|
Title:
|
A SCHEME FOR APPROXIMATE MATCHING EVENT ANNOUNCEMENTS TO A CUSTOMER DATABASE |
|
Author(s):
|
Ray R. Hashemi , John R. Talburt |
|
ISBN:
|
972-99353-0-0 |
|
Editors:
|
Pedro Isaías and Nitya Karmakar |
|
Year:
|
2004 |
|
Edition:
|
1 |
|
Keywords:
|
Approximate record matching, Event processing, Weight-based matching scheme, and accuracy. |
|
Type:
|
Full Paper |
|
First Page:
|
123 |
|
Last Page:
|
130 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
An event is a short and data-rich document and it refers to an instance of an announcement type such as wedding, "birth", "graduation", "auction", "obituary", divorce, etc. In this research we focus on the events reported in newspapers. To obtain data from an event three steps are involved: (1) Obtaining a set of events for a given announcement from a group of newspapers, (2) extracting the features (data) of each event and build event records, and (3) approximate matching of the event records to an existing customer database. We have completed the first and the second steps and reported previously. The completion of the third step is the focus of this paper. The approximate matching scheme that is introduced in this paper is a weight-based scheme in which the degree of memberships for specific attribute values of an event record partially influence the discrimination among the candidate records. This work is an exploratory study that constitutes the last part of a larger research project. Although the data sample for testing the proposed scheme is small, the results reveal the fact that proposed scheme for approximate matching is an effective one. |
|
|
|
|
|
|
