between variables and identify the significance of the correlation between two variables respectively. Results indicate that males in the sample （n = 619） are more likely to acknowledge self-perceived risk than females. This paper concludes that management strategies should be put in place in all universities in order to help the students stay HIV negative. Unless HIV and AIDS are institutionalised, the management of risk behaviour will prove difficult.

Trade Development and Global Crisis Influence

The purpose of this paper is to create a comprehensive analysis of the development of foreign trade in the global world, the process of divergence of exports and imports under the influence political and economic changes in Europe and the economic crisis in the world. Data from world public databases are summarized to a clear and understandable form. We analyzed the share of imports and exports （due to the global trade is presented） and its potential impact on the development of current account balance （CAB） for the United States of America （USA）, the European Union （EU）, and China. Correlation coefficient timeline for the last decade of CAB and world trade is also presented to show the influence of trade flow in the world and the EU with respect to current account. The work emphasizes the clear and understandable processing of the required data, which are then formulated to make arguments and then used to make predictions of further development of world trade. From summarized data, future crisis can be predicted and impacts can be evaluated.

Customer Loyalty： An Empirical Study on Italian E-commerce Websites

In contemporary entrepreneurial environment based on customer retention, the growth of the internet has pushed the most dynamic businesses to compete in the electronic market. In literature, there are many studies showing how a company＇s brand affects customer loyalty. Less attention has been paid to the relationship between e-commerce website loyalty and e-customer loyalty. This paper can be included in a stream of research that applies the construct of loyalty to e-commerce. Thus, this study aims to analyze, within the electronic market, the elements influencing and determining the trust and loyalty of the e-customers. This research consists of a descriptive study carried out utilizing the survey method. The population in this study includes 421 e-customers of Italian online multi-brand portals. The instrument of data gathering was a web-questionnaire and the data analysis was based on regression. Findings show a positive relationship between three factors （website technological level, e-service quality, and brand） and e-customer trust. Furthermore, the correlation, among the three variables, trust as mediating variable, and e-customer＇s loyalty, has been revealed. Therefore, it can be basically assumed that trust builds e-customer loyalty.

Regularized canonical correlation analysis with unlabeled data

In standard canonical correlation analysis (CCA), the data from definite datasets are used to estimate their canonical correlation. In real applications, for example in bilingual text retrieval, it may have a great portion of data that we do not know which set it belongs to. This part of data is called unlabeled data, while the rest from definite datasets is called labeled data. We propose a novel method called regularized canonical correlation analysis (RCCA), which makes use of both labeled and unlabeled samples. Specifically, we learn to approximate canonical correlation as if all data were labeled. Then, we describe a generalization of RCCA for the multi-set situation. Experiments on four real world datasets, Yeast, Cloud, Iris, and Haberman, demonstrate that, by incorporating the unlabeled data points, the accuracy of correlation coefficients can be improved by over 30%.

Quantitative description of steam channels after steam flooding

Steam channeling is one of the main barriers for EOR after steam flooding.In order to enhance the oil recovery in steam flooded reservoirs,steam channel volumes should be precisely known.In this paper,a set of methods has been established in order to study steam channeling quantitatively by using dynamic data.Firstly,steam channeling wells are identified through curves of watercut and temperature.Then,considering the hysteresis phenomenon,channeling relations are identified with the correlation coefficients between injection wells and production wells under different conditions.Lastly,an analytic model,in which steam condensation,pressure and temperature are considered,is established to calculate the steam channel volumes.A production well named L31615 in some block in Henan Oilfield in China is systematically analyzed by using the method established.The whole block is further analyzed,and the distribution map of steam channels and the steam channel volumes are obtained.The results show that steam channeling does not only occur inside a well group,but also sometimes occurs between wells in different well groups.The calculation of the steam channel volumes provides a theoretical basis for bringing a remedial action like plugging into operation.