METHOD FOR COLLABORATIVELY FILTERING INFORMATION IN USE OF PERSONALIZED REGRESSION WITH AUXILIARY INFORMATION TO PREDICT PREFERENCE GIVEN BY USER OF ITEM TO THE ITEM AND COMPUTING APPATARUS APPARATUS USING THE SAME

摘要

The present invention relates to a method of purifying information in order to predict a user's preference given to the item, and a computing device using the method. According to the present invention, the computing device obtains the data r _ui of the preference degree previously assigned by the individual user u with respect to the individual item i, (U? U, i? I) Estimator of Where U denotes a set of individual users, I denotes a set of individual items, and r _ui denotes an observed value of R _ui , which is a random variable indicating the preference given to the individual user u by the individual user u , X _u refers to additional information that depends on the individual user u, z _i refers to additional information that depends on the individual item i, and then the computing device calculates an estimate of the mean u _ui The residual And calculates the preference distribution of each user, which is a dispersion about the preference of the user u, Estimates a matrix [phi] using the residual, And the covariance matrix , And calculates the conditional expectation value of R _ui as the estimated preference data of the specific user u regarding i, which is each item of at least one of the individual items, .