Technique for data mining using a web service

• US 7,657,535 B2
• Filed: 10/01/2004
• Issued: 02/02/2010
• Est. Priority Date: 10/01/2004
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: International Business Machines Corporation
• Inventors: Selby, David A., Schott, Charles J., Thomas, Vince P., Moyaux, Yan
• Primary Examiner(s): Wong; Don
• Assistant Examiner(s): XUE, BIN
• Application Number:
  US10/956,579
• Publication Number:
  US 20060085422A1
• Time in Patent Office: 1,950 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 16/972 Access to data in other repository systems, e.g. legacy data or dynamic Web page generation

Abstract

A technique for the deployment of data mining algorithms on a web service, such as IBM'"'"'s WebSphere Application Server, is disclosed. Rather than having to deploy the data mining models with the data, the data can be transported to the web server as part of a message. Models can be cached on the web server and easily changed by operations executed by the client. This allows for efficient administration of the operational environment. Because a web services environment is inherently scalable, servers can be transparently enabled based on demand. Further, with web services communication is via data objects in memory which allows for ease of implementation and operational efficiency.

Claims

1. A system for performing an enhanced data mining operation over a network, said network including a web server and at least one client device, comprising:

means for receiving, by said web server, a first message from one of said client devices, said message containing data to be subjected to said data mining operation,

wherein said first message identifies a plurality of workspaces cached on said web server to be used in connection with said data mining operation;

wherein said plurality of workspaces cached on said web server identify one or more data mining algorithms stored on said web server to be used to perform said data mining operation and identify model parameters pertinent to said one or more identified algorithms;

and wherein said first message comprises a self-describing data envelope, said self-describing data envelope containing the data to be subjected to said data mining operation and a particular workspace to be used in conducting the data mining operation;

means for performing, by said server, said data mining operation using said workspace and the data to be subjected to said data mining operation, to produce a data mining result; and

means for preparing, and packaging in a results data envelope, a second message, to said one of said client devices, said second message containing said data mining result.