Unnormalized Relations Research Articles

Bridging Weighted Rules and Graph Random Walks for Statistical Relational Models.

The aim of statistical relational learning is to learn statistical models from relational or graph-structured data. Three main statistical relational learning paradigms include weighted rule learning, random walks on graphs, and tensor factorization. These paradigms have been mostly developed and studied in isolation for many years, with few works attempting at understanding the relationship among them or combining them. In this article, we study the relationship between the path ranking algorithm (PRA), one of the most well-known relational learning methods in the graph random walk paradigm, and relational logistic regression (RLR), one of the recent developments in weighted rule learning. We provide a simple way to normalize relations and prove that relational logistic regression using normalized relations generalizes the path ranking algorithm. This result provides a better understanding of relational learning, especially for the weighted rule learning and graph random walk paradigms. It opens up the possibility of using the more flexible RLR rules within PRA models and even generalizing both by including normalized and unnormalized relations in the same model.

Form document management system SPECDOQ its architecture and implementation

Architecture and implementation of a form document management system SPECDOQ are presented. The system automates or semi-automates manipulation of conventional form documents combining alphanumeric and Kanji/Kana character strings, texts, and graphics. SPECDOQ features uniform management of the document database in terms of unnormalized relations, named nested tables. Complicated manipulation of form documents can be accomplished by combinations of abstract, algebraic operations on the nested tables. The system also features graphics-based user friendly manipulation of external form documents on the screen. Architectural key concepts in SPECDOQ and their implementation are explained. Some system-user interaction examples and early experiences on SPECDOQ field use are also given in this paper.

Synthesis of unnormalized relations incorporating more meaning

There are two methods to handle redundancies caused by dependency constraints in relational databases. The first method, decomposition into relations in high-level normal form, has been studied by various authors, since it is very important for logical database design. The second method, representation by unnormalized form, is discussed in this paper. Although the latter method has not been studied except for simple cases, it is considered to be a very important tool for a user interface. We use four basic operations, GROUP-BY, ROW-NEST, COLUMN-NEST and RELATION-NEST to obtain unnormalized relations from a relation in a normal form. It is shown that 1. (1) functional dependencies can be handled by GROUP-BY operations, 2. (2) a join dependency is represented by a combination of GROUP-BY, ROW-NEST and RELATION-NEST operations and 3. (3) COLUMN-NEST operations can be used for a hierarchy of related attribute sets. Interactions among these operations are also discussed. We give procedures to represent an important class of constraints, consisting of one join dependency, functional dependencies satisfied by each component of the join dependency and a hierarchy of related attribute sets.