近年来,隐私保护事务数据发布得到了研究者的广泛关注.事务数据的稀疏性导致个体隐私保护与数据效用性之间很难达到平衡.目前已有的方法大多是基于分组的匿名模型,但该类模型依赖于攻击者背景知识,且发布的数据无法满足事务数据分析任...近年来,隐私保护事务数据发布得到了研究者的广泛关注.事务数据的稀疏性导致个体隐私保护与数据效用性之间很难达到平衡.目前已有的方法大多是基于分组的匿名模型,但该类模型依赖于攻击者背景知识,且发布的数据无法满足事务数据分析任务的需要.针对事务数据隐私保护发布的数据安全性与效用性不足,基于差分隐私与压缩感知理论,提出一种有效的面向应用的事务数据发布策略(transaction data publish strategy,TDPS).首先构建事务数据库的完整Trie项集树,然后基于压缩感知技术对项集树添加满足差分隐私约束的噪音得到含噪Trie项集树,最后在含噪树上进行频繁项集挖掘任务.实验结果表明,TDPS不仅能很好地保护隐私,而且能有效保持数据效用性,满足事务数据分析任务对数据质量的要求.展开更多
In this paper, we first briefly introduce the concepts of clickstream data and data warehouse, analyze twoexisting clickstream star schema click star schema and session star schema in webhouse, then induce a new mod-e...In this paper, we first briefly introduce the concepts of clickstream data and data warehouse, analyze twoexisting clickstream star schema click star schema and session star schema in webhouse, then induce a new mod-el transaction star model based on them, and expressed the method of bringing out the model. Comparing withthe two schemas mentioned above, its most apparent speciality is that it includes a series of meaningful page-view se-quence rather than a single click. Thus, on the one hand it improves the query performance of data, on the other handit is in favor of executing more deepen analysis data mining, and simplifies the process of data pretreatment. Atlast ,the paper verifies its' feasibility and validity using association rules based on the model.展开更多
The author combines statistical thoughts with coarse aggregate theory,statistically describes affair data bank system, forwards the method of compressing affair data bank,and further conducts data mining regarding con...The author combines statistical thoughts with coarse aggregate theory,statistically describes affair data bank system, forwards the method of compressing affair data bank,and further conducts data mining regarding consumption information of mobile telecommunication subscriber by using the method.展开更多
数据流是随着时间顺序快速变化的和连续的,对其进行频繁模式挖掘时会出现概念漂移现象.在一些数据流应用中,通常认为最新的数据具有最大的价值.数据流挖掘会产生大量无用的模式,为了减少无用模式且保证无损压缩,需要挖掘闭合模式.因此,...数据流是随着时间顺序快速变化的和连续的,对其进行频繁模式挖掘时会出现概念漂移现象.在一些数据流应用中,通常认为最新的数据具有最大的价值.数据流挖掘会产生大量无用的模式,为了减少无用模式且保证无损压缩,需要挖掘闭合模式.因此,提出了一种基于时间衰减模型和闭合算子的数据流闭合模式挖掘方式TDMCS(Time-Decay-Model-based Closed frequent pattern mining on data Stream).该算法采用时间衰减模型来区分滑动窗口内的历史和新近事务权重,使用闭合算子提高闭合模式挖掘的效率,设计使用最小支持度-最大误差率-衰减因子的三层架构避免概念漂移,设计一种均值衰减因子平衡算法的高查全率和高查准率.实验分析表明该算法适用于挖掘高密度、长模式的数据流;且具有较高的效率,在不同大小的滑动窗口条件下性能表现是稳态的,同时也优于其他同类算法.展开更多
文摘近年来,隐私保护事务数据发布得到了研究者的广泛关注.事务数据的稀疏性导致个体隐私保护与数据效用性之间很难达到平衡.目前已有的方法大多是基于分组的匿名模型,但该类模型依赖于攻击者背景知识,且发布的数据无法满足事务数据分析任务的需要.针对事务数据隐私保护发布的数据安全性与效用性不足,基于差分隐私与压缩感知理论,提出一种有效的面向应用的事务数据发布策略(transaction data publish strategy,TDPS).首先构建事务数据库的完整Trie项集树,然后基于压缩感知技术对项集树添加满足差分隐私约束的噪音得到含噪Trie项集树,最后在含噪树上进行频繁项集挖掘任务.实验结果表明,TDPS不仅能很好地保护隐私,而且能有效保持数据效用性,满足事务数据分析任务对数据质量的要求.
文摘In this paper, we first briefly introduce the concepts of clickstream data and data warehouse, analyze twoexisting clickstream star schema click star schema and session star schema in webhouse, then induce a new mod-el transaction star model based on them, and expressed the method of bringing out the model. Comparing withthe two schemas mentioned above, its most apparent speciality is that it includes a series of meaningful page-view se-quence rather than a single click. Thus, on the one hand it improves the query performance of data, on the other handit is in favor of executing more deepen analysis data mining, and simplifies the process of data pretreatment. Atlast ,the paper verifies its' feasibility and validity using association rules based on the model.
文摘The author combines statistical thoughts with coarse aggregate theory,statistically describes affair data bank system, forwards the method of compressing affair data bank,and further conducts data mining regarding consumption information of mobile telecommunication subscriber by using the method.
文摘数据流是随着时间顺序快速变化的和连续的,对其进行频繁模式挖掘时会出现概念漂移现象.在一些数据流应用中,通常认为最新的数据具有最大的价值.数据流挖掘会产生大量无用的模式,为了减少无用模式且保证无损压缩,需要挖掘闭合模式.因此,提出了一种基于时间衰减模型和闭合算子的数据流闭合模式挖掘方式TDMCS(Time-Decay-Model-based Closed frequent pattern mining on data Stream).该算法采用时间衰减模型来区分滑动窗口内的历史和新近事务权重,使用闭合算子提高闭合模式挖掘的效率,设计使用最小支持度-最大误差率-衰减因子的三层架构避免概念漂移,设计一种均值衰减因子平衡算法的高查全率和高查准率.实验分析表明该算法适用于挖掘高密度、长模式的数据流;且具有较高的效率,在不同大小的滑动窗口条件下性能表现是稳态的,同时也优于其他同类算法.
