Patent Number: 6,311,173

Title: Pattern recognition using generalized association rules

Abstract: A method and system for predicting an unknown value of an attribute ofinterest of a given item from a population of items, each item in thepopulation having a plurality of variable attributes including theattribute of interest. Known attribute values regarding a training sampleof items within the population including the attribute of interest arestored in a memory. The stored attribute values are processed to determineassociation rules regarding the training sample, including at least onegeneralized association rule, each association rule including one or moreconditions on one or more respective attribute values of the itemspredictive of the value of the attribute of interest, and the at least onegeneralized rule including a logical combination of a plurality of suchconditions using at least one logical operation from the group consistingof disjunction and negation. Data are received from an input device, thedata including values of at least some of the attributes of the givenitem. The association rules including the at least one generalizedassociation rule are applied to the values included in the data so as topredict the unknown value of the attribute of interest of the given item.

Inventors: Levin; Boris (Rishon le Zion, IL), Meidan; Abraham (Tel Aviv, IL), Cheskis; Alex (Ma'ale Adumim, IL)


International Classification:

Expiration Date: 10/32013