Method and system for predicting transaction

• US 6,519,576 B1
• Filed: 02/17/2000
• Issued: 02/11/2003
• Est. Priority Date: 09/25/1999
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Freeman, Clive Alan
• Primary Examiner(s): Black, Thomas
• Assistant Examiner(s): Booker, Kelvin
• Application Number:
  US09/506,163
• Time in Patent Office: 1,090 Days
• Field of Search: 706/48, 706/21, 706/10, 706/20, 706/50, 705/43, 705/35, 705/42, 707/6
• US Class Current: 706/21
• CPC Class Codes: G06F 16/958 Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 40/02 Banking, e.g. interest calculation or account maintenance credit or loans G06Q40/03
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

The invention provides a system for predicting a transaction a customer may wish to make. In, for example, an Internet banking system, when the customer clicks to bring up an “inter-account transfer” panel, the system may recognise that usually, with a given balance in their savings and cheque accounts, and at this time of the month, at this point in the financial year, the customer will want to transfer a given amount to their cheque account. So, as well as providing a window enabling the customer to manually enter the details for an inter-account transfer, the system also provides a personalised window containing the system'"'"'s anticipated transaction so that with a single-click the customer can complete the transaction.

Claims

1. A personalisation sub-system for an Internet transaction processing system comprising a web server responsive to requests from a web client computing device to execute a transaction, said sub-system comprising:

means, responsive to said client request, for obtaining historical records corresponding to previous client transactions;

means for partially populating a target record from data available to said sub-system, said target record comprising a plurality of attributes;

means for measuring a similarity of each of said historical records to said partially populated target record, wherein said means for measuring comprises means for calculating a normalised Euclidian distance between each of said historical records and said partially populated target record;

means for selecting at least one similar historical record;

means for populating remaining target record attributes from at least one corresponding attribute of the at least one similar historical record; and

means for providing said populated target record to said system as a prediction of said transaction to be executed.

2. A personalisation sub-system as claimed in claim 1 wherein said sub-system is adapted to reside on said client computing device.

3. A personalisation sub-system as claimed in claim 1 wherein said selecting means is adapted to select a pre-determined number of similar historical records;

and said populating means is adapted to populate said remaining target record attributes as a function of said pre-determined number of similar historical records'"'"' corresponding attributes.

4. A personalisation sub-system as claimed in claim 3 wherein said function is a median of said pre-determined number of similar historical records'"'"' continuous attributes or a mode of said pre-determined number of similar historical records'"'"' discrete attributes.

5. A personalisation sub-system as claimed in claim 1 comprising:

means for calculating a weighting for each remaining target record attribute, said weighting being indicative of a remaining target record attribute'"'"'s dominance; and

wherein said selecting means is adapted to select a most similar historical record; and

said populating means is adapted to populate said remaining target record attributes in order of said remaining target record attributes'"'"' dominance.

6. A personalisation sub-system as claimed in claim 5 wherein said target record attributes comprise a combination of absolute value attributes and attributes whose value is derived from other attributes, and said populating means is responsive to a derived value attribute being the most dominant remaining attribute to populate said target record'"'"'s corresponding derived value attribute and any remaining target record attributes that can be calculated from said derived value attribute.

7. A personalisation sub-system as claimed in claim 1 comprising:

means, responsive to said client accepting said prediction, for increasing the weighting of the most dominant remaining target record attribute.

8. A personalisation sub-system as claimed in claim 1 wherein said transaction processing system is an Internet Bank System.

9. A personalisation sub-system as claimed in claim 8 wherein said available data comprises any combination of a current time, a current date or one or more current account balances.

10. A personalisation sub-system as claimed in claim 1 wherein said measuring means comprises means for calculating a normalised Euclidian distance between each of said historical records and said partially populated target record.

11. A method, operable in an Internet transaction processing system comprising a web server responsive to requests from a web client computing device to execute a transaction, of personalising said system, comprising the steps of:

responsive to a client request from a web client computing device, obtaining historical records corresponding to previous client transactions;

partially populating a target record from data available to said system, said target record comprising a plurality of attributes;

measuring a similarity of each of said historical records to said partially populated target record, wherein said measuring step comprises calculating a normalised Euclidian distance between each of said historical records and said partially populated target record;

selecting at least one similar historical record;

populating remaining target record attributes from at least one corresponding attribute of the at least one similar historical record; and

providing said populated target record to said system as a prediction of said transaction to be executed.

12. Computer readable code stored on a computer readable storage medium for an Internet transaction processing system comprising a web server responsive to requests from a web client computing device to execute a transaction, said computer readable code comprising:

first subprocesses, responsive to said client request, for obtaining historical records corresponding to previous client transactions;

second subprocesses for partially populating a target record from data available to said sub-system, said target record comprising a plurality of attributes;

third subprocesses for measuring a similarity of each of said historical records to said partially populated target record, said third subprocesses calculating a normalised Euclidian distance between each of said historical records and said partially populated target record to identify highly similar historical records from said historical records;

fourth subprocesses for selecting at least one of said highly similar historical records;

fifth subprocesses for populating remaining target record attributes from at least one corresponding attribute of the selected highly similar historical records; and

sixth subprocesses for providing said populated target record to a personalisation sub-system as a prediction of said transaction to be executed.