METHOD FOR PROBABILISTIC ANALYSIS OF MOST FREQUENTLY OCCURRING ELECTRONIC MESSAGE ADDRESSES WITHIN PERSONAL STORE(,PST) FILES TO DETERMINE OWNER WITH CONFIDENCE FACTOR BASED ON RELATIVE WEIGHT AND SET OF USER-SPECIFIED FACTORS

Patent №

US 7,636,734

Granted

2009-12-22

Filed 2005

Owner

MICROSOFT CORPORATION

AI components

2

kr · hardware

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

11159449

A probabilistic process to determine the owner of an electronic file, such as a Personal Store (.pst). A weighted analysis of multiple factors is performed including the operating system file owner, a user running the process, a “top Y most frequently occurring addresses” when analyzing “X number of sent then received messages,” and a number of occurrences of each “top Y most frequently occurring address.” Other factors, such as the ability to resolve against a directory service may be used. Each of the “top Y most frequently occurring addresses” is analyzed to calculate its weight according to a predetermined relationship and the address is compared to the operating system file owner and a logged-on user value. If there is a match, that value is returned as the probable owner of the file.

Knowledge representationAI hardwareH04L 67/06G06F 21/604H04L 51/00Y10S 707/99937Y10S 707/99943Y10S 707/99945Y10S 707/99948

AI classification

AI hardware0.97
Knowledge representation0.81
Machine learning0.10
Natural language0.09
Vision0.02
Evolutionary computation0.00
Planning0.00
Speech0.00

Ownership

MICROSOFT CORPORATION

assignment · 170630799

Assignors

JOHNSON, CHRISTOPHER L., WEBB, RICHARD DEAN, KEARNEY, DAVID THOMAS

On an employer assignment, the assignors are typically the inventors.

© 2026 NYSGPT2525 LLC