RU2595530C1 - System and method for selection of content for displaying to user - Google Patents

Abstract

FIELD: information technology.

SUBSTANCE: invention relates to data processing systems and methods for selecting optimum and personalised content for display to users. System includes a database of users, database of content, database of templates, internal logic unit, statistics unit and interface unit. In sets of characteristics of user profiles and units of content include constant, time and calculated characteristics. Invention provides calculation of relevance of content units of users and storage of content lists for display of each user in queue tables. Data related with history showing content, is stored and processed to make changes to field of statistical data of user profiles and content with possibility of periodical analysis and calculation of characteristics of users and corresponding content.

EFFECT: technical result consists in reduction of time for processing a request, simple and cheap system, reducing required amount of memory and processor performance.

25 cl, 2 dwg

Description

The invention relates to systems and methods for processing data intended for managerial, commercial and prognostic purposes, in particular for selecting optimal and personalized content for display to users, as well as for predicting and controlling the response to such a display.

Recently, the use of mobile devices and the Internet as information channels has become a necessary component of any advertising campaign. It is known that the advertiser evaluates the effectiveness of the information channel, in particular, by indicators such as clickthrough rate and conversion rate. The higher these indicators, the more effective the information channel. The choice of personalized content for display to the user allows you to increase these indicators, that is, the user is more likely to respond to the proposed content. However, forecasting and managing these indicators is still one of the most important tasks in the field of marketing.

There is a known system and method of providing advertising (RF patent 2459256, IPC Н04М 11/00, priority 10.11.2010, publ. 08.20.2012), allowing to select content, in particular personalized advertising messages. The known system includes a database of user profiles and an advertising selection module that selects an advertising message taking into account the user profile and requirements of the advertiser. The known method includes selecting an advertising message using the advertising selection module immediately after receiving a request for information about a user's balance. The known system and method does not allow to evaluate the response to the proposed advertising and, if necessary, adjust accordingly the selection algorithm and the content of the selected content. A known system and method for providing personalized content, in particular advertising messages to users (patent EP 1940120, IPC H04L 29/06, G06Q 30/00, priority 12/12/2007, publ. 04/14/2010), allowing you to select advertising messages that take into account the user profile. The known system includes a database of user profiles, a content database that serves to store multiple units of content with characteristics associated with them, and a module for selecting advertising messages that selects an advertising message based on the user profile. When implementing the known method, the selection of an advertising message is performed using the module for selecting advertising also immediately after receiving a request from the provider for the selection of an advertising message.

As the closest analogue, the well-known system and method of selecting an advertising message for displaying to the user were selected (patent US 8438159, IPC G06F 17/30, G06F 3/00, priority September 26, 2011, publ. May 7, 2013), allowing to choose the optimal content taking into account display conditions and user profile. To implement the known method, a system is used that includes a database of user profiles, a content base and a module for selecting an advertising message, which makes it possible to select an advertising message for display to the user taking into account the user profile and characteristics of the content. The choice of an advertising message for display to the user, as well as in the above-mentioned known systems, is carried out immediately after receiving a request to select an advertising message.

A common drawback of all known methods is the selection of content for display in real time, that is, immediately after receiving the request. This method requires very fast processing of the incoming request and, therefore, for its implementation, a system with a very high performance and a large amount of RAM is required. When selecting content in real time, it is difficult to use selection algorithms that require analysis of large amounts of data and time. In addition, none of the known systems allows predicting the response to the proposed content from an individual user or a specific group of users in advance.

The authors were faced with the task of developing a system and method of selecting content for display to a user that would overcome the disadvantages of known technical solutions, using complex content selection algorithms taking into account user profile characteristics, content characteristics, statistical data, predicting the performance of displaying content and managing these indicators, while reducing query processing time, simplifying the system and its requirements, and, as a result, cheaper lenii implementation of the proposed technical solutions.

The problem is solved in that the content selection system for displaying to the user includes a user base consisting of a plurality of user profiles, each specified profile includes at least a user identifier and a set of fields containing sets of user characteristics and statistics, a content base consisting of multiple profiles of content units, each specified profile includes at least the identifier of the content unit and a set of fields containing sets of characteristics of the content unit statistics, a template database containing, at a minimum, a user profile template, a content unit profile template and a characteristics catalog, an internal logic block containing at least a controller, a content selection module, an analysis and control module, and a plurality of queue tables used to store lists of content units to be displayed to users, a statistics block containing an event log and an event log processing module that performs the functions of storing and processing data related to the history of displaying user content m, and changes in the field of statistics, and the interface unit, adapted to obtain information from an external device on the results of the content display to the user. At the same time, sets of characteristics of user profiles and sets of characteristics of profiles of content units include at least permanent, temporary and calculated characteristics, and lists of content units in the queue tables include attributes that determine the priority of delivery. The system may further comprise a cluster block containing a target table and a plurality of clusters. A method for selecting content to be shown to users, for which a content selection system is used for showing to users, includes creating a user base, a content base, a base of templates and a catalog of characteristics, calculating the relevance of a content unit to a user using the content selection module, and saving content lists for display for each user in the queue tables, receiving a request to select content for display to a user, selecting content from a pre-prepared list for that user In the queue tables, retrieving the selected content from the content database and transferring the specified content for display to the user, storing data related to the history of the display of the specified content in the event log, and processing the specified data using the event log processing module, making changes to the fields of the statistical data of user profiles and content according to the control signal from the event log processing module, periodic calculation of user and content characteristics by means of the analysis and control module and making changes to the fields of characteristics of user profiles and content by a control signal from the analysis and control module. The sets of characteristics of user profiles and profiles of content units include, at a minimum, permanent, temporary and calculated characteristics, which are either assigned manually or calculated automatically by the analysis and control module. The relevance of the content unit to the user is calculated taking into account the characteristics of the user profiles and content, and the lists of content for display to the user are stored in the queue tables with the assignment of attributes that determine the priority of the display. When supplying the system with an additional cluster of clusters, the method of selecting content for display to users also includes the creation of a target table for the cluster block and the formation of user clusters and content clusters based on a set of characteristics using the analysis and control module and calculating for each user cluster the most relevant content cluster using the selection module content.

The technical result of the claimed invention is to create a system and method, providing:

a) an increase in the performance indicators of the information channel when displaying content to users, in particular, clickthrough rate, in comparison with known solutions, due to the possibility of using complex and time-consuming algorithms;

b) the ability to predict the effectiveness of the information channel and manage these indicators;

c) reducing the processing time of the request, simplifying and cheapening the system, reducing the required amount of memory and processor performance.

In order to describe the claimed invention, the following concepts are used:

User - a person using a mobile device or personal computer, to whom the content is shown during the use of the information channel. The user has a unique user identifier, such as MSISDN or IP address;

Content - informational or meaningful content of the message sent to the user, for example, an advertising message. For the purposes of the claimed invention, the content contains at least one attribute that allows you to evaluate the user's reaction to the display of content, for example, a phone number, URL link, etc .;

Information channel - a network or system of a provider, for example, a mobile network or a website on the Internet. The network or system of the provider is external to the claimed system, is mentioned solely for the purpose of a better understanding of the claimed technical solution and their functioning is not included in the claimed method;

Performance indicators - characteristics of the information channel, calculated on the basis of statistical data related to the history of the content display. For example, the clickability rating of an advertising message, calculated as the ratio of the number of clicks on a given message to the total number of impressions of this message, forms an indicator of the effectiveness of the information channel through which the specified message was displayed.

In FIG. 1 shows a block diagram of a system used to implement the proposed method.

In FIG. 2 shows a block diagram of a system including a block of clusters used to implement the proposed method.

In FIG. 3 shows a timing diagram of the functioning of the claimed system in real time.

The inventive system is a combination of software and hardware interacting with each other and consists of the following elements: 1 - interface unit, 2 - content base, 3 - user base, 4 - template base, 5 - characteristics catalog, 6 - internal logic block, 7 - controller of the internal logic block, 8 - content selection module, 9 - analysis and control module, 10 - queue tables, 11 - statistics block, 12 - event log, 13 - event log processing module.

With large volumes of users and content, to simplify data processing and reduce RAM, the system is additionally equipped with a block of 14 clusters containing a table of 15 goals.

The interface unit 1 provides the interaction of the claimed system with external devices of networks and systems of providers, receiving requests for the selection of content, transferring selected content, receiving information from systems of providers about the results of displaying content.

The content base 2 is intended for creating, editing and storing multiple profiles of content units, each such profile includes at least the identifier of the content unit and a set of fields containing sets of characteristics of the content unit and statistical data.

The 3 user database is designed to create, edit, and store many user profiles; each such user profile includes at least a user identifier and a set of fields containing sets of user characteristics and statistical data.

The sets of characteristics and statistical data stored in the respective fields of the profiles of content units and user profiles contain information characterizing individual users or units of content, which makes it possible to predict the performance indicators of the information channel and manage these indicators, as well as improve the accuracy of calculating the relevance of content to the user and, as a result, increased response rates to content display.

Base 4 templates contains at least one user profile template, at least one content unit profile template and a catalog of 5 characteristics.

The catalog of characteristics 5 contains at least three tables of characteristics: a table of constant characteristics, a table of time characteristics and a table of calculated characteristics. The permanent characteristics include, in particular, the type of product or service (refrigerator, cosmetics, legal services, etc.), demographic characteristics (gender, age, etc.). Temporary characteristics include, in particular, areas of interest (hunting, traveling, luxury brands, modern music, etc.), target audience (adolescents, high-income people, children's goods, etc.). Calculated characteristics include criteria that can be calculated based on an analysis of the history of the display of content and relevant statistics, for example, innovator, conservative, easily responding to advertising, etc. The characteristics contained in the catalog tables 5 are added to user and content profiles, and the same characteristics can be linked to both the user profile and the content profile. The algorithm for adding characteristics is constructed in such a way that permanent characteristics can be added to the user profile or content only manually, for example, by the system administrator or provider. Time characteristics can be added to the user or content profile either manually, for example by the system administrator or provider, or automatically, based on the analysis of the history of the content display and the application of various algorithms and methods for calculating the optimal content. The calculated characteristics can be added to the user or content profile only automatically based on the analysis of the history of content displays and the application of various algorithms and methods for calculating optimal content. For example, to add a specific characteristic, the statistics in the user profile or content must match the predefined criteria. This automatic assignment of characteristics avoids subjectivity in the administration of content and user profiles, as well as dynamically adjust the changing interests and needs of users, the target and price categories of goods, etc. Which, in turn, helps to increase the accuracy of calculations and improve the performance indicators of the information channel.

Block 6 of the internal logic contains a controller 7, a module 8 for selecting content, a module 9 for analysis and control, and table 10 queues.

The controller 7 provides the interaction of all modules and blocks of the system.

The content selection module 8 periodically calculates the relevance of the content unit to users and compiles a list of content to display separately for each user. The list contains identifiers of a content unit, for example, a serial number and an attribute that determines the priority of displaying a given content unit. The relevance is calculated and the lists are compiled off-line or during the period of the least load using various algorithms containing the know-how of the authors and taking into account the characteristics of user profiles and content, in particular: determining the degree of relevance of a given unit of content to a given user, that is, calculating the percentage regarding the likelihood that the user will respond to the proposed content, for example, follow the link or make a call by number, and / or set the priority of the show, while There are advertiser requirements and display strategies, frequency, volume, etc.

All calculated lists are stored in tables 10 of the queue.

The analysis and control module 9, offline or during the period of the least load, on the basis of statistical data in user profiles and content, periodically calculates the time and calculated characteristics of users and content in the long term and generates control signals for making changes to the fields of characteristics of user profiles and content.

It should be noted that the applied calculation algorithms in modules 8 and 9 can be changed both by the administrator and automatically, for example, different formulas for calculation, different weighting criteria, etc. can be used.

In general, the functions of block 6 of internal logic include selecting relevant content, saving lists of selected content, transferring the corresponding unit of content from a pre-prepared list in response to a request, analyzing statistics related to the history of displaying content, and making changes to the fields of characteristics of user profiles and content. Computing that requires the involvement of significant system resources is carried out through module 8 for selecting content offline or during the period of the least load, which allows the use of algorithms of any complexity and provides more accurate results and, therefore, increases the efficiency of the information channel when displaying content to users , in particular clickthrough rate. The obtained calculation results for further use upon receipt of a request to display the content are stored in the queue tables 10. Thus, immediately at the time of receipt of the request for the content to be shown, the claimed system does not need to perform calculations, but only needs to transfer in response to the request a content unit, the first in the queue in a pre-prepared list, which reduces the processing time of the request, simplifies and cheapens the system, reduces required memory and processor performance. Using module 9 analysis and management makes it possible to predict the performance indicators of the information channel and manage these indicators by storing, analyzing and processing statistical data in user profiles and content.

The statistics block 11 contains an event log 12 and an event log processing module 13. In the event log 12, data related to the content display is stored, in particular, which unit of content was shown to which user, whether information was received from the provider about the user's reaction to the display, what the reaction was, and so on. The event log processing module 13 periodically in the off-line mode or during the period of the least load performs reading the event log 12, extracting data from it and, in accordance with the content of these data, generates control signals for making changes to the statistical data fields of user profiles and content. User profiles store information about the characteristics of the content shown to this user and the reactions of the specified user to the displayed content. In the profiles of content units, they store information about the characteristics of users who were shown the specified unit of content, and user reactions to the display of this content. This method of preserving the history of displaying content can significantly reduce the required memory by storing partially processed statistics.

The inventive system can be additionally equipped with a block of 14 clusters, consisting of many user lists and content lists (clusters) and table 15 goals. Table 15 goals is managed by the administrator of the system and contains a list of goals, strategies and / or features by which these clusters are formed. The use of block 14 clusters is advisable for large volumes of users and content since it allows to simplify data processing and reduce RAM by grouping users and content into large groups according to various criteria.

The implementation of the proposed method provides for two modes of functioning of the system, proceeding in parallel and independently from each other - work in real time and work in the background. A real-time timing diagram is shown in FIG. 3, where 16 is a request for showing content to a user, the specified request includes, at a minimum, the user ID received by the interface unit 1 and transmitted to the controller 7 of the internal logic block, 17 - the request in the table 10 of the queue identifies the content unit for the specified user, 18 - a response message containing the identifier of the content unit, the first in the list for the specified user, 19 - retrieving the content unit from the database 2 of the content using the specified identifier, 20, 21 - transmitting the corresponding content unit in response to request, 22 - signal to save data about the event in the log of 12 events, 23 - information from the provider about the results of the show, 24 - signal to save data about the results of the show. In the background, the creation and editing of a database of 2 content, a base of 3 users, a base of 4 templates, a catalog of 5 characteristics, assignment of permanent and temporary characteristics in user and content profiles, calculation of relevant content in module 8 for selecting content, saving lists of selected content in tables 10 are performed queues, analysis in module 9 of analysis and management of statistical data related to the history of displaying content and making changes to the fields of characteristics and statistical data of user profiles and content and data processing from the log 12 events in the module 13, event log processing. All operations in the background are carried out offline or during the period of least load. Elements of the system interact with each other during operation in both modes, in particular, use the results of individual actions.

The implementation of the proposed method includes, at a minimum, the following operations, the implementation of which can be performed both sequentially and in parallel, independently of each other:

a) create a base of 2 content, a base of 3 users, a base of 4 templates, a catalog of 5 characteristics;

b) calculate, by means of module 8, the selection of content, the relevance of the units of content to users and compile lists of content for display to users with assignment of priority to display in the table 10 queue;

c) calculate, using the analysis and control module 9, the characteristics of users and content, based on the statistics in user and content profiles, generate a control signal and, based on the specified signal, make changes to the characteristics fields of user and content profiles in content base 2 and base 3 users

d) receive a request, by means of interface unit 1, to select the content to be shown to the user, select 17 content identifiers to be shown to the specified user from a pre-prepared list in the queue table 10 for the specified user, extract 19, 20 the selected content by the specified identifier from base 2 content, transmit 21, by means of the interface unit 1, the specified content;

e) storing 22, 24 in the event log 12 data related to the history of the display of the specified content, and periodically process the specified data by means of the event log processing module 13;

f) make changes to the fields of statistical data of user and content profiles in the content database 2 and the user database 3 based on a control signal from the event log processing module 13.

When using block 14 clusters, the following operations are additionally performed:

g) create a table of 15 goals;

h) form, by means of analysis and management module 9, user lists and content lists (clusters), said formation includes, but is not limited to, grouping users and / or content according to a set of characteristics of user profiles and content;

i) calculate, using the module 8 content selection, for each cluster of users the most relevant cluster of content with the subsequent calculation for each user from the cluster list of the most relevant units of content for the formation of the table 10 queue.

To form clusters, the characteristics of user profiles and content are divided into two types: the first type includes permanent and temporary characteristics, and the second includes the calculated characteristics of user profiles and content. The formation of clusters according to the first type of characteristics to a greater extent affects the simplification of data processing and a decrease in the amount of required memory with large volumes of processed information. The use of clusters, the formation of which take into account the characteristics of the second type, provides the ability to predict the effectiveness of the information channel and manage these indicators.

Claims (25)

1. The system of selecting content for display to users, including:
a user database consisting of many user profiles, each specified profile includes at least a user identifier and a set of fields containing sets of user characteristics and statistical data,
a content base consisting of a plurality of content unit profiles, each specified profile includes at least a content unit identifier and a set of fields containing sets of content unit characteristics and statistics,
a template database containing, at a minimum, a user profile template, a content unit profile template, a catalog of characteristics,
block of internal logic containing at least
a controller that ensures the interaction of all modules and blocks of the system,
a content selection module that performs the function of calculating relevant content and compiling a list of content for display to users,
an analysis and control module that performs the functions of analyzing statistical data, calculating user characteristics and content units and generating a control signal for making changes to the profile characteristics fields of the specified users and content units,
many queue tables used to store lists
units of content to be shown to users,
a statistics unit containing an event log and an event log processing module that performs the functions of storing and processing data related to the history of displaying content to users and making changes to the statistics fields of user profiles and content units, and
an interface unit that allows the system to interact with external devices. 2. The system of claim 1, wherein the sets of user profile characteristics include, at a minimum, permanent, temporary, and calculated characteristics. 3. The system of claim 1, wherein the sets of characteristics of the content unit profiles include, at a minimum, permanent, temporary, and calculated characteristics. 4. The system of claim 1, wherein the lists of units of content in the queue tables include attributes that determine the priority of delivery. 5. The system of claim 1, wherein the interface unit is configured to receive information from an external device about the results of displaying content to a user. 6. The method of selecting content for display to users, for the implementation of which they use the content selection system for display to users and perform the following actions:
create a user database consisting of many user profiles, each specified profile includes at least a user identifier and a set of fields containing sets of user characteristics and statistical data,
create a content database consisting of many profiles of content units, each specified profile includes at least the identifier of the content unit and a set of fields containing sets of characteristics of the content unit and statistical data, create a database of templates and a catalog of characteristics,
by means of the content selection module, the relevance of the content unit to the user is calculated and the lists of content for display for each user in the queue tables are stored,
receive a request, by means of the interface unit, to select content for showing to the user, select content to be shown to the specified user from a pre-prepared list in the queue tables for the specified user, select the selected content from the content database, transmit, by the means of the interface unit, the specified content,
save the data related to the history of the display of the specified content in the event log, and process the specified data by means of the event log processing module,
by a control signal from the event log processing module, make changes to the fields of statistical data of user profiles and content,
the analysis and control module means perform periodic calculation of user characteristics and content,
by a control signal from the analysis and control module, make changes to the fields of characteristics of user profiles and content. 7. The method of selecting content according to claim 6, the implementation of which includes at least permanent, temporary, and calculated characteristics in the sets of characteristics of user profiles. 8. The method of selecting content according to any one of paragraphs. 6 and 7, during the implementation of which the characteristics of user profiles are either assigned manually or calculated automatically by the analysis and control module. 9. The method of selecting content according to claim 6, the implementation of which includes at least permanent, temporary, and calculated characteristics in the sets of characteristics of profiles of content units. 10. The method of selecting content according to any one of paragraphs. 6 and 9, during the implementation of which the characteristics of the profiles of content units are either assigned manually or automatically calculated by the analysis and control module. 11. The method of selecting content according to claim 6, the implementation of which the relevance of the content unit to the user is calculated taking into account the characteristics of user profiles and content. 12. The method of selecting content according to claim 6, the implementation of which lists of content for display to users are stored in the tables of the queue with the assignment of attributes that determine the priority of the display. 13. A system for selecting content for display to users, including:
a user database consisting of many user profiles, each specified profile includes at least a user identifier and a set of fields containing sets of user characteristics and statistical data,
a content base consisting of a plurality of content unit profiles, each specified profile includes at least a content unit identifier and a set of fields containing sets of content unit characteristics and statistics,
a template database containing, at a minimum, a user profile template, a content unit profile template, a catalog of characteristics,
block of internal logic containing at least
a controller that ensures the interaction of all modules and blocks of the system,
a content selection module that performs the function of calculating relevant content and compiling a list of content for display to users,
an analysis and control module that performs the functions of analyzing statistical data, calculating user characteristics and content units and generating a control signal for making changes to the profile characteristics fields of the specified users and content units,
many queue tables used to store lists of content units for display to users,
a statistics block containing an event log and an event log processing module that performs the functions of storing and processing data related to the history of displaying content to users and making changes to the fields of statistical data of user profiles and content units,
an interface unit for interfacing the system with external devices, and
cluster block containing a goal table and multiple clusters. 14. The system of claim 13, wherein the sets of characteristics of user profiles include, at a minimum, permanent, temporary, and calculated characteristics. 15. The system of claim 13, wherein the sets of characteristics of the content unit profiles include, at a minimum, permanent, temporary, and calculated characteristics. 16. The system of claim 13, wherein the lists of content units in the queue tables include attributes defining a display priority. 17. The system of claim 13, wherein the interface unit is configured to receive information from an external device about the results of displaying content to a user. 18. The method of selecting content for display to users, for the implementation of which they use the content selection system for display to users and perform the following actions:
create a user database consisting of many user profiles, each specified profile includes at least a user identifier and a set of fields containing sets of user characteristics and statistical data,
create a content database consisting of many profiles of content units, each specified profile includes at least the identifier of the content unit and the set of fields containing sets of characteristics of the content unit and statistical data,
create a template database and a catalog of characteristics,
by means of the content selection module, the relevance of the content unit to the user is calculated and the lists of content for display for each user in the queue tables are stored,
receive a request, by means of the interface unit, to select content for showing to the user, select content to be shown to the specified user from a pre-prepared list in the queue tables for the specified user, select the selected content from the content database, transmit, by the means of the interface unit, the specified content,
save the data related to the history of the display of the specified content in the event log, and process the specified data by means of the event log processing module, by the control signal from the event log processing module, make changes to the statistical data fields of user and content profiles,
the analysis and control module means perform periodic calculation of user characteristics and content,
by a control signal from the analysis and control module, make changes to the fields of characteristics of user profiles and content,
create a target table of the cluster block,
using the analysis and management module means form user clusters and content clusters,
using the content selection module, for each user cluster, the most relevant content cluster is calculated. 19. The method of selecting content according to claim 18, the implementation of which includes at least permanent, temporary and calculated characteristics in the sets of characteristics of user profiles. 20. The method of selecting content according to any one of paragraphs. 18 and 19, during the implementation of which the characteristics of user profiles are either assigned manually or calculated automatically by the analysis and control module. 21. The method of selecting content according to claim 18, the implementation of which includes at least permanent, temporary, and calculated characteristics in the sets of characteristics of profiles of content units. 22. The method of selecting content according to any one of paragraphs. 18 and 21, during the implementation of which the characteristics of the profiles of content units are either assigned manually or calculated automatically by the analysis and control module. 23. The method of selecting content according to claim 18, the implementation of which the relevance of the content unit to the user is calculated taking into account the characteristics of user profiles and content. 24. The method of selecting content according to claim 18, the implementation of which lists of content for display to users are stored in the queue tables with the assignment of attributes that determine the priority of the display. 25. The method of selecting content according to claim 18, the implementation of which clusters of users and clusters of content form a set of characteristics of user profiles and content.

Images