Invention Grant
- Patent Title: Recommending media items based on take rate signals
-
Application No.: US15791392Application Date: 2017-10-23
-
Publication No.: US10405016B2Publication Date: 2019-09-03
- Inventor: Angadh Singh , Carlos Gomez-Uribe
- Applicant: NETFLIX, INC.
- Applicant Address: US CA Los Gatos
- Assignee: NETFLIX, INC.
- Current Assignee: NETFLIX, INC.
- Current Assignee Address: US CA Los Gatos
- Agency: Artegis Law Group, LLP
- Main IPC: G06F3/00
- IPC: G06F3/00 ; G06F13/00 ; H04N5/445 ; H04N21/25 ; H04N21/258 ; H04N21/61 ; H04N21/262 ; H04N21/2668 ; H04N21/482

Abstract:
In an approach, a method comprises using a server computer in a media content delivery system that is configured to selectively deliver a particular media title from among a library of titles, for a source title, generating title data that specifies an order of a plurality of titles that are related to the source title based on a plurality of stored probability values; wherein each probability value in the plurality of probability values represents, for each particular title of the plurality of titles, a likelihood of selecting the particular title after playing the source title; using the server computer, receiving title impression data, wherein the title impression data specifies a plurality of browsed titles that were browsed from among the plurality of titles but may have not been selected for interaction; using the server computer, receiving title interaction data, wherein the title interaction data specifies a plurality of selected titles that were selected for interaction from the plurality of browsed titles; based on the title interaction data, the title impression data and a statistical model, re-calculating the plurality of probabilities.
Public/Granted literature
- US20180048925A1 RECOMMENDING MEDIA ITEMS BASED ON TAKE RATE SIGNALS Public/Granted day:2018-02-15
Information query