Ming sequential associations is becoming increasing essential in many scientific and commercial domains.Developing parallel algorithm becomes quite challenging depending on enormous size of available dataset and possi...Ming sequential associations is becoming increasing essential in many scientific and commercial domains.Developing parallel algorithm becomes quite challenging depending on enormous size of available dataset and possiblylarge number of mined associations ,the nature of input data and the timing constraints imposed on the desired associa-tions. In this paper ,we discuss several different parallel algorithms that cater to various situations to speed up thecurrent mining process.展开更多
Rough sets is one important method of data mining. Data mining processes such a great quantity of data inlarge database that the speed of Rough Sets Data Mining Algorithm is critical to Data Mining System. Utilizing n...Rough sets is one important method of data mining. Data mining processes such a great quantity of data inlarge database that the speed of Rough Sets Data Mining Algorithm is critical to Data Mining System. Utilizing net-work computing resources is an effective approach to improve the performance of Data Mining System. This paperproposes the concept of meta-information,which is used to describes the result of Rough Sets Data Mining in informa-tion system,and a meta-information-based method for rule parallel mining. This method decomposes the information-system into a lot of sub-information-system,dispatchs the task of generating meta-information of sub-information-sys-tem to some task performer in the network,and lets them parallel compute meta-information,then synthesizes themeta-information of sub-information-system to the meta-information of information system in the task synthesizer,and finally produces the rule according to the meta-information.展开更多
为满足大数据实时处理的需求,提出了一种基于划分的关联规则并行分层挖掘算法(Parallel Hierarchical Association Rule Mining,PHARM)。首先,将整个数据库D随机分割成若干个非重叠区域,并行挖掘出局部频繁项集;然后利用先验性质,连接...为满足大数据实时处理的需求,提出了一种基于划分的关联规则并行分层挖掘算法(Parallel Hierarchical Association Rule Mining,PHARM)。首先,将整个数据库D随机分割成若干个非重叠区域,并行挖掘出局部频繁项集;然后利用先验性质,连接局部频繁项集得全局候选项集;再次扫描D统计出每个候选项集的实际支持度,以确定全局频繁项集。最后,建模分析了该算法的高效性。展开更多
文摘Ming sequential associations is becoming increasing essential in many scientific and commercial domains.Developing parallel algorithm becomes quite challenging depending on enormous size of available dataset and possiblylarge number of mined associations ,the nature of input data and the timing constraints imposed on the desired associa-tions. In this paper ,we discuss several different parallel algorithms that cater to various situations to speed up thecurrent mining process.
文摘Rough sets is one important method of data mining. Data mining processes such a great quantity of data inlarge database that the speed of Rough Sets Data Mining Algorithm is critical to Data Mining System. Utilizing net-work computing resources is an effective approach to improve the performance of Data Mining System. This paperproposes the concept of meta-information,which is used to describes the result of Rough Sets Data Mining in informa-tion system,and a meta-information-based method for rule parallel mining. This method decomposes the information-system into a lot of sub-information-system,dispatchs the task of generating meta-information of sub-information-sys-tem to some task performer in the network,and lets them parallel compute meta-information,then synthesizes themeta-information of sub-information-system to the meta-information of information system in the task synthesizer,and finally produces the rule according to the meta-information.
文摘为满足大数据实时处理的需求,提出了一种基于划分的关联规则并行分层挖掘算法(Parallel Hierarchical Association Rule Mining,PHARM)。首先,将整个数据库D随机分割成若干个非重叠区域,并行挖掘出局部频繁项集;然后利用先验性质,连接局部频繁项集得全局候选项集;再次扫描D统计出每个候选项集的实际支持度,以确定全局频繁项集。最后,建模分析了该算法的高效性。
