Showing papers by "Jiawei Han published in 1991"

Learning in relational databases: an attribute-oriented approach

Abstract: The development of efficient algorithms for learning from large relational databases is an important task in applicative machine learning. In this paper, we study knowledge discovery in relational databases and develop an attribute-oriented learning method which extracts generalization rules from relational databases. The method adopts the artificial intelligence “learning-from-examples” paradigm and applies in the learning process an attribute-oriented concept tree ascending technique which integrates database operations with the learning process and provides a simple and efficient way of learning from databases. The method learns both characteristic rules and classification rules of a learning concept, where a characteristic rule characterizes the properties shared by all the facts of the class being learned; while a classification rule characterizes the properties that distinguish the class being learned from other classes. The learning result could be a conjunctive rule or a rule with a small number of disjuncts. Moreover, learning can be performed with databases containing noisy data and exceptional cases using database statistics. Our analysis of the algorithms shows that attribute-oriented induction substantially reduces the computational complexity of the database learning process. Le developpement d'algorithmes efficaces permettant l'apprentissage a partir de bases de donnees relationnelles est une fonction importante de l'apprentissage automatique applicatif. Dans cet article, les auteurs examinent la decouverte des connaissances dans les bases de donnees relationnelles et elaborent une methode d'apprentissage orientee sur l'attribut qui extrait des bases de donnees relationnelles les regies de generalisation. La methode adopte le paradigme d'apprentissage a partir d'exemples et applique au processus d'apprentissage la technique de l'arbre des concepts orientes sur l'attribut qui incorpore les operations de base de donnees au processus d'apprentissage, ce qui permet d'obtenir une methode simple et efficace d'apprentissage a partir des bases de donnees. La methode fait l'apprentissage des regies caracteristiques et des regies de classification d'un concept d'apprentissage; la regie caracteristique qualifie les pro-prietes communes a tous les faits d'une categorie faisant l'objet d'un apprentissage alors que la regie de classification caracterise les proprietes qui distinguent la categorie faisant l'objet d'un apprentissage des autres categories. Le resultat peut ětre une regie conjonctive ou une regie ayant un petit nombre de disjonctifs. Qui plus est, 1′apprentissage peut se faire avec des bases de donnees contenant des donnees bruitees et des cas exceptionnels utilisant des statistiques de bases de donnees. L'analyse des algorithmes demontre que l'induction orientee sur l'attribut reduit considerablement la complexite informatique du processus d'apprentissage des bases de donnees.

Constraint-based reasoning in deductive databases

Abstract: Constraint-based reasoning in deductive databases is studied, with the focus on set-oriented, constraint-based processing of functional linear recursions. A technique is developed which compiles a functional linear recursion into chain or bounded forms and analyzes efficient processing of the compiled chains based on different kinds of constraints. It is shown that rule constraints should be compiled together with the rectified recursions; finiteness constraints and monotonicity constraints should be used in the analysis of finite evaluability and termination; and query constraints should be pushed into the compiled chains, when possible, for efficient set-oriented evaluation. Constraint-based processing can be enhanced by dynamic constraint enforcement in query evaluation. The method is illustrated using a typical traversal recursion problem. It is concluded that the principles developed are useful for a large set of deductive database application problems. >

Efficient access methods in deductive and object-oriented databases

Abstract: Recently, a new access method, the Hierarchical-Tree (H-tree) was proposed as an efficient access method for object-oriented database that supports superclass-subclass relationship. We show that, the H-tree indexing scheme is indeed a general access method for the new generation DBMS that most probably will support the object-oriented concept and have the deductive capability. The simplest form of the H-tree, like the widely used B+-tree, can be used to index simple objects for efficient associative search. Its nesting capability provides efficient support to different types of queries that reference objects in superclass-subclass hierarchy. Furthermore, by dynamically nesting the indexes to the objects generated iteratively during recursive query processing, it also supports the least fixed point computation in object-oriented database with deductive capability. This paper examines the use of the H-tree as uniform indexing structure to index objects for efficient query retrieval and, as an efficient implementation of the semi-naive evaluation of least fixed point computation.

Efficient evaluation of multiple linear recursions

Abstract: The authors study the efficient evaluation of side-coherent multiple linear recursions, which can be further classified into three types: multiple one-sided, multiple balanced k-sided, and multiple mixed k-sided. New techniques are developed by integrating the existing single-linear recursive query evaluation methods with the idea of side-relation unioned processing, which leads to a set of efficient query evaluation algorithms such as a side-relation unioned transitive closure algorithm for the processing of Type I ML recursions and a generalized side-relation unioned magic sets method for the processing of Types II and III ML recursions. The authors describe the processing of single-probe queries on side-coherent ML recursions. They outline the processing of complex queries on ML recursions. >

On the Power of Query-Independent Compilation

Abstract: Recursive query processing techniques can be classified into three categories: interpretation, query-dependent compilation and query-independent compilation Query-dependent compilation compiles IDB programs based on possible query instantiations into query-specific EDB programs, while query-independent compilation compiles IDB programs into query-independent and easily analyzable EDB expressions Previous studies show that linear recursions can be query-independently compiled into highly regular forms This study analyzes the power of query-independent compilation and shows that (i) query-independent compilation captures more binding information than other methods for irregular linear recursions; (ii) the compilation provides succinct information for the selection of efficient query processing methods; and (iii) it facilitates the constraint-based processing of complex queries Finally, query-independent compilation can be applied to more complex recursions as well

Efficient Evaluation of Linear Recursions with Function Sysmbols in Deductive Databases

Abstract: We study the compilation and efficient evaluation of functional linear recursions in deductive databases and develop a compiled chain-based query evaluation technique. A functional linear recursion is transformed to a function-free one by a function-predicate transformation. It is then compiled to a highly regular chain/bounded form which can be evaluated efficiently by the incorporation of finiteness, monotonicity and query constraints. Compilation greatly facilitates the analysis of functional recursions.