Abstract
This paper presents STARBASE, a new algorithm for recursive query processing on deductive databases base on the Chart Parsing algorithm. It is distinct from the other applications of parsing to deduction, namely Earley Deduction and Rosenblueth's method, because it removes variables from literals and extends the Chart Parsing engine to handle all possible variations in the pattern of arguments in the literals of deduction rules. Like other tabling methods, STARBASE avoids redundant computation by storing and reusing partial results but, in contrast with them, it does not depend on subsumption and uses syntactic equality checking, instead. Because STARBASE takes a strongly graph-oriented view of both the database and the deduction rules, the evaluation of a query on a database can be viewed as a process of traversing paths in the graph representing the database. A prototype of the STARBASE system has been implemented in the C language. Performance results show that STARBASE, even in prototype form, lies within the performance range of the most advanced existing systems.
Cite
CITATION STYLE
Pulido, E. (1996). Recursive query processing using graph traversal techniques. In International Conference on Information and Knowledge Management, Proceedings (pp. 37–44). https://doi.org/10.1145/238355.238371
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.