Abstract

To reduce the human intervention in the preference measure process,thisarticle proposes a preference collaborative measure framework based on anupdated belief system,which is also capable of improving the accuracy andefficiency of preferen-ce measure algorithms.Firstly,the distance of rules andthe average internal distance of rulesets are proposed for specifying therelationship between the rules.For discovering the most representativepreferences that are common in all users,namely common preference,a algorithmbased on average internal distance of ruleset,PRA algorithm,is proposed,whichaims to finish the discoveryprocess with minimum information lossrate.Furthermore,the concept of Common belief is proposed to update the beliefsystem,and the common preferences are the evidences of updated beliefsystem.Then,under the belief system,the proposed belief degree and deviationdegree are used to determine whether a rule confirms the belief system or notand classify the preference rules into two kinds(generalized orpersonalized),and eventually filters out Top-K interesting rules relying onbelief degree and deviation degree.Based on above,a scalable interestingnesscalculation framework that can apply various formulas is proposed foraccurately calculating interestingness in different conditions.At last,IMCosalgorithm and IMCov algorithm are proposed as exemplars to verify the accuracyand efficiency of the framework by using weighted cosine similarity andcorrelation coefficients as belief degree.In experiments,the proposedalgorithms are compared to two state-of-the-art algorithms and the results showthat IMCos and IMCov outperform than the other two in most aspects.