次世代パーソナル情報管理システムにおけるデータモデル・検索に関する研究

近年来，人们对传统关系数据库的关键字搜索越来越感兴趣。通常，关系数据库表示为图结构;计算适当的子图作为关键字查询的答案。在这些研究中，查询结果的单位是基于“最小”的概念，即每个答案子图是包含每个查询关键字的最小子图。然而，在个人信息管理系统中，“最小性”的概念不能应用于在表示异构信息的图上的关键字搜索。这是因为适当的答案不一定是包括每个查询至少出现一次的节点的子图。因此，需要考虑进一步扩大“最低限度”的概念。在这项研究中，我们提出了“超越最小”的概念，计算答案单位的关键字搜索在一个图表示异构信息。我们还探索了代数， 关于我们 在传统的信息检索系统中，通常按照相关性的降序对搜索结果进行排序。换句话说，搜索结果，而不仅仅是一个元素的集合，是一个排名列表类型。因此，由于传统数据库系统中的查询结果是元组元素的简单集合，因此查询优化是通过各种策略来实现的--其中之一是代数运算的操作。但是，在个人信息管理系统中，在关键字检索中，显然无法保证同样的优化。电子教科书和引用这些教科书的PowerPoint幻灯片在个人信息管理系统中具有特定的特征。这种数据可以被归类为一种特殊类型的同步媒体流，可以包含额外的时间信息。为了高效和有效地找到信息在一个大的收集同步的媒体流，我们提出了一个数据模型，表示在一个二维空间中的信息。这个空间可以由我们的查询处理器操作。少

In recent years, there has been growing interest in keyword search over traditional relational databases. Usually a relational database is represented as a graph structure; an appropriate subgraph is computed as an answer to a keyword query. In these studies, the units of query results is based on the concept of 'minimality', that is, each answer subgraph is the smallest subgraph containing each query keyword. However, the concept ofminimality' cannot be applied to keyword search over a graph representing heterogeneous information in a personal information management system. It is because an appropriate answer is not necessarily a subgraph comprising the nodes having at least one occurrence of each query. As a result, further expansion of the concept of 'minimality' needs to be need to be considered. In this research, we proposed the concept of "beyond minimality" for computing answer units in keyword search over a graph representing heterogeneous information. We also explored algebrai … More c operations to generate such answer units.In a traditional information retrieval system, it is common to rank search result in a descending order of relevance. In other words, search result, rather than just a collection of elements, is a ranked list type. Thus, as query result in a traditional database system is a simple collection of tuple elements, query optimization is achieved through various strategies---one of them being the manipulation of algebraic operations. However, it is obvious that the similar kind of optimization cannot be guaranteed in keyword search over a personal information management system.Electronic textbooks and PowerPoint slides which refer to those textbooks have specific characteristics in a personal information management system. Such data can be categorized as a special kind of synchronized media streams can contain additional temporal information. In order to efficiently and effectively find information in a large collection of synchronized media streams, we have proposed a data model that represents information in a two-dimensional space. This space can be manipulated by our query processor. Less