A Decision Tree Interface Based on Predicate Calculus

Abstract

The paper presents automated relational technology ART Studio that generates end-user decision trees from relational data automatically. End-users navigate down these tree structures to pinpoint information managed by the RDBMS. This decision tree reflects and selfreferences data in the database; it turns the data inside out. To generate this visualization, the Studio models a uniform pattern of data in the RDBMS called the Aleph data relation. Earlier research reports that the Aleph is a mechanically derived parent-child data relationship that is ubiquitous throughout the database. The paper will now argue that the Aleph is a complex mathematical object that is dense and compact. Subsequently, its identity is multifaceted. On the computer, its physical symbols represent both an IF-THEN construct, as well as a rich and expressive predicate that replace the need for creating SQL SELECT commands. To support this claim the author draws on two sources: 1) Codd's view of the relational database as a form of predicate calculus, and 2) the author's nascent theory of meta-symbols. The bridging metaphor between these two ideas is the fact that data symbols can be viewed in linguistic terms as tokens and types. The paper will show how the Alephbased decision tree models database tables and their data. It has a logical correspondence with the tables attributes and tuples. The paper shows how predicates in its pathways map to predicates in Codd's tuple components. This point is made to highlight how Alephbased decision trees project Codd's view of the RDBMS as a form of predicate calculus.