Collaborative filtering system and collaborative filtering method

When there are no evaluation values present from a user that has evaluated both content (X, Z), an indirect degree of similarity computation unit (32) of a computation processing unit (30) of an information processing center (10a) employs evaluation values of content (Y), wherein evaluation values are present from a user that has evaluated content (X, Y) and evaluation values are present from a user that has evaluated content (Y, Z), to indirectly compute degrees of similarity between the content (X, Z). A predicted evaluation value computation unit (33) employs the degrees of similarity between the content (X, Z) computed by the indirect degree of similarity computation unit (32) and the evaluation values of the content (X, Z) to compute predicted evaluation values for either of the content (X, Z) that a user has not evaluated. It is thereby possible to compute predicted evaluation values for the content (X, Z) that could not be computed directly, thus allowing an expanded range of content whereof evaluation