Randall J Calistri-Yeh

US Patent:

7912868, Mar 22, 2011

Filed:

May 11, 2005

Appl. No.:

11/126184

Inventors:

Randall Jay Calistri-Yeh - Florham Park NJ, US
Wen Ruan - Manlius NY, US
Clinton P. Mah - Fairfax VA, US
Mary Louise Howatt - Toronto, CA
Mary E. McKenna - Natick MA, US
Maurice J. Forrester - Syracuse NY, US
Bo Yuan - Webster NY, US
David L. Snyder - Pittsford NY, US

Assignee:

Textwise LLC - Syracuse NY

International Classification:

G06F 17/30
G06F 17/27

US Classification:

707803, 704 9

Abstract:

An advertisement placement method and system for relating an advertisement to a dataset based on a trainable semantic vector (TSV) associated with the dataset and respective semantic representations of the advertisements. The trainable semantic vector associated with the dataset is generated based on at least one data point included in the dataset and known relationships between predetermined data points and predetermined categories. A comparison process is performed to determine a similarity between the trainable semantic vector associated with the dataset and the semantic representation associated with each of the plurality of advertisements. The system selectively relates one or more of the advertisements with the dataset based on a result of the comparison process. The selected advertisement or advertisements may be displayed with the dataset.