Currently, there is tremendous interest in providing ad-hoc mining capabilities in database management systems. As a first step towards this goal, in [15] we proposed an architecture for supporting constraint-based, human-centered, exploratory mining of various kinds of rules including associations, introduced the notion of constrained frequent set queries (CFQs), and developed effective pruning optimizations for CFQs with 1-variable (1-var) constraints.
While 1-var constraints are useful for constraining the antecedent and consequent separately, many natural examples of CFQs illustrate the need for constraining the antecedent and consequent jointly, for which 2-variable (2-var) constraints are indispensable. Developing pruning optimizations for CFQs with 2-var constraints is the subject of this paper. But this is a difficult problem because: (i) in 2-var constraints, both variables keep changing and, unlike 1-var constraints, there is no fixed target for pruning; (ii) as we show, "conventional" monotonicity-based optimization techniques do not apply effectively to 2-var constraints.
The contributions are as follows. (1) We introduce a notion of
当前,在数据库管理系统中提供即席挖掘功能引起了极大的兴趣。作为朝着这个目标迈出的第一步,我们在[15]中提出了一种用于支持基于约束的,以人为中心的,探索性挖掘各种规则(包括关联)的体系结构,引入了约束频繁集查询(CFQ)的概念,并开发了具有1变量(1-var)约束的CFQ的有效修剪优化。 P>
尽管1-var约束对于分别约束先决条件和结果很有用,但许多自然CFQ实例说明了共同约束前因和结果的必要性,对于这些变量,2-变量(2-var)约束是必不可少的。本文针对具有2-var约束的CFQ开发修剪优化是本文的主题。但这是一个困难的问题,因为:(i)在2-var约束中,两个变量都不断变化,并且与1-var约束不同,没有固定的修剪目标; (ii)正如我们所展示的那样,基于“传统”单调性的优化技术不能有效地应用于2-var约束。 P>
贡献如下。 (1)我们引入了
机译:使用时变终端约束集的最小-最大反馈MPC,并评论“具有时变终端约束集的有效鲁棒约束模型预测控制”
机译:约束多目标风电场布局优化:基于约束规划的新型约束处理方法
机译:约束频繁集的有效动态挖掘
机译:用2变量约束优化约束频繁设置查询
机译:高效有效地探索受限频繁集。
机译:具有熵多样性约束的基数约束均值方差投资组合优化问题的萤火虫算法
机译:具有2变量约束的约束频繁集查询的优化
机译:航天器约束机动规划使用正不变约束允许集(后印刷)。