Investment analysis tool and service for making investment decisions

• US 7,225,174 B2
• Filed: 02/19/2002
• Issued: 05/29/2007
• Est. Priority Date: 07/14/1999
• Status: Expired due to Term

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Schreckengast, James O., Skaanning, Claus, Gilbert, Douglas M.
• Primary Examiner(s): Knight, Anthony
• Assistant Examiner(s): Holmes, Michael B.
• Application Number:
  US10/078,971
• Publication Number:
  US 20020128943A1
• Time in Patent Office: 1,925 Days
• Field of Search: 706/45, 705/36
• US Class Current: 706/45
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis

Abstract

A decision support system supports investment decisions. The decision support system includes a decision support engine and a Bayesian network. The Bayesian network is traversed by the decision support engine. The Bayesian network includes an investment decision node, potential investment nodes, and information nodes. The investment decision node identifies an investment decision. The potential investment nodes identify potential investments. The information nodes identify information to be obtained. The information to be obtained is matched to potential investments. Reliability of the information is estimated, for example, with the help of domain experts.

Claims

1. A method for using a Bayesian network to diagnose an issue in a stock brokering domain, the method comprising:

displaying via a user interface one or more questions for guiding a user to identify at least one issue in the stock brokering domain to be solved by traversing questions connected within the Bayesian network responsive to answers to the questions;

responsive to identifying the at least one issue, identifying a cause connected via at least one directed edge in the Bayesian network with the identified issue;

traversing at least one directed edge from the identified cause to an action in the Bayesian network;

displaying via the user interface the action for the user to perform; and

responsive to the user having performed the action, receiving user input indicating whether or not the action solved the issue.

2. The method of claim 1 further comprising:

responsive to the action not solving the identified issue, determining an optimal sequence of one or more actions from a set of actions connected in the Bayesian network to the identified cause which have not already been performed; and

computing an expected cost of executing the optimal sequence.

3. The method of claim 2 further comprising:

computing an expected cost of first asking a question from a set of questions which have not already been answered and the cost of performing the optimal sequence of actions given an answer to the question.

4. A system for using a Bayesian network to diagnose an issue in a stock brokering domain, the system comprising:

means for displaying via a user interface one or more questions for guiding a user to identify at least one issue in the stock brokering domain to be solved by traversing questions connected within the Bayesian network responsive to answers to the questions;

responsive to identifying the at least one issue, means for identifying a cause connected via at least directed edge in the Bayesian network with the identified issue;

means for traversing at least one directed edge from the identified cause to an action in the Bayesian network;

means for displaying via the user interface the action for the user to perform; and

responsive to the user having performed the action, means for receiving user input indicating whether or not the action solved the issue.