Content recommendation device and program

Abstract

To provide a content recommendation device and a program which can effectively utilize information on content length for selection of content.SOLUTION: A preference vector generation unit generates a preference vector on the basis of a feature vector acquired on the basis of viewing history of a user. An organization vector generation unit generates an organization vector on the basis of a feature vector acquired on the basis of content to be distributed on the basis of organization. A retrieval query generation unit generates a retrieval query on the basis of the preference vector and the organization vector. A similarity retrieval unit acquires content of similar features from a content database by using the retrieval query. A recommendation time length calculation unit calculates the total recommendation time length of content to be recommended on the basis of organization content information. A recommendation content selection unit selects content to be settled within the recommendation time length as recommendation content on the basis of the content acquired by the similarity retrieval unit.SELECTED DRAWING: Figure 1

Images