Representing Incomplete and Uncertain Information in Graph Data

Abstract

A method for representing and querying incomplete and uncertain information in graph data receives a plurality of graphs containing subject nodes, object nodes and predicates extending between subject and object nodes. The subject nodes and predicates can be URIs or blank, and the object nodes can be URIs, literals or blank. Incomplete graph data sets are created by a variable into each blank subject node, each blank predicate and each blank object node, and uncertain graph data sets are created by substituting alternative values for all variables in the incomplete data graph. A query is received from a user and a naïve search of the graph data is performed for certain data. The incomplete and uncertain graphs are then used to determine potential answers and certain potential answers based on user-specified requirements. The certain answers and potential certain answers are returned to the user.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS[0001]The present application is a continuation of co-pending U.S. patent application Ser. No. 13 / 476,316 filed May 21, 2012. The entire disclosure of that application is incorporated herein by reference.STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH[0002]The invention disclosed herein was made with U.S. Government support under Contract No. W911NF-09-2-0053 awarded by the U.S. Department of Defense. The Government has certain rights in this invention.FIELD OF THE INVENTION[0003]The present invention relates to data management in graph databases.BACKGROUND OF THE INVENTION[0004]A goal of any dataset covering any domain or type of data is to have as complete, certain and accurate a set of data as possible given limitations on the collection and storage of the data. Scientific data, for example, are often incomplete, e.g., only certain portions of the sky have been studied by astronomers, or inaccurate, e.g., instrument readings vary by the sen...

AI Technical Summary

Benefits of technology

The present invention provides a system and method for representing and querying incomplete and uncertain information in graph data. The system uses models that clearly define the representation of incompleteness and the way in which data is probed. The system includes a graph data input module for receiving graph data, an uncertainty and incompleteness specification module for creating incomplete and uncertain graph data, and a query processing module for processing user-defined queries. The system can create incomplete and uncertain graph data sets by inserting variables into blank subject nodes, predicates, and object nodes, and can also create partial uncertain graph data sets by substituting alternative values for variables in different parts of the graph. The system can also use a set of alternative values to represent each variable in the graph data. Overall, the system provides a more efficient and effective way to represent and query incomplete and uncertain information in graph data.

Problems solved by technology

Scientific data, for example, are often incomplete, e.g., only certain portions of the sky have been studied by astronomers, or inaccurate, e.g., instrument readings vary by the sensitivity of each instrument.

Similarly, business decisions are made using uncertain business data, e.g., when not all sales data have been reported in advance of a decision being made.

The advent of the Semantic Web has rendered these previous approaches useless.

Incomplete and inaccurate data are available everywhere, in massive amounts, and there is no centralized control.

Incompleteness in the data cannot be completely eliminated.

Due to the lack of clear semantics, this support is inadequate.

The use of probabilities is impractical as they need to be determined and associated with the data within RDF.

Images