Claims:We claim:

1. A method for generating a personalized Television (TV) content profile 107 for a user 105, the method comprising:
receiving, by a detection system 101, digital footprint data 209 of the user 105;
determining, by the detection system 101, one or more direct features and one or more derived features associated with the user 105 based on the digital footprint data 209;
detecting, by the detection system 101, affinity of the user 105 for one or more content types using one or more trained models 245 based on the one or more direct features and the one or more derived features,
wherein each of the one or more trained models 245 are associated with a content type, wherein, each of the one or more trained models 245 are trained based on affinity of one or more users for each of the one or more content types which is determined based on informative features,
wherein the informative features are identified based on relationship between TV viewership data 215 and digital footprint data 209 of the one or more users; and
generating, by the detection system 101, the personalized TV content profile 107, comprising list of one or more content types preferred by the user 105, based on the detected affinity of the user 105 for each the one or more content types.

2. The method as claimed in claim 1, wherein the digital footprint data 209 is obtained by at least one of capturing ad-request data, capturing cookie-based information, and e-commerce data from one or more websites and applications used by the user 105.

3. The method as claimed in claim 1, wherein the digital footprint data 209 is associated with location data 211, and travel data 213 of the user 105.

4. The method as claimed in claim 1, wherein the one or more direct features comprises at least one of, features associated with connectivity of a device for performing digital activity by the user 105, frequency of use of the device in performing digital activity by the user 105, time duration of the usage of the device in performing digital activity by the user 105, and usage patterns of one or more applications in the device associated with the user 105.

5. The method as claimed in claim 1, wherein the one or more derived features comprises at least one of, features associated with home and office location of the user 105, travelling behavior of the user 105, demographics data associated with location of the user 105, device preferences, and application preferences.

6. The method as claimed in claim 5, wherein the device preferences include at least one of, make of device, category of the device, carrier information, and internet connectivity features, wherein the category of the device includes one of premium, medium price and economy price.

7. The method as claimed in claim 1, wherein the one or more content types are associated with one or more industries comprising at least one of automobile, health, finance, power, media and entertainment, manufacturing, agriculture, telecommunication, tourism, demographics and Fast-Moving Consumer Goods (FMCG).

8. The method as claimed in claim 7, wherein the personalized TV content profile 107 comprises list of one or more content types, preferred by the user, which are associated with the media and entertainment comprising genres, sub-genres, and specific events.

9. The method as claimed in claim 1, wherein the television viewership data 215 is obtained by at least one of audio fingerprinting-based tracking, set-top box-based tracking, tracking by setting up personal panel across one or more geographical locations, and tracking through one or more third parties and survey data.

10. The method as claimed in claim 1, wherein the one or more trained models 245 associated with the content type are optimal models for each of the one or more content types, wherein the one or more trained models 245 comprises one or more supervised and unsupervised machine learning models 243, and wherein the optimal models are identified from one or more models trained for each of the one or more content types.

11. The method as claimed in claim 8, wherein the optimal models are identified by:
determining a deviation in affinity predicted for content types for each of the one or more users by each of the one or more models from actual affinity of each of the one or more users for each of the one or more content types; and
identifying the one or more models as the optimal models when the determined deviation is within a predefined threshold for each content type.

12. The method as claimed in claim 1 comprises performing pre-processing of the television viewership data 215 and digital footprint data 209 of the one or more users by at least one of handling null values, treatment of outliers, transformation of encoded variables, normalization of direct features and derived features prior to identifying the informative features.

13. The method as claimed in claim 1, wherein the informative features are identified by removing uncorrelated features and redundant features from the television viewership data 215 and the digital footprint data 209.

14. A detection system 101 for generating a personalized Television (TV) content profile 107 for a user 105, the system comprises:
a processor 203; and
a memory 205 communicatively coupled to the processor 203, wherein the memory 205 stores the processor-executable instructions, which, on execution, causes the processor 203 to:
receive digital footprint data 209 of the user 105;
determine one or more direct features and one or more derived features associated with the user 105 based on the digital footprint data 209;
detect affinity of the user 105 for one or more content types using one or more trained models 245 based on the one or more direct features and the one or more derived features,
wherein each of the one or more trained models 245 are associated with a content type, wherein each of the one or more trained models 245 are trained based on affinity of one or more users for each of the one or more content types which is determined based on informative features,
wherein the processor 203 identifies informative features based on relationship between TV viewership data 215 and digital footprint data 209 of the one or more users; and
generate the personalized TV content profile 107, comprising list of one or more content types preferred by the user 105, based on the detected affinity of the user 105 for each the one or more content types.
15. The detection system 101 as claimed in claim 14, wherein the processor 203 obtains digital footprint data 209 by at least one of capturing ad-request data, capturing cookie-based information, and e-commerce data from one or more websites and applications used by the user 105.

16. The detection system 101 as claimed in claim 14, wherein the processor 203 digital footprint data 209 is associated with location data 211, and travel data 213 of the user 105.

17. The detection system 101 as claimed in claim 14, wherein the one or more direct features comprises at least one of features associated with connectivity of a device in performing digital activity by the user 105, frequency of use of the device in performing digital activity by the user 105, time duration of the usage of the device in performing digital activity by the user 105, and usage patterns of one or more applications in the device associated with the user 105.

18. The detection system 101 as claimed in claim 14, wherein the one or more derived features comprises at least one of features associated with home and office location of the user 105, travelling behavior of the user 105, demographics data associated with location of the user 105, device preferences and application preferences.

19. The detection system 101 as claimed in claim 18, wherein the device preferences include at least one of make of device, category of the device, carrier information and internet connectivity features, wherein the category of the device includes one of premium, medium price and economy price.

20. The detection system 101 as claimed in claim 14, wherein the one or more content types are associated with one or more industries comprising at least one of automobile, health, finance, power, media and entertainment, manufacturing, agriculture, telecommunication, tourism, demographics and Fast-Moving Consumer Goods (FMCG).

21. The detection system 101 as claimed in claim 20, wherein the personalized TV content profile 107 comprises list of one or more content types, preferred by the user, which are associated with the media and entertainment comprising genres, sub-genres, and specific events.

22. The detection system 101 as claimed in claim 14, wherein the processor 203 obtains television viewership data 215 by at least one of audio fingerprinting-based tracking, set-top box-based tracking, tracking by setting up personal panel across one or more geographical locations, and tracking through one or more third parties and survey data.

23. The detection system 101 as claimed in claim 14, wherein the one or more trained models 245 associated with the content type are optimal models for each of the one or more content types, wherein the one or more trained models 245 comprises one or more supervised and unsupervised machine learning models 243, and wherein the processor 203 identifies the optimal models from one or more models trained for each of the one or more content types.

24. The detection system 101 as claimed in claim 23, wherein the processor 203 identifies the optimal models by performing one or more steps comprising:
determining a deviation in affinity predicted for content types for each of the one or more users by each of the one or more models from actual affinity of each of the one or more users for each of the one or more content types; and
identifying the one or more models as the optimal models when the determined deviation is within a predefined threshold for each content type.

25. The detection system 101 as claimed in claim 14, wherein the processor 203 performs pre-processing of the television viewership data 215 and digital footprint data 209 of the one or more users by at least one of handling null values, treatment of outliers, transformation of encoded variables, normalization of direct features and derived features prior to identifying the informative features.

26. The detection system 101 as claimed in claim 14, wherein the processor 203 identifies the informative features by removing uncorrelated features and redundant features from the television viewership data 215 and the digital footprint data 209.
, Description:TECHNICAL FIELD
The present subject matter is generally related to data processing and more particularly, but not exclusively, to a method and a detection system for generating a personalized Television (TV) content profile for a user.
BACKGROUND
TV viewership persona of a user is a vital information which is of great importance specifically in the area of digital marketing. At present, existing systems track TV viewership data of users to predict preferences/ affinity of users which may be further used for personalizing the TV contents. The existing systems utilize custom hardware or software to accumulate real time TV viewership data associated with the users. However, the current methods for tracking TV viewership data are limited by installation of dedicated hardware for each TV channel subscribers, penetration of set-top boxes of partner companies and usage of partner applications. This imposes scalability challenges for predicting preferred TV contents of users. Further, existing methods of predicting users’ preferences are computationally expensive and complex, and also time consuming due to requirement of continuous monitoring of users’ activities on TV channels.
Further, a television may be used by multiple viewers in a home environment, or at commercial or public places. Although each of the viewers has unique preferences for TV contents, the existing TV content tracking system accumulate TV viewership data of all viewers cumulatively. As a result, predicted preferred TV contents become erroneous as it indicates preferences of multiple viewers instead of single viewer. In such cases, to ensure accuracy in prediction for preferred TV contents of each viewers, identification of individual TV viewership data of each viewer from the accumulated real time data is required. The existing systems fail to provide individual TV viewership data of the viewers. Further the methods to extract individual TV viewership data is time consuming and not economical.
Further, in some scenarios, users may not subscribe to TV channels for watching various TV contents which the user might be interested. In this case, the existing systems fail to predict preferred TV contents for users due to unavailability of TV viewership data for corresponding users.
The information disclosed in this background of the disclosure section is only for enhancement of understanding of the general background of the invention and should not be taken as an acknowledgement or any form of suggestion that this information forms the prior art already known to a person skilled in the art.
SUMMARY
The present disclosure discloses a method for generating a personalized Television (TV) content profile for a user. The method comprises receiving, by a detection system, digital footprint data of the user. Thereafter, the method comprises determining, by the detection system, one or more direct features and one or more derived features associated with the user based on the digital footprint data. Further, the method comprises detecting, by the detection system, affinity of the user for one or more content types using one or more trained models based on the one or more direct features and the one or more derived features. Each of the one or more trained models are associated with a content type. Each of the one or more trained models are trained based on affinity of one or more users for each of the one or more content types which is determined based on informative features. The informative features are identified based on relationship between TV viewership data and digital footprint data of the one or more users. Further, the method comprises generating, by the detection system, the personalized TV content profile based on the detected affinity of the user for each the one or more content types. The personalized TV content profile comprises list of one or more content types preferred by the user.
Further, the present disclosure discloses a detection system for generating a personalized Television (TV) content profile for a user. The detection system comprises a processor and a memory communicatively coupled to the processor. To generate a personalized TV content profile, the processor receives digital footprint data of the user. Thereafter, the processor determines one or more direct features and one or more derived features associated with the user based on the digital footprint data. Further, the processor detects affinity of the user for one or more content types using one or more trained models based on the one or more direct features and the one or more derived features. Each of the one or more trained models, associated with a content type, are trained based on affinity of one or more users for each of the one or more content types which is determined based on informative features. The processor identifies informative features based on relationship between TV viewership data and digital footprint data of the one or more users. Further, the processor generates the personalized TV content profile based on the detected affinity of the user for each the one or more content types. The personalized TV content profile comprises list of one or more content types preferred by the user.
The foregoing summary is illustrative only and is not intended to be in any way limiting. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features will become apparent by reference to the drawings and the following detailed description.
BRIEF DESCRIPTION OF THE ACCOMPANYING DRAWINGS
The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate exemplary embodiments and, together with the description, explain the disclosed principles. In the figures, the left-most digit(s) of a reference number identifies the figure in which the reference number first appears. The same numbers are used throughout the figures to reference like features and components. Some embodiments of system and/or methods in accordance with embodiments of the present subject matter are now described, by way of example only, and regarding the accompanying figures, in which:
Fig.1 shows an exemplary architecture for generating a personalized TV content profile for a user in accordance with some embodiments of the present disclosure.
Fig.2 shows a block diagram of a detection system in accordance with some embodiments of the present disclosure.
Fig.3 shows a flow chart illustrating a method for generating a personalized TV content profile for a user in accordance with some embodiments of the present disclosure.
Fig.4 shows a block diagram of an exemplary computer system for implementing embodiments consistent with the present disclosure.
It should be appreciated by those skilled in the art that any flow diagrams and timing diagrams herein represent conceptual views of illustrative device embodying the principles of the present subject matter. Similarly, it will be appreciated that any flow charts, flow diagrams, state transition diagrams, pseudo code, and the like represent various processes which may be substantially represented in computer readable medium and executed by a computer or processor, whether such computer or processor is explicitly shown.

DETAILED DESCRIPTION
In the present document, the word "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any embodiment or implementation of the present subject matter described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments.
While the disclosure is susceptible to various modifications and alternative forms, specific embodiment thereof has been shown by way of example in the drawings and will be described in detail below. It should be understood, however that it is not intended to limit the disclosure to the specific forms disclosed, but on the contrary, the disclosure is to cover all modifications, equivalents, and alternative falling within the scope of the disclosure.
The terms “comprises”, “comprising”, “includes”, “including” or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a setup, device, or method that comprises a list of components or steps does not include only those components or steps but may include other components or steps not expressly listed or inherent to such setup or device or method. In other words, one or more elements in a system or apparatus proceeded by “comprises… a” does not, without more constraints, preclude the existence of other elements or additional elements in the system or method.
The present disclosure relates to a method and a detection system [also referred as system] for generating a personalized TV content profile for a user. Each user may have unique preference for viewing various TV contents. To generate individual personalized TV content profile for the user, the disclosed detection system may receive digital footprint data of the user. The digital footprint data may be captured from ad-request data, cookie-based information, and e-commerce data, which is easily available from one or more websites and applications used by the user. In an embodiment, the digital footprint data may include, but not limited to, location data and travel data associated with the user. The location data and travel data may be collected to enhance accuracy in determination of affinity of users for TV content types. Upon receiving the digital footprint data, the detection system may determine one or more direct features and one or more derived features associated with the user based on the digital footprint data. The one or more direct features comprises at least one of features associated with connectivity of a device for performing digital activity by the user, frequency of use of the device in performing digital activity by the user, time duration of the usage of the device in performing digital activity by the user, and usage patterns of one or more applications in the device associated with the user. The one or more derived features comprises at least one of features associated with home and office location of the user, travelling behavior of the user, demographics data associated with location of the user, device preferences and application preferences. The device preferences may include, but not limited to, make of device, category of the device, carrier information and internet connectivity features, wherein the category of the device may be based on price of the device such highest price device may be under premium category, the devices with low price may be under medium category and devices with least price may be under economy category. The detection system may utilize the determined one or more direct features and one or more derived features to detect affinity of the user for one or more content types. The one or more content types may be associated with one or more industries. The one or more industries may include, but not limited to, automobile, health, finance, power, media and entertainment, manufacturing, agriculture, telecommunication, tourism, demographics, and Fast-Moving Consumer Goods (FMCG) industries. The one or more content types associated with media and entertainment industry may be genres, sub-genres, and specific events. As an example, the genres may include, but not limited to, sports, sitcom, news, documentary, soap opera, cartoon, travel, makeover, and kids. As an example, the sub-genres may include, but not limited to, action, drama, adventure, comedy, crime, fantasy, historical, horror, magical realism, mystery, paranoid fiction, philosophical, political, saga, satire, science fiction, social, speculative, thriller, romance and urban. As an example, the specific events may include, but not limited to, sports leagues, award shows, music concerts. The detection system may detect the affinity of the user by one or more trained models. Each of the one or more trained models are associated with a content type. Each of the one or more trained models are trained based on affinity of one or more users for each of the one or more content types. The affinity of one or more users may be determined based on informative features, which are identified based on relationship between TV viewership data and digital footprint data of the one or more users. Upon detecting affinity of the user for one or more content types, the detection system may generate the personalized TV content profile for the user. The personalized TV content profile comprises list of one or more content types preferred by the user. The personalized TV content profile is generated based on the detected affinity of the user for each the one or more content types. In an embodiment, the personalized TV content profile may be generated such as preference of the user for one or more genres, one or more sub-genres and events. In another embodiment, the personalized profile generated for the user for the content type health may be, preference of the user for one or more doctors, one or more health applications and the like.
In this manner, the present disclosure provides a method and a detection system for generating a unique personalized TV content profile for the user. in an embodiment, the disclosed detection system may not depend on TV viewership data associated with the user. Rather, it utilizes the digital footprint data associated with the user. Hence, the disclosed system’s non reliability on costly custom hardware and software for tracking TV viewership data associated with the users in real time reduces the scalability challenges associated with implementation. Thus, the disclosed system provides a cost-effective solution for generating the personalized TV content profile for the user. Further, the disclosed system generates personalized TV content profiles for users irrespective of whether the users have opted for TV channels subscriptions or not, for viewing various types of contents. In this manner, the technical limitation of the conventional system for predicting preferred TV contents of users who have unsubscribed or have not subscribed to TV related services is addressed by the disclosed system. Further, the user experience may be improved by recommending subscriptions for only preferred contents to the user based on the generated personalized TV content profile of the user by the detection system.
Fig.1 shows an exemplary architecture for generating a personalized TV content profile for a user in accordance with some embodiments of the present disclosure.
As shown in Fig.1, the architecture 100 may include a detection system 101 and a user 105. In an embodiment, the detection system 101 may be configured inside one or more devices [not shown in Fig.1] associated with the user 105 as an application. In another embodiment, the detection system 101 may be a server to which the one or more devices may be associated with. As an example, the one or more devices may include, but not limited to, smartphone, laptop, tablet, smart watch, Internet of Things (IOT) devices, and desktop computer.
In an exemplary scenario, the user 105 may not be subscribed to any Television (TV) channels for viewing various TV contents. Nevertheless, the user 105 may have unique preference for viewing one or more TV contents. In this scenario, recommendation of preferred TV contents based on TV viewership data 215 associated with the user 105 may not be feasible due to unavailability of the TV viewership data 215. To overcome this limitation, the disclosed system may utilize digital footprint data associated with the user 105 which may be obtained from partners applications, websites, and third parties to generate the personalized TV content profile 107 for the user 105. Firstly, the detection system 101 may receive digital footprint data 103 of the user 105. The digital footprint data 103 may be obtained by at least one of capturing ad-request data, cookie-based information, and e-commerce data from one or more websites and applications used by the user 105. An ad-request may occur, when information from an advertisement server is requested by web browser of a device associated with the user 105. Out of all ad-requests, viewing advertisements may provide user’s preferences for various types of advertisement contents. Cookie-based information may provide long-term records of user’s browsing activities on internet. This may indicate user’s preferences for different webpages and corresponding contents. E-commerce data may provide activities of the user 105 related to online shopping services, from which user’s preferences for various shopping items may be analyzed. The ad-request data, cookie-based information, and e-commerce data may be obtained by external partnerships of the detection system 101 with e-commerce websites, smart devices, smart home assistants, data from one or more partner applications associated with the detection system 101. In an embodiment, the digital footprint data 103 of the user 105 may also include, but not limited to, location data 211, and travel data 213 of the user 105.
Upon detecting the digital footprint data 103 of the user 105, the detection system 101 may determine one or more direct features and one or more derived features associated with the user 105 based on the digital footprint data 103. The one or more direct features may comprise at least one of features associated with connectivity of a device for performing digital activity by the user 105, frequency of use of the device in performing digital activity by the user 105, time duration of the usage of the device in performing digital activity by the user 105, and usage patterns of one or more applications in the device associated with the user 105. The one or more derived features may comprise at least one of features associated with home and office location of the user 105, travelling behavior of the user 105, demographics data associated with location of the user 105, device preferences and application preferences. The device preferences may include at least one of make of device, category of the device, carrier information and internet connectivity features. The category of the device may include one of premium, medium price and economy price. As an example, the carrier information may include one or more wireless service provider that supplies cellular network connectivity services to mobile phone and tablet subscribers.
Upon determining one or more direct features and one or more derived features, the detection system 101 may detect affinity of the user 105 for one or more content types using one or more trained models 245 based on the one or more direct features and the one or more derived features. The one or more trained models 245 may be machine learning models 243. Each of the one or more trained models 245 may be associated with a content type. As an example, each of the one or more trained models 245 may be associated with sports, news, music, action, drama, adventure, comedy, automobile, health, finance, and ecommerce. The detection system 101 may train each of the one or more trained models 245 based on affinity of one or more users for each of the one or more content types. The detection system 101 may determine the affinity of one or more users based on informative features. The detection system 101 may identify the informative features based on relationship between TV viewership data 215 and digital footprint data 103 of the one or more users. The detection system 101 may obtain the television viewership data 215 by at least one of audio fingerprinting-based tracking, set-top box-based tracking, tracking by setting up personal panel across one or more geographical locations, and tracking through one or more third parties and survey data. Further, in an embodiment, the detection system 101 may perform pre-processing of the television viewership data 215 and digital footprint data 103 of the one or more users by at least one of handling null values, treatment of outliers, transformation of encoded variables, normalization of direct features and derived features prior to identifying the informative features. In an embodiment, upon pre-processing the television viewership data 215 and digital footprint data 103 of the one or more users, the detection system 101 may identify the informative features by removing uncorrelated features and redundant features from the television viewership data 215 and the digital footprint data 103 of the one or more users. The detection system 101 may utilize identified informative features for training one or more supervised and unsupervised machine learning models 243. Upon completion of training, the detection system 101 may identify optimal models from one or more models trained for each of the one or more content types by evaluating affinity predicted by the one or more models which are trained. To identify the optimal models, the detection system 101 may determine a deviation in predicted affinity for each of the one or more content types from corresponding actual affinity for each of the one or more users. Thereafter, the detection system 101 may identify the one or more models as the optimal models when the determined deviation is within a predefined threshold for each content type.
Upon detecting affinity of the user 105, the detection system 101 may generate the personalized TV content profile 107, comprising list of one or more content types preferred by the user 105, based on the detected affinity of the user 105 for each the one or more content types. The one or more content types may indicate behavior of the user towards one or more industries. The one or more industries may include, but not limited to, automobile, health, finance, power, media and entertainment, manufacturing, agriculture, telecommunication, tourism, demographics and Fast-Moving Consumer Goods (FMCG). The one or more content types associated with media and entertainment industry may be genres, sub-genres, and specific events. As an example, the genres may be sports, sitcom, news, documentary, soap opera, cartoon, travel, makeover, and kids. As an example, the sub-genres may be action, drama, adventure, comedy, crime, fantasy, historical, historical fiction, horror, magical realism, mystery, paranoid fiction, philosophical, political, saga, satire, science fiction, social, speculative, thriller, and urban. As an example, the specific events may be sports leagues, award shows, music concerts.
The generated personalized TV content profile 107 of the user 105 may be utilized by marketing agencies, and survey agencies for customer relation management, digital marketing, content recommendation, subscription packages or personalized product offering to the user 105. The marketing agencies and survey agencies may determine target customers accurately based on the generated personalized TV content profile 107 for the user 105. The detection system is scalable and comprehensive and facilitates in identifying the personalized TV content profile 107 more accurately. The disclosed system may also generate personalized content profile 107 for the user 105 related to one or more industries for example automobile, health, finance for advertising. The disclosed system 101 may determine impact of specific events like sports leagues, large scale marketing, campaigns, geopolitical changes, economic policy changes and the like on at least one of TV viewership behavior of the user 105 and interest of the user 105 in different sectors such as automobile, finance, e-commerce. The detection system 101 may also predict traction of multiple users on large TV/online/offline events based on their digital footprint data 103.
Fig.2 shows a block diagram of a detection system in accordance with some embodiments of the present disclosure.
In some implementations, the detection system 101 may include an I/O interface 201, a processor 203, and a memory 205. The I/O interface 201 may be configured to receive digital footprint data 209, location data 211, and travel data 213 of a user 105 for whom a personalized TV content profile 107 is to be generated. Further, the I/O interface 201 may be configured to receive TV viewership data 215 and digital footprint data 209 of one or more users to train one or more machine learning models 243. The processor 203 may be configured to receive the digital footprint data 209, the location data 211, the travel data 213, and the TV viewership data 215 through the I/O interface 201. Further, the processor 203 may retrieve data from the memory 205 and interact with modules 219 and the one or more machine learning models 243 to generate personalized TV content profile 107. In the detection system 101, the memory 205 may store data 207 received through the I/O interface 201, modules 219 and the processor 203. In one embodiment, the data 207 may include digital footprint data 209, location data 211, travel data 213, TV viewership data 215 and other data 217. The other data 217 may store data, including temporary data and temporary files, generated by the modules 219 for performing the various functions of the detection system 101.
In some embodiments, the data 207 stored in the memory 205 may be processed by the modules 219 of the detection system 101. As an example, the modules 219 may be communicatively coupled to the processor 203 configured in the detection system 101. The modules 219 may be present outside the memory 205 as shown in Fig.2 and implemented as hardware. As used herein, the term modules 219 may refer to an Application Specific Integrated Circuit (ASIC), an electronic circuit, a processor 203 (shared, dedicated, or group) and memory 205 that execute one or more software or firmware programs, a combinational logic circuit, and/or other suitable components that provide the described functionality.
In some embodiments, the modules 219 may include, for example, a receiving module 221, a data enrichment module 223, a pre-processing module 225, a feature selection module 227, a modelling engine 229, an inferencing module 237, a post-processing module 239 and other modules 241. The other modules 241 may be used to perform various miscellaneous functionalities of the detection system 101. It will be appreciated that aforementioned modules 219 may be represented as a single module or a combination of different modules 219. Furthermore, a person of ordinary skill in the art will appreciate that in an implementation, the one or more modules 219 may be stored in the memory 205, without limiting the scope of the disclosure. The said modules 219 when configured with the functionality defined in the present disclosure will result in a novel hardware.
In an embodiment, the receiving module 221 may be configured to receive digital footprint data 209 of the user 105 for whom personalized TV content profile 107 is to be generated. The digital footprint data 209 may include, but not limited to, ad-request data, capturing cookie-based information, and e-commerce data from one or more websites and applications used by the user 105. The digital footprint data 209 may be obtained by external partnerships with one or more websites, application, smart devices, smart home assistants. The digital footprint data 209 may also include, but not limited to health measurement data, data from smart wearables, email data, Global Positioning System (GPS) data, data from voice assistant, which may be obtained from smart devices or Personal Computer (PC) used by the user 105 for performing digital activity. The received digital footprint data 209 may be stored as digital footprint data 209 in the memory 205.
In an embodiment, the receiving module 221 may be configured to receive TV viewership data 215 and digital footprint data 209 of one or more users for training one or more machine learning models 243. The TV viewership data 215 of one or more users may be obtained by at least one of audio fingerprinting-based tracking, set-top box-based tracking, tracking by setting up personal panel across one or more geographical locations, and tracking through one or more third parties. Personal panel may comprise of people spread across geographical location who send TV viewership data 215 samples. This may be accomplished through in-house application associated with the detection system 101, which tracks TV viewership data 215 of one or more users across country. Third party data may be obtained from vendors/partners who have setup for personal panel to obtain TV viewership data 215. The TV viewership data 215 of one or more users may be received through I/O interface 201 and may be stored as TV viewership data 215 in the memory 205.
In an embodiment, the data enrichment module 223 may receive location data 211, and travel data 213 of the user to enrich the digital footprint data 209. In another embodiment, the location data 211, and travel data 213 of the user may be part of the digital footprint data 209. . The location data 211 and travel data 213 may be utilized for determining latitude and longitude information associated with the user. As an example, the latitude and latitude and longitude information may be utilized for determining which includes, but not limited to, education levels, male-female ratio, income levels of the area in which the user resides thereby enriching the digital footprint data 209. The location data 211 may provide location of digital activity of the user 105. The location data 211 may include, but not limited to, home and office location of the user 105. The travel data 213 may provide latitude-longitude information associated with digital activities of the user 105 during travelling. The location data 211, and travel data 213 of the user 105 may be obtained from available geotagging information. The obtained location data 211, and travel data 213 may be stored as location data 211 and travel data 213 in the memory 205 respectively. Further, the data enrichment module 223 may determine one or more direct features and one or more derived features associated with the user 105 based on the digital footprint data 209. The one or more direct features and one or more derived features may be stored in other data 217 in the memory 205. The one or more direct features may comprise at least one of features associated with connectivity of a device for performing digital activity by the user 105, frequency of use of the device in performing digital activity by the user 105, time duration of the usage of the device in performing digital activity by the user 105, and usage patterns of one or more applications in the device associated with the user 105. As an example, the features associated with connectivity of the device may include, but not limited to, connectivity with 5G, 4G, 3G base stations or WiFi connectivity. The one or more derived features comprises at least one of features associated with home and office location of the user 105, travelling behavior of the user 105, demographics data associated with location of the user 105, device preferences and application preferences. As an example, travelling behavior of the user 105 may include, but not limited to, distance travelled by the user 105, mode of transport, frequency and location of travel, and daily travel distance associated with the user 105. The device preferences may include, but not limited to, make of device, category of the device, carrier information and internet connectivity features. The category of the device may include one of premium, medium price and economy price. The application preferences may include, but not limited to, category, price, popularity, and public opinion of applications preferred by users. The category of application may include, but not limited to, applications related to cooking, games, finance, utility, beauty, and communication.
In an embodiment, the pre-processing module 225 may perform pre-processing of the television viewership data 215 and digital footprint data 209 of the one or more users for training one or more machine learning models 243. The pre-processing may be performed by at least one of handling null values, treatment of outliers, transformation of encoded variables, normalization of direct features and derived features prior to identifying the informative features. As an example, if demographics data associated with the user 105 are not available at sub-district level, then those values may be imputed with the median values of the city level demographics. As an example, outliers related to distance between home and office may be handled by capping them to a predetermined value. As an example, data with missing values may be imputed with zeros.
In an embodiment, the feature selection module 227 may identify the informative features by removing uncorrelated features and redundant features from the television viewership data 215 and the digital footprint data 209 of the one or more users. As an example, to determine interest of the user 105 towards sports channels, user's activity on sports-based webpages/applications may be utilized. However, the features related to sports may not be relevant for determining interest of the user 105 towards music genre. The informative features for each of one or more content types may be identified by one or more feature selection techniques, which may include, but not limited to, Variance Inflation Factor (VIF), removing collinear data, and correlation. As an example, to determine affinity of one or more users towards one or more content types, features which have variance below a predetermined threshold, may be dropped. Further, collinearity associated with the features may be determined using correlation metrics and the features with highest degree of collinearity may be identified. For the identified features VIF may be determined and based on the determined VIF, one of collinear features may be provided as input to the modelling engine 229. Upon removing collinear features, features may be correlated based on relationship between TV viewership data 215 and digital footprint data 209 of one or more users. These correlated features may be provided as input to the modelling engine 229.
In an embodiment, the feature selection module 227 may identify informative features from the preprocessed digital footprint data 209 of the user 105. As an example, to determine whether the user 105 has affinity towards sports genre or not, the features having best correlation for sports genre may be identified from the determined one or more direct features and one or more derived features.
Example illustration for feature selection
As an example, informative features related to sports genre may be identified from direct features and derived features associated with one or more users. For illustration purpose, the number of users may be two, a first user and a second user. Direct features and derived features associated with the first user and the second user may include ‘number of ad-requests’, ‘percentage of ad- requests during prime time’, ‘percentage of ad- requests during non-prime time’, ‘percentage of ad- requests during weekdays’, ‘percentage of ad- requests during weekends’, ‘type of mobile device’ associated with the first user and the second user, ‘class of mobile device’ associated with the first user and the second user, ‘mobile network carrier’, ‘type of connection’ associated with the first user and the second user, ‘type of Operating System (OS)’ of the device associated with the first user and the second user, and ‘age of OS’ of the device associated with the first user and the second user associated with the mobile device. For example, the first user may have 150 ad-requests, 20% of ad- requests during prime time, 80% of ad- requests during non-prime time, 70% of ad- requests during weekdays, 30% of ad- requests during weekends, mobile ‘A’, high range of mobile class, ‘A’ as mobile network carrier, 3G connection, Android OS, and age of OS may be slightly old. The second user may have 300 ad-requests, 40% of ad- requests during prime time, 60% of ad- requests during non-prime time, 55% of ad- requests during weekdays, 45% of ad- requests during weekends, mobile ‘B’, medium range of mobile class, ‘B’ as mobile network carrier, 2G connection, Android OS, and age of OS may be old. Here, variance may be calculated for each of the features and the features with very low variance may be dropped. From the above example, type of OS is Android for all the users. As the feature ‘type of OS’ has zero variance, hence may be eliminated.
Upon eliminating redundant features based on variance, collinearity may be evaluated for the remaining list of features by performing correlation among the listed features. In the above example, ‘percentage of ad-requests during prime time’ may be collinear with ‘percentage of ad-requests during non-prime time’ and ‘percentage of ad- requests during weekdays’ may be collinear with ‘percentage of ad- requests during weekends’. The features ‘percentage of ad-requests during non-prime time’ and ‘percentage of ad- requests during weekdays’ may be eliminated due to high degree of collinearity. Upon eliminating the redundant features, correlation of each of the features with sports genre may be determined. The features with low coefficient of correlation may be eliminated. For example, if ‘age of OS’ (version of mobile OS) shows very low coefficient of correlation with sports genre, then it may be eliminated from the remaining list of features. Thus, the uncorrelated features may be eliminated based on correlation coefficients. Upon eliminating redundant features and uncorrelated features, the remaining list of features may be considered as informative features which may be passed on to the modeling engine.
In an embodiment, the modelling engine 229 may train one or more machine learning models 243 based on the informative features obtained from the TV viewership data 215 and digital footprint data 209 of one or more users. The modelling engine 229 may comprise a modelling framework 231, an evaluation module 233 and a selection module 235. The modelling framework 231 may train one or more supervised and unsupervised machine learning models 243 for one or more content types based on the informative features. As an example, the machine learning models 243 may include, but not limited to, Logistic Regression, Random Forest, XGBoost, Neural Networks, Support Vector Machine (SVM). Upon completion of training the one or more machine learning models 243, the evaluation module 233 may determine affinity of the one or more users utilizing the one or more trained models 245 for each of the one or more content types. Further, the evaluation module 233 may determine a deviation in the predicted affinity from actual affinity of each of the one or more users for each of the one or more content types. Thereafter, the selection module 235 may identify the one or more models as the optimal models when the determined deviation is within a predefined threshold for each content type. Further, the modelling framework 231 may include sequential modelling for handling of multiclass problem. As an example, one or more machine learning models 243 may be trained for sports content type for predicting affinity of the user for sports genre. Further, the one or more machine learning models 243 may be trained to determine degree of association of the user with sports when affinity of the user is predicted for sports genre. As an example, the determined degree of association of the user with sports may be one of high, medium, and low. Further, the one or more machine learning models 243 may be trained to identify users with high degree of association with sports. Thereafter, the one or more machine learning models 243 may be trained to classify remaining users based on their degree of association with sports between medium and low. In an embodiment, the present disclosure may use multiclass models instead of sequential modelling for handling multiclass problems.
In an embodiment, the inferencing module 237 may detect affinity of the user 105 for one or more content types using one or more trained models 245 based on the one or more direct features and the one or more derived features. The detected affinity may be provided to the post-processing module 239.
In an embodiment, the post-processing module 239 may generate the personalized TV content profile 107 based on the detected affinity of the user 105 for each the one or more content types. The personalized TV content profile 107 may comprise list of one or more content types preferred by the user 105. The one or more content types are associated with one or more industries comprising at least one of automobile, health, finance, power, media and entertainment, manufacturing, agriculture, telecommunication, tourism, demographics, and Fast-Moving Consumer Goods (FMCG). The one or more content types associated with media and entertainment industry may be genres, sub-genres, and specific events. As an example, the genres may be sports, sitcom, news, documentary, soap opera, cartoon, travel, makeover, and kids. As an example, the sub-genres may be action, drama, adventure, comedy, crime, fantasy, historical, historical fiction, horror, magical realism, mystery, paranoid fiction, philosophical, political, saga, satire, science fiction, social, speculative, thriller, and urban. As an example, the specific events may be sports leagues, award shows, music concerts.
In an embodiment, the machine learning models 243 may be trained for detecting affinity of a user 105 for one or more content types generating personalized TV. The machine learning models 243 which are trained may be verified by the evaluation module 233 for identifying the optimal models by the selection module 235. The optimal models may be utilized by the inferencing module 237 to detect affinity of the user 105 for one or more content types.
Fig.3 shows a flow chart illustrating a method for generating a personalized TV content profile 107 for a user in accordance with some embodiments of the present disclosure.
As illustrated in Fig.3, the method 300 includes one or more blocks illustrating a method for generating a personalized TV content profile 107 for a user 105. The order in which the method 300 is described is not intended to be construed as a limitation, and any number of the described method blocks can be combined in any order to implement the method. Additionally, individual blocks may be deleted from the methods without departing from the spirit and scope of the subject matter described herein. Furthermore, the method can be implemented in any suitable hardware, software, firmware, or combination thereof.
At block 301, the method may include receiving, by a detection system 101, digital footprint data 209 of the user 105. The digital footprint data 209 may be obtained by at least one of capturing ad-request data, capturing cookie-based information, and e-commerce data from one or more websites and applications used by the user 105. In an embodiment, the digital footprint data 209 may include, but not limited to, location data 211, and travel data 213 of the user 105.
At block 303, the method may include determining, by the detection system 101, one or more direct features and one or more derived features associated with the user 105 based on the digital footprint data 209. The one or more direct features may comprise at least one of features associated with connectivity of a device for performing digital activity by the user 105, frequency of use of the device in performing digital activity by the user 105, time duration of the usage of the device in performing digital activity by the user 105, and usage patterns of one or more applications in the device associated with the user 105. The one or more derived features comprises at least one of features associated with home and office location of the user 105, travelling behavior of the user 105, demographics data associated with location of the user 105, device preferences and application preferences. The device preferences include at least one of make of device, category of the device, carrier information and internet connectivity features, wherein the category of the device includes one of premium, medium price and economy price.
At block 305, the method may include detecting, by the detection system 101, affinity of the user 105 for one or more content types using one or more trained models 245 based on the one or more direct features and the one or more derived features. Each of the one or more trained models 245 may be associated with a content type. Each of the one or more trained models 245 may be trained based on affinity of one or more users for each of the one or more content types. The affinity may be determined based on informative features. The informative features may be identified based on relationship between TV viewership data 215 and digital footprint data 209 of the one or more users. The television viewership data 215 may be obtained by at least one of audio fingerprinting-based tracking, set-top box-based tracking, tracking by setting up personal panel across one or more geographical locations, and tracking through one or more third parties. The television viewership data 215 and digital footprint data 209 of the one or more users may be pre-processed by at least one of handling null values, treatment of outliers, transformation of encoded variables, normalization of direct features and derived features prior to identify the informative features. The informative features may be identified by removing uncorrelated features and redundant features from the television viewership data 215 and the digital footprint data 209. The identified informative features may be used for training one or more supervised and unsupervised machine learning models 243 to obtain one or more trained models 245. The one or more trained models 245 may be optimal models for each of the one or more content types. The optimal models may be identified from one or more models trained for each of the one or more content types. To identify the optimal models, affinity may be predicted for content types for each of the one or more users using each of the one or more models. Thereafter, a deviation in the predicted affinity from actual affinity of each of the one or more users for each of the one or more content types may be determined. When the determined deviation is within a predefined threshold for each content type, the one or more models may be identified as the optimal models.
At block 307, the method may include generating, by the detection system 101, the personalized TV content profile 107, comprising list of one or more content types preferred by the user 105, based on the detected affinity of the user 105 for each the one or more content types. The one or more content types are associated with one or more industries comprising at least one of automobile, health, finance, power, media and entertainment, manufacturing, agriculture, telecommunication, tourism, demographics, and Fast-Moving Consumer Goods (FMCG). The one or more content types associated with media and entertainment industry may be genres, sub-genres, and specific events. As an example, the genres may be sports, sitcom, news, documentary, soap opera, cartoon, travel, makeover, and kids. As an example, the sub-genres may be action, drama, adventure, comedy, crime, fantasy, historical, historical fiction, horror, magical realism, mystery, paranoid fiction, philosophical, political, saga, satire, science fiction, social, speculative, thriller, and urban. As an example, the specific events may be sports leagues, award shows, music concerts.
Computer System

Fig.4 illustrates a block diagram of an exemplary computer system 400 for implementing embodiments consistent with the present disclosure. In an embodiment, the computer system 400 may be a detection system 101 for generating a personalized TV content profile 107 for a user 105. The computer system 400 may include a central processing unit (“CPU” or “processor”) 402. The processor 402 may comprise at least one data processor for executing program components for executing user or system-generated business processes. The processor 402 may include specialized processing units such as integrated system (bus) controllers, memory management control units, floating point units, graphics processing units, digital signal processing units, etc.
The processor 402 may be disposed in communication with one or more input/output (I/O) devices (411 and 412) via I/O interface 401. The I/O interface 401 may employ communication protocols/methods such as, without limitation, audio, analog, digital, stereo, IEEE-1394, serial bus, Universal Serial Bus (USB), infrared, PS/2, BNC, coaxial, component, composite, Digital Visual Interface (DVI), high-definition multimedia interface (HDMI), Radio Frequency (RF) antennas, S-Video, Video Graphics Array (VGA), IEEE 802.n /b/g/n/x, Bluetooth, cellular (e.g., Code-Division Multiple Access (CDMA), High-Speed Packet Access (HSPA+), Global System For Mobile Communications (GSM), Long-Term Evolution (LTE) or the like), etc. Using the I/O interface 401, the computer system 400 may communicate with one or more I/O devices 411 and 412.
In some embodiments, the processor 402 may be disposed in communication with a communication network 409 via a network interface 403. The network interface 403 may communicate with the communication network 409. The network interface 403 may employ connection protocols including, without limitation, direct connect, Ethernet (e.g., twisted pair 10/100/1000 Base T), Transmission Control Protocol/Internet Protocol (TCP/IP), token ring, IEEE 802.11a/b/g/n/x, etc.
The communication network 409 can be implemented as one of the several types of networks, such as intranet or Local Area Network (LAN) and such within the organization. The communication network 409 may either be a dedicated network or a shared network, which represents an association of several types of networks that use a variety of protocols, for example, Hypertext Transfer Protocol (HTTP), Transmission Control Protocol/Internet Protocol (TCP/IP), Wireless Application Protocol (WAP), etc., to communicate with each other. Further, the communication network 409 may include a variety of network devices, including routers, bridges, servers, computing devices, storage devices, etc.
In some embodiments, the processor 402 may be disposed in communication with a memory 405 (e.g., RAM 413, ROM 414, etc. as shown in Fig. 4) via a storage interface 404. The storage interface 404 may connect to memory 405 including, without limitation, memory drives, removable disc drives, etc., employing connection protocols such as Serial Advanced Technology Attachment (SATA), Integrated Drive Electronics (IDE), IEEE-1394, Universal Serial Bus (USB), fiber channel, Small Computer Systems Interface (SCSI), etc. The memory drives may further include a drum, magnetic disc drive, magneto-optical drive, optical drive, Redundant Array of Independent Discs (RAID), solid-state memory devices, solid-state drives, etc.
The memory 405 may store a collection of program or database components, including, without limitation, user /application 406, an operating system 407, a web browser 408, mail client 415, mail server 416, web server 417 and the like. In some embodiments, computer system 400 may store user /application data 406, such as the data, variables, records, etc. as described in this invention. Such databases may be implemented as fault-tolerant, relational, scalable, secure databases such as OracleR or SybaseR.
The operating system 407 may facilitate resource management and operation of the computer system 400. Examples of operating systems include, without limitation, APPLE MACINTOSHR OS X, UNIXR, UNIX-like system distributions (E.G., BERKELEY SOFTWARE DISTRIBUTIONTM (BSD), FREEBSDTM, NETBSDTM, OPENBSDTM, etc.), LINUX DISTRIBUTIONSTM (E.G., RED HATTM, UBUNTUTM, KUBUNTUTM, etc.), IBMTM OS/2, MICROSOFTTM WINDOWSTM (XPTM, VISTATM/7/8, 10 etc.), APPLER IOSTM, GOOGLER ANDROIDTM, BLACKBERRYR OS, or the like. A user interface may facilitate display, execution, interaction, manipulation, or operation of program components through textual or graphical facilities. For example, user interfaces may provide computer interaction interface elements on a display system operatively connected to the computer system 400, such as cursors, icons, check boxes, menus, windows, widgets, etc. Graphical User Interfaces (GUIs) may be employed, including, without limitation, APPLE MACINTOSHR operating systems, IBMTM OS/2, MICROSOFTTM WINDOWSTM (XPTM, VISTATM/7/8, 10 etc.), UnixR X-Windows, web interface libraries (e.g., AJAXTM, DHTMLTM, ADOBE® FLASHTM, JAVASCRIPTTM, JAVATM, etc.), or the like.

Furthermore, one or more computer-readable storage media may be utilized in implementing embodiments consistent with the present invention. A computer-readable storage medium refers to any type of physical memory on which information or data readable by a processor may be stored. Thus, a computer-readable storage medium may store instructions for execution by one or more processors, including instructions for causing the processor(s) to perform steps or stages consistent with the embodiments described herein. The term “computer-readable medium” should be understood to include tangible items and exclude carrier waves and transient signals, i.e., non-transitory. Examples include Random Access Memory (RAM), Read-Only Memory (ROM), volatile memory, nonvolatile memory, hard drives, Compact Disc (CD) ROMs, Digital Video Disc (DVDs), flash drives, disks, and any other known physical storage media.
Advantages of the embodiment of the present disclosure are illustrated herein.
In an embodiment, the present disclosure provides a method and a detection system for generating a personalized TV content profile for a user.

In an embodiment, the present disclosure provides a detection system which requires very less amount of TV viewership data to train one or more machine learning models for predicting affinity of the user for one or more content types. This eliminates requirement of tracking TV Viewership data for larger segment.

In an embodiment, the present disclosure provides a method to cover a larger segment of digitally active audience, without the need for exclusive technology to track their TV viewership data.

In an embodiment, the present disclosure provides a detection system for generating a personalized TV content profile for a user without tracking TV viewership data. As the disclosed system does not rely on custom hardware and software, the disclosed system is computationally efficient and cost effective.

In an embodiment, the present disclosure provides a detection system which is scalable and comprehensive. This provides an economical solution to marketing agencies and survey agencies to reduce overall expenditure including marketing cost associated with advertisements and services. The detection system determines target customers without need of custom technology to be present with users. Also, the present disclosure facilitates in identifying the personalized TV content profile 107 more accurately.

In an embodiment, in the present disclosure, the detection system improves the user experience by recommending subscriptions for only preferred contents to the user based on the generated personalized TV content profile of the user by the detection system.

The terms "an embodiment", "embodiment", "embodiments", "the embodiment", "the embodiments", "one or more embodiments", "some embodiments", and "one embodiment" mean "one or more (but not all) embodiments of the invention(s)" unless expressly specified otherwise.

The terms "including", "comprising", “having” and variations thereof mean "including but not limited to", unless expressly specified otherwise. The enumerated listing of items does not imply that any or all the items are mutually exclusive, unless expressly specified otherwise.

The terms "a", "an" and "the" mean "one or more", unless expressly specified otherwise.
A description of an embodiment with several components in communication with each other does not imply that all such components are required. On the contrary, a variety of optional components are described to illustrate the wide variety of possible embodiments of the invention.

When a single device or article is described herein, it will be clear that more than one device/article (whether they cooperate) may be used in place of a single device/article. Similarly, where more than one device or article is described herein (whether they cooperate), it will be clear that a single device/article may be used in place of the more than one device or article or a different number of devices/articles may be used instead of the shown number of devices or programs. The functionality and/or the features of a device may be alternatively embodied by one or more other devices which are not explicitly described as having such functionality/features. Thus, other embodiments of the invention need not include the device itself.

Finally, the language used in the specification has been principally selected for readability and instructional purposes, and it may not have been selected to delineate or circumscribe the inventive subject matter. It is therefore intended that the scope of the invention be limited not by this detailed description, but rather by any claims that issue on an application based here on. Accordingly, the embodiments of the present invention are intended to be illustrative, but not limiting, of the scope of the invention, which is set forth in the following claims.

While various aspects and embodiments have been disclosed herein, other aspects and embodiments will be apparent to those skilled in the art. The various aspects and embodiments disclosed herein are for purposes of illustration and are not intended to be limiting, with the true scope and spirit being indicated by the following claims.

Referral Numerals:

Reference Number Description
100 Architecture
101 Detection system
103,209 Digital footprint data
105 User
107 Personalized TV content profile
201 I/O interface
203 Processor
205 Memory
207 Data
211 Location data
213 Travel data
215 Television viewership data
217 Other data
219 Modules
221 Receiving module
223 Data enrichment module
225 Pre-processing module
227 Feature selection module
229 Modelling engine
231 Modelling framework
233 Evaluation module
235 Selection module
237 Inferencing module
239 Post-processing module
241 Other modules
243 Machine learning models
245 Trained models
400 System
401 I/O Interface
402 Processor
403 Network interface
404 Storage interface
405 Memory
406 User/Application
407 Operating system
408 Web browser
409 Communication network
411 Input device
412 Output device
413 RAM
414 ROM
415 Mail client
416 Mail server
417 Web server