METHOD AND APPARATUS FOR ORGANIZING DATA SOURCES

• US 20080259084A1
• Filed: 06/27/2008
• Published: 10/23/2008
• Est. Priority Date: 08/14/2006
• Status: Abandoned Application

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Zhang, Zhen, Wang, Min, Lim, Lipyeow, Chang, Yuan-chi
• Application Number:
  US12/163,485
• Publication Number:
  US 20080259084A1
• Time in Patent Office: Days
• US Class Current: 345/440
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99953 Recoverability

Abstract

A method and apparatus for organizing deep Web services are provided. In one aspect, the method and apparatus obtains a collection of sources and their associated attributes and/or input modes, for instance, using a crawling algorithm. The method and apparatus uses this information to organize the sources into communities. A mining algorithm such as the hyperclique mining algorithm is used to obtain cliques of highly correlated attributes. A clustering algorithm such as the hierarchical agglomerative clustering algorithm is used to further cluster the cliques of attributes into larger cliques, which in the present disclosure is referred to as signatures. The sources that are associated with each signature form a community and a graph representation of the communities is constructed, where the vertices are communities and the edges are the shared attributes.

Claims

1. A computer-implemented method of organizing data sources, comprising:

grouping a plurality of items including input attributes, output attributes, or keywords or combination thereof from one or more sources into a plurality of cliques of highly correlated items;

clustering the plurality of cliques into one or more signatures; and

for each of the one or more signatures,

selecting one or more sources that are associated with a signature; and

forming the selected sources into a community.

2. The method of claim 1, further including:

constructing a graph representation of a plurality of communities, the graph representation including at least a plurality of vertices representing the plurality of communities respectively and one or more edges connecting the plurality of vertices, the one or more edges representing one or more input attributes, output attributes, or keywords or combination thereof that are shared between the communities represented in the connecting vertices.

3. The method of claim 2, further including:

navigating the graph representation including at least one of;

starting at a vertex representing a community, following one or more edges to one or more second vertices representing one or more related communities;

starting at a source, traversing to one or more associated vertices; and

starting from an attribute or a keyword or combination thereof, traversing one or more associated edges and connected vertices.

4. The method of claim 1, wherein the step of grouping is performed using a hyperclique mining algorithm.

5. The method of claim 1, wherein the step of clustering is performed using a hierarchical agglomerative clustering algorithm.

6. The method of claim 1, further including:

obtaining the plurality of items including input attributes, output attributes, or keywords or combination thereof from one or more sources using a crawling algorithm.

7. An apparatus for organizing data sources, comprising:

a means for grouping a plurality of items including input attributes, output attributes, or keywords or combination thereof from one or more sources into a plurality of cliques of highly correlated items; and

a means for clustering the plurality of cliques into one or more signatures,

a means for selecting one or more sources that are associated with a signature and forming the selected sources into a community for each of the one or more signatures; and

a means for constructing a graph representation of a plurality of communities, the graph representation including at least a plurality of vertices representing the plurality of communities respectively and one or more edges connecting the plurality of vertices, the one or more edges representing one or more input attributes, output attributes, or keywords or combination thereof that are shared between the communities represented in the connecting vertices.