# US 11,451,875 B2 · Machine learning-based approach to demographic attribute inference using time-sensitive features > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,451,875 B2 - Document ID: US11451875B2 - Status: Granted - Kind: B2 - Application number: 16/359,684 - Filed: 2019-03-20 - Published: 2019-12-05 - Granted: 2022-09-20 - Earliest claimed: 2018-06-04 - Assignees: Samsung Electronics Co., Ltd. - Applicants: Samsung Electronics Co., Ltd. - Inventors: Hyun-Woo Kim - Web page: https://patentgrep.com/patent/US11451875B2 ## CPC - H04N 21/466: Learning process for intelligent management, e.g. learning user preferences for recommending movies {(services using the results of monitoring in broadcast systems H04H60/61)} - H04N 21/4826: {using recommendation lists, e.g. of programmes or channels sorted out according to their score} - G06N 20/00: Machine learning - G06N 20/20: Ensemble learning - H04N 21/44204: {Monitoring of content usage, e.g. the number of times a movie has been viewed, copied or the amount which has been watched (monitoring of user activities for profile generation for accessing a video database G06F16/739; protecting generic digital content where the protection is independent of the precise nature of the content G06F21/10; arrangements for monitoring the use made of the broadcast services in broadcast systems H04H60/31)} - H04N 21/44213: {Monitoring of end-user related data (arrangements for monitoring the users' behaviour or opinions in broadcast systems H04H60/33)} - H04N 21/4751: {for defining user accounts, e.g. accounts for children} - H04N: PICTORIAL COMMUNICATION, e.g. TELEVISION - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract A system and method for predicting user demographics based on a user's television, or media, viewing habits using machine learning algorithms is provided. A method of predicting a user's demographics comprises acquiring training data including one or more household data, person identification data, program title data, or watch time data. The method includes assessing a set of features. In addition, the method includes training one or more models based on the training data and set of features. The method includes acquiring viewing history data associated with at least one user. The method further includes determining one or more attributes associated with the at least one user based on inputting the viewing history data into the one or more models. ## Claims ### Claim 1 A method for determining user attributes based on television viewing habits, the method comprising: acquiring, by a computing system, training data including one or more of household data, person identification data, program title data, or watch time data, assessing, by a processor of the computing system, a set of features including a first set of features associated with one or more specific time slices, training, by the processor of the computing system, a learning model to determine demographic attributes specific to particular users within a household based on the training data, the set of features, and demographic labels corresponding to possible demographic attributes, transferring, by the processor of the computing system, the learning model to a predictive model, acquiring, by the computing system, viewing history data from at least one communal electronic device that is viewed by multiple users within a specified household, each of the multiple users having one or more unknown demographic attributes associated with and specific to that user, and determining, by the processor of the computing system, the one or more unknown demographic attributes, from among the possible demographic attributes, associated with and specific to at least one of the multiple users within the specified household based on inputting the viewing history data from the at least one communal electronic device into the predictive model. ### Claim 2 (depends on 1) The method of claim 1, wherein the one or more specific time slices include at least one of: (1) 48 time units of 30 minutes each, (2) 24 time units of one hour each, (3) parts of a day, (4) days of a week, (5) a weekday, or (6) a weekend. ### Claim 3 (depends on 1) The method of claim 1, wherein the set of features includes a second set of features associated with one or more keywords for one or more television programs. ### Claim 4 (depends on 1) The method of claim 1, wherein the set of features includes a second set of features associated with one or more television program titles. ### Claim 5 (depends on 1) The method of claim 1, wherein the viewing history data comprises at least one of electronic programing guide (EPG) data or automatic content recognition (ACR) data. ### Claim 6 (depends on 5) The method of claim 5, wherein the viewing history data is transformed into user-centric smart data based on sessionization or hierarchical sessionization. ### Claim 7 (depends on 1) The method of claim 1, wherein the one or more unknown demographic attributes associated with and specific to the at least one of the multiple users comprises at least one of gender or age. ### Claim 8 A computing system comprising: at least one processor configured to: acquire training data including one or more of household data, person identification data, program title data, or watch time data, assess a set of features including a first set of features associated with one or more specific time slices, train a learning model to determine demographic attributes specific to particular users within a household based on utilizing machine learning with the training data, the set of features, and demographic labels corresponding to possible demographic attributes, transfer the learning model to a predictive model, acquire viewing history data from at least one communal electronic device that is viewed by multiple users within a specified household, each of the multiple users having one or more unknown demographic attributes associated with and specific to that user, and determine the one or more unknown demographic attributes, from among the possible demographic attributes, associated with and specific to at least one of the multiple users within the specified household based on inputting the viewing history data from the at least one communal electronic device into the predictive model. ### Claim 9 (depends on 8) The computing system of claim 8, wherein the one or more specific time slices include at least one of: (1) 48 time units of 30 minutes each, (2) 24 time units of one hour each, (3) parts of a day, (4) days of a week, (5) a weekday, or (6) a weekend. ### Claim 10 (depends on 8) The computing system of claim 8, wherein the set of features includes a second set of features associated with one or more keywords for one or more television programs. ### Claim 11 (depends on 8) The computing system of claim 8, wherein the set of features includes a second set of features associated with one or more television program titles. ### Claim 12 (depends on 8) The computing system of claim 8, wherein the viewing history data comprises at least one of electronic programing guide (EPG) data or automatic content recognition (ACR) data. ### Claim 13 (depends on 12) The computing system of claim 12, wherein the at least one processor is configured to transform the viewing history data into user-centric smart data based on sessionization or hierarchical sessionization. ### Claim 14 (depends on 8) The computing system of claim 8, wherein the one or more unknown demographic attributes associated with and specific to the at least one of the multiple users comprises at least one of gender or age. ### Claim 15 A non-transitory computer readable medium storing instructions that, when executed by at least one processor, cause the at least one processor to: acquire training data including one or more of household data, person identification data, program title data, or watch time data, assess a set of features including a first set of features associated with one or more specific time slices, train a learning model to determine demographic attributes specific to particular users within a household based on utilizing machine learning with the training data, the set of features, and demographic labels corresponding to possible demographic attributes, transfer the learning model to a predictive model, acquire viewing history data from at least one communal electronic device that is viewed by multiple users within a specified household, each of the multiple users having one or more unknown demographic attributes associated with and specific to that user, and determine the one or more unknown demographic attributes, from among the possible demographic attributes, associated with and specific to at least one of the multiple users within the specified household based on inputting the viewing history data from the at least one communal electronic device into the predictive model. ### Claim 16 (depends on 15) The non-transitory computer readable medium of claim 15, wherein the one or more specific time slices include at least one of: (1) 48 time units of 30 minutes each, (2) 24 time units of one hour each, (3) parts of a day, (4) days of a week, (5) a weekday, or (6) a weekend. ### Claim 17 (depends on 15) The non-transitory computer readable medium of claim 15, wherein the set of features includes a second set of features associated with one or more keywords for one or more television programs. ### Claim 18 (depends on 15) The non-transitory computer readable medium of claim 15, wherein the set of features includes a second set of features associated with one or more television program titles. ### Claim 19 (depends on 15) The non-transitory computer readable medium of claim 15, wherein the viewing history data comprises at least one of electronic programing guide (EPG) data or automatic content recognition (ACR) data. ### Claim 20 (depends on 19) The non-transitory computer readable medium of claim 19, wherein the instructions when executed further cause the at least one processor to transform the viewing history data into user-centric smart data based on sessionization or hierarchical sessionization.