基于数据关联度的迁移策略研究

以电子政务云存储环境为背景,研究了已有迁移算法中用到的基本迁移因子,包括静态因素和动态因素.在计算迁移函数的M(d)值时引入数据关联度这一新的因子.得出新的数据迁移函数后,根据混合云的迁移算法设计出新的适合于电子政务云存储的迁移策略.新策略提高了数据的预计迁移值,降低大数据管理成本,对于准备采用云存储管理大数据的企业或相关部门有一定的参考价值.

基于数据挖掘地域性强关联规则数据提取

针对数据提取方法无法进行海量挖掘,且挖掘结果不准确,挖掘时间较长的问题,提出一种基于数据挖掘算法的地域性强关联规则数据提取方法。结合地域性强关联规则数据管理系统,采集用户需求信息,检索特征关联性,收集地域特征。利用数据关联度,分析地域检索中地域特征间的关联,计算相似标签信息参数,并对支持度和置信度实施计算,从地域性强关联规则数据库中挖掘关联规则。利用Kulczynski量度和不平衡率实施相关性分析和过滤,最终获取到具有实际意义的强关联规则。实验结果表明,该方法挖掘效率较高,且具有广泛的应用价值。

基于大数据分析的气象观测数据质量控制算法研究

目前,常用的气象数据质量控制算法均是依据统计学原理对数据的变化趋势进行分析,针对其算力无法对大数据进行处理以及实现质量控制的问题。文中将BP神经网络模型与粒子滤波算法结合,有效增强了神经网络的综合计算性能,将该算法应用于海量气象数据处理,可以显著减少外在因素对数据精确度的影响。实验结果表明,所提模型的预测数据误差平均值和异常数据监测误差平均值在对比算法中最小,证明该模型可以准确地监测气象数据,同时能够对异常气象数据进行修补拟合,进而实现对气象数据的质量控制。

基于日志数据的局域底层软件运行故障定位

传统的软件故障定位方法定位耗时过长,且定位准确率较低。为此,研究基于日志数据设计了一种新的局域底层软件运行故障定位方法。首先基于日志数据采集局域底层软件运行信息,通过设计线程函数运算程序采集系统日志、防火墙日志以及网络访问日志数据。在明确日志数据之间的关联度后得到映射矩阵,然后根据检测参数确定运行周期,在相关性处理后构建故障检测模型。最后根据用户系统的日志类型,对日志数据采集程序的参数进行设置与调整,从而实现故障精准定位。实验结果表明:上述故障定位方法能够有效缩短定位时间,且其定位准确率较高,证明上述方法具有较强的实用性能。

Study on multiple maneuvering targets tracking based on JPDA algorithm

A tracking algorithm for multiple-maneuvering targets based on joint probabilistic data association（JPDA）is proposed to improve the accuracy for tracking algorithm of traditional multiple maneuvering targets.The interconnection probability of the two targets is calculated,the weighted value is processed and the target tracks are obtained.The simulation results show that JPDA algorithm achieves higher tracking accuracy and provides a basis for more targets tracking.

Measurement and Correlation for Solubility of 2-Chloro-5-chloromethylpyridine in Different Solvents

2-Chloro-5-chloromethylpyridine is a crucial intermediate of pesticides.Its solid-liquid equilibrium data will provide essential support for industrial design and further theoretical studies.The solubilities of 2-chloro-5-chloromethylpyridine in water,methanol,ethanol,ethyl acetate,acetone,trichloromethane and toluene at different temperatures were measured using the synthetic method by a laser monitoring observation technique. The solubility data were correlated with the modified Apelblat equation and Wilson equation.The calculated values were in good agreement with the experimental values.

A New Thinking of the Objects Served Relationship Management in Complex System

Take a digital libraries＇ service system for example, Objects Served Relationship Management （OSRM） in complex systems is proposed firstly as a new concept, and its connotation is explained. The significances and constructions of OSRM are analyzed. Both the fundamental facts and the important natures that the things which are interested by Objects Served （OS） （e. g. publishers and readers） and the server （e. g. digital libraries are the servers of publishers and readers） will not be the same completely although there are a lot of common benefits between OS and servers, are indeed clarified. The valuable information,which should be used by OS and their server, is often hidden behind them. Thus, how to find, manage and control the relationship among OS and their servers is very necessary and important for the common benefits among all of them.（e. g. the three dimensions of OSRM in digital library system and its overall framwork are proposed. The different strategies to different cases in the digital library＇s multidimensional framework are analyzed.）