Value-added service recommendation method and system

Abstract

The embodiment of the invention discloses a value-added service recommendation method and system, and the method comprises the steps: obtaining a user portrait of a user and song information of a target song; matching a related value-added service set according to the user portrait and the song information, and taking the value-added service set as an initial value-added service queue; calculating the priority and trigger probability of each value-added service according to the user portrait, the song information and the value-added services, and rearranging the initial value-added service queue according to the priority from high to low to obtain a current value-added service queue; displaying the value-added service according to the current value-added service queue, the triggering probability and the cooling time; wherein the cooling time is preset. And the value-added service pushing accuracy and the user experience are improved.

Description

Technical Field

[0001] The present invention relates to the field of multimedia, and in particular to a value-added service recommendation method and system. Background Art

[0002] With the development of the entertainment industry, KTV song ordering systems, online KTV platforms, and music streaming platforms have gradually become an indispensable part of people's daily lives. These platforms not only provide a wealth of song resources, but also introduce various value-added services (such as membership upgrades, high-definition song unlocking, accompaniment downloads, personalized recommendations, etc.) to meet the diverse needs of users and improve the profitability of the platform. However, the existing value-added service recommendation mechanism has the following problems: it usually pushes the same value-added service frequently, which is easy to cause user disgust. In addition, it cannot recommend personalized content according to the specific situation of the user, resulting in poor user experience.

[0003] There is currently no effective solution to the above problems in the prior art. Summary of the invention

[0004] To solve the above problems, the present invention provides a value-added service recommendation method and system, which dynamically calculates the triggering probability and priority of each value-added service by combining user identity, song information and value-added service rules, and optimizes the value-added service triggering effect by combining the cooling time mechanism and adaptive algorithm, so as to solve the problems of single trigger mechanism and poor user experience in the prior art.

[0005] To achieve the above-mentioned purpose, the present invention provides a method for recommending value-added services, comprising: obtaining a user portrait of a user and song information of a target song; matching a relevant value-added service set according to the user portrait and song information, and using the value-added service set as an initial value-added service queue; calculating the priority and trigger probability of each value-added service according to the user portrait, song information and value-added service, and rearranging the initial value-added service queue in order from high to low priority to obtain a current value-added service queue; displaying the value-added services according to the current value-added service queue, trigger probability and cooling time; wherein the cooling time is pre-set.

[0006] Further optionally, the calculating the priority and trigger probability of each value-added service based on the user portrait, song information and value-added services includes: determining a preset priority of the value-added service; calculating a user information score based on the user portrait; calculating a song information score based on the song information; calculating a content relevance score based on the song ordering preference in the user portrait and the preference type in the song information; setting weight coefficients for the preset priority, user information score, song information score and content relevance score, respectively, and performing weighted summation to obtain the priority of the value-added service; and calculating the trigger probability based on the priority of the value-added service and a preset trigger threshold.

[0007] Further optionally, the display of the value-added service according to the current value-added service queue, the trigger probability and the cooling time includes: selecting the value-added service in order of priority from the current value-added service queue as the current value-added service; calculating the time difference between the trigger time corresponding to the current value-added service and the current time; when the time difference is not less than the cooling time, the cooling time ends, and judging whether the current value-added service is triggered according to the trigger probability, if triggered, the current value-added service is displayed and the trigger time is updated, if not triggered, the next value-added service is selected for cooling time judgment; otherwise, the cooling time has not ended, and the next value-added service is selected for cooling time judgment.

[0008] Further optionally, after displaying the value-added service according to the current value-added service queue, trigger probability and cooling time, it also includes: calculating user feedback information, value-added service trigger rate and value-added service conversion rate; adjusting the cooling time and / or preset priority of the corresponding value-added service according to the user feedback information; adjusting the preset trigger threshold and / or the weight coefficient according to the value-added service trigger rate; adjusting the content of the value-added service according to the value-added service conversion rate.

[0009] Further optionally, matching the relevant value-added service set based on the user portrait and song information includes: combining features in the user portrait with features in the song information according to preset rules to obtain matching conditions; matching value-added services in a value-added service database according to the matching conditions, and adding any value-added service to the value-added service set when the matching conditions meet the matching rules corresponding to any value-added service; wherein the matching rules are pre-set and are combinations of features of the user portrait and features of the song information according to preset rules.

[0010] On the other hand, the present invention also provides a value-added service recommendation system, including: an initial data acquisition module, used to obtain a user portrait of a user and song information of a target song; a value-added service matching module, used to match a related value-added service set according to the user portrait and song information, and use the value-added service set as an initial value-added service queue; a sorting module, used to calculate the priority and trigger probability of each value-added service according to the user portrait, song information and value-added service, and rearrange the initial value-added service queue in order from high to low priority to obtain a current value-added service queue; a display module, used to display the value-added service according to the current value-added service queue, trigger probability and cooling time; wherein the cooling time is pre-set.

[0011] Further optionally, the sorting module includes: a preset priority determination submodule, used to determine the preset priority of the value-added service; a user information score calculation submodule, used to calculate the user information score based on the user portrait; a song information score calculation submodule, used to calculate the song information score based on the song information; a relevance score calculation submodule, used to calculate the content relevance score based on the song request preference in the user portrait and the preference type in the song information; a weighted summation submodule, used to set weight coefficients for the preset priority, user information score, song information score and content relevance score, and then perform weighted summation to obtain the priority of the value-added service; a trigger probability calculation submodule, used to calculate the trigger probability based on the priority of the value-added service and the preset trigger threshold.

[0012] Further optionally, the display module includes: a value-added service selection submodule, which is used to select value-added services from the current value-added service queue in order of priority as the current value-added service; a time difference calculation submodule, which is used to calculate the time difference between the trigger time corresponding to the current value-added service and the current time; a judgment submodule, which is used to determine whether to trigger the current value-added service based on the trigger probability when the time difference is not less than the cooling time. If triggered, the current value-added service is displayed and the trigger time is updated. If not triggered, the next value-added service is selected for cooling time judgment; otherwise, the cooling time has not ended, and the next value-added service is selected for cooling time judgment.

[0013] Further optionally, the device also includes: an effect data calculation module, used to calculate user feedback information, value-added service trigger rate and value-added service conversion rate; a first adjustment module, used to adjust the cooling time and / or preset priority of the corresponding value-added service according to the user feedback information; a second adjustment module, used to adjust the preset trigger threshold and / or the weight coefficient according to the value-added service trigger rate; a third adjustment module, used to adjust the content of the value-added service according to the value-added service conversion rate.

[0014] Further optionally, the value-added service matching module includes: a feature combination submodule, used to combine the features in the user portrait with the features in the song information according to preset rules to obtain matching conditions; a matching submodule, used to match the value-added services in the value-added service database according to the matching conditions, and when the matching conditions meet the matching rules corresponding to any value-added service, the any value-added service is added to the value-added service set; wherein the matching rules are pre-set and are a combination of the features of the user portrait and the features of the song information according to preset rules.

[0015] The above technical solution has the following beneficial effects: through comprehensive analysis of user portraits and song information, the types of recommended song value-added services are more in line with user needs; by setting a cooling-off time, it is ensured that the same value-added service will not be recommended to users frequently, thereby ensuring the diversification of value-added service recommendations and improving user experience; by adaptively adjusting various parameters, the accuracy of value-added service recommendations is ensured, thereby improving user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 is a flow chart of a value-added service recommendation method provided by an embodiment of the present invention;

[0018] Figure 2 is a flow chart of a method for calculating priority and trigger probability provided by an embodiment of the present invention;

[0019] Figure 3 is a flow chart of a value-added service display method provided by an embodiment of the present invention;

[0020] Figure 4 is a flow chart of an adaptive adjustment method provided by an embodiment of the present invention;

[0021] Figure 5 is a flow chart of a value-added service matching method provided by an embodiment of the present invention;

[0022] Figure 6 is a structural diagram of a value-added service recommendation system provided by an embodiment of the present invention;

[0023] Figure 7 is a schematic diagram of the structure of a sorting module provided in an embodiment of the present invention;

[0024] Figure 8 is a schematic diagram of the structure of a display module provided by an embodiment of the present invention;

[0025] Fig. 9 is a schematic diagram of the structure of a module for adaptive adjustment provided by an embodiment of the present invention;

[0026] Fig.10 It is a structural diagram of a value-added service matching module provided in an embodiment of the present invention.

[0027] Figure markings: 100-initial data acquisition module; 200-value-added service matching module; 2001-feature combination submodule; 2002-matching submodule; 300-sorting module; 3001-preset priority determination submodule; 3002-user information score calculation submodule; 3003-song information score calculation submodule; 3004-correlation score calculation submodule; 3005-weighted summation submodule; 3006-trigger probability calculation submodule; 400-display module; 4001-value-added service selection submodule; 4002-time difference calculation submodule; 4003-judgment submodule; 500-effect data calculation module; 600-first adjustment module; 700-second adjustment module; 800-third adjustment module. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] In order to solve the problems in the prior art, the present invention provides a value-added service recommendation method. Figure 1 is a flow chart of a method for recommending value-added services provided by an embodiment of the present invention. Figure 1 As shown, the method includes:

[0030] S1. Obtain the user profile of the user and the song information of the target song.

[0031] User portrait is a comprehensive description of the user's multi-dimensional characteristics, including: user identity, song preference, historical on-demand behavior, consumption level, and social relationships.

[0032] The target song refers to the song that the user currently selects, plays on demand, or recommends in the system, such as the song currently selected by the user, the song that the user browses in the recommendation list, the song that the user searches in the search list, etc.

[0033] Song information includes: song number, song type, song popularity, song rating, copyright status, availability of high-definition version, singing difficulty, song language, and singer.

[0034] In order to increase the data processing speed, user portraits and song information can be cached so that they can be directly retrieved from the cache when needed, reducing the number of times the database is accessed.

[0035] S2. Match the relevant value-added service set according to the user portrait and song information, and use the value-added service set as the initial value-added service queue.

[0036] Value-added services refer to additional services provided for specific users or songs, such as free VIP song downloads, high-definition song playback, mobile phone song requests, song ratings, gift giving, coupons, song teaching videos, AI song accompaniment, online chorus, social interaction, etc.

[0037] The value-added service database contains all available value-added services; eligible services are matched from the value-added service database based on user portraits and song information. For example, if the user's user identity is VIP, advanced value-added services (unlocking high-definition resources, exclusive accompaniment services, etc.) can be matched, and songs that are difficult to sing can be matched with professional singing practice services.

[0038] All the relevant value-added services that have been screened are collected to form an initial queue. At this time, the order of the value-added services in the queue has not yet been prioritized.

[0039] S3. Calculate the priority and trigger probability of each value-added service based on the user portrait, song information and value-added services, and rearrange the initial value-added service queue in descending order of priority to obtain the current value-added service queue.

[0040] For each value-added service, its priority is calculated based on the matching of user portrait and song information, and the priority of each value-added service is sorted based on the relevance between the user and the song.

[0041] The trigger probability of each value-added service is calculated based on the priority, and a formula based on the priority and a preset trigger threshold is usually used to determine the probability. For example, if the priority of a value-added service is higher and the trigger probability is higher, then the service is more likely to be recommended.

[0042] The initial value-added service queue is sorted according to priority, such as sorting each value-added service in the initial value-added service queue in descending order of priority to obtain the current value-added service queue, ensuring priority recommendation of those value-added services that are highly relevant to user needs and song content.

[0043] S4. Display the value-added service according to the current value-added service queue, trigger probability and cooling time; wherein the cooling time is preset.

[0044] Each value-added service has a preset cooling time. The cooling time means that after a value-added service is displayed or triggered once, the system must wait for a certain period of time before recommending or triggering the same service again. This is used to avoid repeatedly recommending the same value-added service in a short period of time. The system will check whether the cooling time has passed. If the cooling time has not ended, the current value-added service will be skipped and the next value-added service will be selected for judgment.

[0045] When the cooling time is over, the system will decide whether to display the current value-added service based on the trigger probability. If the trigger probability is high, the value-added service is more likely to be triggered, otherwise it is not easy to be triggered. The triggered value-added service will be displayed to the user in various display forms, such as pop-up windows, labels, advertising spaces, QR codes, etc. The specific display form depends on the system design and user interface.

[0046] If the user activates the value-added service, the user will be provided with downloading, playing and other operations according to the content of the value-added service.

[0047] By setting a cooling-off period, we can avoid recommending the same value-added service to users frequently over a period of time, thereby improving user experience. By setting priorities, we can ensure that value-added services that meet user needs are recommended to users first, thereby improving the conversion rate of value-added services. By setting a trigger probability for each value-added service, we can ensure that each value-added service has the possibility of being recommended, thereby improving the diversity of recommendations.

[0048] As an optional implementation, Figure 2 is a flow chart of a method for calculating priority and trigger probability provided by an embodiment of the present invention. Figure 2 As shown, the priority and triggering probability of each value-added service are calculated based on the user portrait, song information and value-added services, including:

[0049] S301. Determine a preset priority of a value-added service.

[0050] Each value-added service will be pre-set with a preset priority, which can be set manually or determined based on historical data and user feedback. For example, for commonly used value-added services (such as VIP member recommendations) or newly launched value-added services, a higher priority can be set; while some infrequently used value-added services can be set with a lower preset priority.

[0051] When calculating the priority corresponding to the value-added service, it is necessary to select the preset priority of the value-added service for calculation.

[0052] S302: Calculate the user information score based on the user portrait.

[0053] The score of one or more features in the user portrait is calculated to obtain a user information score, that is, the user information may include one score or multiple scores.

[0054] The score can be calculated directly using the assignment method. For example, when calculating the user identity score, SVIP is assigned 80 points, VIP is assigned 50 points, and ordinary members are assigned 20 points.

[0055] Specific features can also be combined for calculation. For example, when calculating the score of a user's historical behavior, the score is calculated based on the user's click behavior, conversion times (the number of times the value-added service is successfully activated by the user) and the length of stay. As an optional implementation, the ratio of the number of clicks of the user to the total number of recommendations within a period of time is calculated as the click score; the ratio of the number of successful conversions of the user to the total number of recommendations within a period of time is calculated as the conversion score; the ratio of the user's stay time in a value-added service to the average stay time is calculated as the length of stay score. Finally, the click score, conversion score and length of stay score are weighted and summed to obtain the historical behavior score.

[0056] The user identity score and historical behavior score mentioned above are both user information scores.

[0057] S303: Calculate a song information score based on the song information.

[0058] Calculate the score of one or more features in the song information as the song information score. For example, the song type or song popularity can be assigned a value to calculate the score. For example, a pop music song is assigned a higher score, and a song with high popularity is assigned a higher score.

[0059] S304: Calculate a content relevance score based on the song request preference in the user portrait and the preference type in the song information.

[0060] The user's song preferences record the user's favorite singers, song types, preferred languages, etc. The preferred type in the song information is the type corresponding to the song preference. For example, when the singer is used as the calculation basis, the user preference and the preferred type are both singers. If the singer in the user's song preference matches the singer of the current song, a higher content relevance score is assigned. If they do not match, a lower content relevance score is assigned.

[0061] In this way, if a user prefers to click on a song of a certain singer, and the song happens to be sung by the singer, it will have a higher content relevance score, and the related value-added services will also have a higher priority.

[0062] S305: Set weight coefficients for the preset priority, user information score, song information score, and content relevance score respectively, and perform weighted summation to obtain the priority of the value-added service.

[0063] As an optional implementation, the priority of each value-added service is calculated by the following formula.

[0064] Priority = user identity weight * user identity score + song type weight * song type score + song popularity weight * song popularity score + user behavior weight * user behavior score + content relevance weight * content relevance score + preset priority.

[0065] Among them, user identity weight, song type weight, song popularity weight, user behavior weight and content relevance weight are all weight coefficients, and the weight coefficients of each item are pre-set.

[0066] S306: Calculate the trigger probability according to the priority of the value-added service and a preset trigger threshold.

[0067] As an optional implementation, the trigger probability is calculated according to the priority of the value-added service and the preset trigger threshold value by the following formula:

[0068]

[0069] The trigger threshold is a preset parameter used to control the trigger probability, P t is the trigger probability, P pri is the priority, T thr is the trigger threshold, and e is a constant.

[0070] The priority in the formula - the trigger threshold (P pri -T thr ) determines the probability weight of the current value-added service. If the priority is higher than the trigger threshold, the value-added service is more likely to be triggered (i.e., P t If the priority is lower than the trigger threshold, the trigger probability of the value-added service is low (i.e., P t smaller).

[0071] As an optional implementation, Figure 3 is a flow chart of a value-added service display method provided by an embodiment of the present invention, such as Figure 3 As shown, the value-added services are displayed according to the current value-added service queue, trigger probability and cooling time, including:

[0072] S401. Select a value-added service from the current value-added service queue in order of priority as the current value-added service.

[0073] Select value-added services from the current value-added service queue in sequence. Since the queue is sorted according to the priority of the value-added service, the value-added service with a higher priority is placed at the front of the queue. Therefore, the value-added service with a higher priority will be given priority when selecting.

[0074] S402: Calculate the time difference between the trigger time corresponding to the current value-added service and the current time.

[0075] Each value-added service has a cool-down period, which is to avoid the same value-added service being triggered to the user too frequently. If the cool-down period of the value-added service has not yet ended (that is, the value-added service has been triggered recently and the recommended time interval has not yet been reached), the value-added service cannot be displayed to the user immediately. At this time, the system will skip the value-added service and continue to select the next value-added service in the queue for cool-down period judgment.

[0076] When the value-added service is triggered, the trigger time is recorded. Then, when judging the cooling time, the judgment can be made based on the time difference between the trigger time and the current time.

[0077] S403: When the time difference is not less than the cooling time, the cooling time ends, and whether to trigger the current value-added service is determined based on the trigger probability and the trigger time is updated after the trigger; otherwise, the cooling time has not ended, and the next value-added service is selected for cooling time determination.

[0078] When the time difference is greater than or equal to the cooling time, the cooling time is considered to have ended, and the trigger probability of the current value-added service is used to determine whether to trigger the value-added service. After the value-added service is triggered, the value-added service is displayed and the trigger time is updated using the current time. If it is not triggered, the next value-added service is selected for cooling time determination. On the contrary, if the time difference is less than the cooling time, the cooling time is considered not to have ended, and the next value-added service is also selected for cooling time determination.

[0079] For example, if the cooling time is 5 minutes, the trigger time is 12:50, the current time is 13:00, and the time difference is 10 minutes, which exceeds the cooling time, then the cooling time has ended.

[0080] After the cooling-off period, the system will determine whether to display the value-added service based on the triggering probability of the value-added service.

[0081] As an optional implementation, a random number in the range of [0,1] can be randomly generated during probability judgment, and the value-added service can be determined by comparing the random number with the trigger probability. Specifically, if the random number is greater than the trigger probability, the trigger is not performed; otherwise, the trigger is performed. For example, if the trigger probability is 80% and the random number is 0.9, then 0.9>0.8, that is, the value-added service cannot be triggered.

[0082] As an optional implementation, if the value-added service is triggered, the system will display the value-added service and update the triggering time, that is, record that the value-added service has been triggered and start the next cooling cycle.

[0083] As an optional implementation, Figure 4 is a flow chart of an adaptive adjustment method provided by an embodiment of the present invention. Figure 4As shown, after displaying the value-added services according to the current value-added service queue, trigger probability and cooling time, it also includes:

[0084] S5. Calculate user feedback information, value-added service trigger rate and value-added service conversion rate.

[0085] User feedback information is the probability that the user turns off the value-added service;

[0086] The value-added service trigger rate is the ratio of the number of value-added service triggers to the number of times selected into the value-added service queue within a certain period of time;

[0087] The value-added service conversion rate is the ratio of the number of value-added service conversions to the number of triggers within a certain period of time.

[0088] The above values ​​can be calculated based on analysis of data over a certain period of time.

[0089] S6. Adjust the cooling time and / or preset priority of the corresponding value-added service according to the user feedback information.

[0090] If the probability of a user closing a value-added service is high, the cooling time of the corresponding value-added service can be increased, or its preset priority can be lowered; conversely, if the probability of a user closing a value-added service is low, the cooling time of the corresponding value-added service can be reduced, or its preset priority can be increased.

[0091] S7. Adjust the preset trigger threshold and / or weight coefficient according to the value-added service trigger rate.

[0092] If the trigger rate of a value-added service is too low, the weight coefficient of each item in the priority calculation process can be increased, or the preset trigger threshold of the corresponding value-added service can be lowered; conversely, if the trigger rate of a value-added service is too high, the corresponding weight can be lowered, or the preset trigger threshold of the corresponding value-added service can be increased.

[0093] S8. Adjust the content of value-added services according to the conversion rate of value-added services.

[0094] If the conversion rate of value-added services is too low, the content of the value-added services can be adjusted to make them more in line with user needs and with more reasonable conditions.

[0095] As an optional implementation, Figure 5 is a flow chart of a value-added service matching method provided by an embodiment of the present invention. Figure 5 As shown, the relevant value-added service set is matched according to the user portrait and song information, including:

[0096] S201. Combine features in the user portrait and features in the song information according to preset rules to obtain matching conditions.

[0097] Preset rules are a set of logical conditions used to combine user profile features and song information features. The rules define how user features and song features are associated, such as the combination of user identity and song clarity, the combination of user language preference and song language, and the combination of the number of times a user has requested a song and the difficulty of singing a song. User profile features and song information features can be a logical combination of multiple features, and multiple conditions can be connected using logical operators such as AND, OR, and NOT.

[0098] S202. Match the value-added services in the value-added service database according to the matching conditions. When the matching conditions meet the matching rules corresponding to the value-added services, add the value-added services to the value-added service set. The matching rules are pre-set and are a combination of the characteristics of the user portrait and the characteristics of the song information according to the preset rules.

[0099] The value-added service database stores all the value-added services that can be recommended and their corresponding matching rules. Each value-added service has a unique ID in the database, and the matching process is the process of traversing all the value-added services in the value-added service database. Each value-added service corresponds to a preset matching rule, which is also a feature combination based on the preset rule. For example, when the value-added service is high-definition accompaniment download, its corresponding matching rule is that the user identity is VIP and the song clarity is high-definition. If the matching condition meets the matching rule during the matching process, the high-definition accompaniment download value-added service is added to the value-added service set.

[0100] The embodiment of the present invention also provides a value-added service recommendation system. Figure 6 is a schematic diagram of the structure of the value-added service recommendation system provided by an embodiment of the present invention. Figure 6 As shown, the system includes:

[0101] The initial data acquisition module 100 is used to obtain the user portrait of the user and the song information of the target song.

[0102] User portrait is a comprehensive description of the user's multi-dimensional characteristics, including: user identity, song preference, historical on-demand behavior, consumption level, and social relationships.

[0103] The target song refers to the song that the user currently selects, plays on demand, or recommends in the system, such as the song currently selected by the user, the song that the user browses in the recommendation list, the song that the user searches in the search list, etc.

[0104] Song information includes: song number, song type, song popularity, song rating, copyright status, availability of high-definition version, singing difficulty, song language, and singer.

[0105] In order to increase the data processing speed, user portraits and song information can be cached so that they can be directly retrieved from the cache when needed, reducing the number of times the database is accessed.

[0106] The value-added service matching module 200 is used to match the relevant value-added service set according to the user portrait and song information, and use the value-added service set as the initial value-added service queue.

[0107] Value-added services refer to additional services provided for specific users or songs, such as free VIP song downloads, high-definition song playback, mobile phone song requests, song ratings, gift giving, coupons, song teaching videos, AI song accompaniment, online chorus, social interaction, etc.

[0108] The value-added service database contains all available value-added services; eligible services are matched from the value-added service database based on user portraits and song information. For example, if the user's user identity is VIP, advanced value-added services (unlocking high-definition resources, exclusive accompaniment services, etc.) can be matched, and songs that are difficult to sing can be matched with professional singing practice services.

[0109] All the relevant value-added services that have been screened are collected to form an initial queue. At this time, the order of the value-added services in the queue has not yet been prioritized.

[0110] The sorting module 300 is used to calculate the priority and trigger probability of each value-added service according to the user portrait, song information and value-added services, and rearrange the initial value-added service queue in descending order of priority to obtain the current value-added service queue.

[0111] For each value-added service, its priority is calculated based on the matching of user portrait and song information, and the priority of each value-added service is sorted based on the relevance between the user and the song.

[0112] The trigger probability of each value-added service is calculated based on the priority, and a formula based on the priority and a preset trigger threshold is usually used to determine the probability. For example, if the priority of a value-added service is higher and the trigger probability is higher, then the service is more likely to be recommended.

[0113] The initial value-added service queue is sorted according to priority, such as sorting each value-added service in the initial value-added service queue in descending order of priority to obtain the current value-added service queue, ensuring priority recommendation of those value-added services that are highly relevant to user needs and song content.

[0114] The display module 400 is used to display the value-added services according to the current value-added service queue, trigger probability and cooling time; wherein the cooling time is preset.

[0115] Each value-added service has a preset cooling time. The cooling time means that after a value-added service is displayed or triggered once, the system must wait for a certain period of time before recommending or triggering the same service again. This is used to avoid repeatedly recommending the same value-added service in a short period of time. The system will check whether the cooling time has passed. If the cooling time has not ended, the current value-added service will be skipped and the next value-added service will be selected for judgment.

[0116] When the cooling time is over, the system will decide whether to display the current value-added service based on the trigger probability. If the trigger probability is high, the value-added service is more likely to be triggered, otherwise it is not easy to be triggered. The triggered value-added service will be displayed to the user in various display forms, such as pop-up windows, labels, advertising spaces, QR codes, etc. The specific display form depends on the system design and user interface.

[0117] If the user activates the value-added service, the user will be provided with downloading, playing and other operations according to the content of the value-added service.

[0118] By setting a cooling-off period, we can avoid recommending the same value-added service to users frequently over a period of time, thereby improving user experience. By setting priorities, we can ensure that value-added services that meet user needs are recommended to users first, thereby improving the conversion rate of value-added services. By setting a trigger probability for each value-added service, we can ensure that each value-added service has the possibility of being recommended, thereby improving the diversity of recommendations.

[0119] As an optional implementation, Figure 7 is a schematic diagram of the structure of the sorting module provided in an embodiment of the present invention, such as Figure 7 As shown, the sorting module 300 includes:

[0120] The preset priority determination submodule 3001 is used to determine the preset priority of the value-added service.

[0121] Each value-added service will be pre-set with a preset priority, which can be set manually or determined based on historical data and user feedback. For example, for commonly used value-added services (such as VIP member recommendations) or newly launched value-added services, a higher priority can be set; while some infrequently used value-added services can be set with a lower preset priority.

[0122] When calculating the priority corresponding to the value-added service, it is necessary to select the preset priority of the value-added service for calculation.

[0123] The user information score calculation submodule 3002 is used to calculate the user information score according to the user portrait.

[0124] The score of one or more features in the user portrait is calculated to obtain a user information score, that is, the user information may include one score or multiple scores.

[0125] The score can be calculated directly using the assignment method. For example, when calculating the user identity score, SVIP is assigned 80 points, VIP is assigned 50 points, and ordinary members are assigned 20 points.

[0126] Specific features can also be combined for calculation. For example, when calculating the score of a user's historical behavior, the score is calculated based on the user's click behavior, conversion times (the number of times the value-added service is successfully activated by the user) and the length of stay. As an optional implementation, the ratio of the number of clicks of the user to the total number of recommendations within a period of time is calculated as the click score; the ratio of the number of successful conversions of the user to the total number of recommendations within a period of time is calculated as the conversion score; the ratio of the user's stay time in a value-added service to the average stay time is calculated as the length of stay score. Finally, the click score, conversion score and length of stay score are weighted and summed to obtain the historical behavior score.

[0127] The user identity score and historical behavior score mentioned above are both user information scores.

[0128] The song information score calculation submodule 3003 is used to calculate the song information score according to the song information.

[0129] Calculate the score of one or more features in the song information as the song information score. For example, the song type or song popularity can be assigned a value to calculate the score. For example, a pop music song is assigned a higher score, and a song with high popularity is assigned a higher score.

[0130] The relevance score calculation submodule 3004 is used to calculate the content relevance score according to the song request preference in the user portrait and the preference type in the song information.

[0131] The user's song preferences record the user's favorite singers, song types, preferred languages, etc. The preferred type in the song information is the type corresponding to the song preference. For example, when the singer is used as the calculation basis, the user preference and the preferred type are both singers. If the singer in the user's song preference matches the singer of the current song, a higher content relevance score is assigned. If they do not match, a lower content relevance score is assigned.

[0132] In this way, if a user prefers to click on a song of a certain singer, and the song happens to be sung by the singer, it will have a higher content relevance score, and the related value-added services will also have a higher priority.

[0133] The weighted sum submodule 3005 is used to set weight coefficients for the preset priority, user information score, song information score and content relevance score respectively, and then perform weighted summation to obtain the priority of the value-added service.

[0134] As an optional implementation, the priority of each value-added service is calculated by the following formula:

[0135] Priority = user identity weight * user identity score + song type weight * song type score + song popularity weight * song popularity score + user behavior weight * user behavior score + content relevance weight * content relevance score + preset priority.

[0136] Among them, user identity weight, song type weight, song popularity weight, user behavior weight and content relevance weight are all weight coefficients, and the weight coefficients of each item are pre-set.

[0137] The trigger probability calculation submodule 3006 is used to calculate the trigger probability according to the priority of the value-added service and the preset trigger threshold.

[0138] As an optional implementation, the trigger probability is calculated according to the priority of the value-added service and the preset trigger threshold value by the following formula:

[0139]

[0140] The trigger threshold is a preset parameter used to control the trigger probability, P t is the trigger probability, P pri is the priority, T thr is the trigger threshold, and e is a constant.

[0141] The priority in the formula - the trigger threshold (P pri -T thr ) determines the probability weight of the current value-added service. If the priority is higher than the trigger threshold, the value-added service is more likely to be triggered (i.e., P t If the priority is lower than the trigger threshold, the trigger probability of the value-added service is low (i.e., P t smaller).

[0142] As an optional implementation, Figure 8 is a schematic diagram of the structure of a display module provided by an embodiment of the present invention, such as Figure 8 As shown, the display module 400 includes:

[0143] The value-added service selection submodule 4001 is used to select a value-added service from the current value-added service queue in order of priority as the current value-added service.

[0144] Select value-added services from the current value-added service queue in sequence. Since the queue is sorted according to the priority of the value-added service, the value-added service with a higher priority is placed at the front of the queue. Therefore, the value-added service with a higher priority will be given priority when selecting.

[0145] The calculation submodule 4002 is used to calculate the time difference between the trigger time corresponding to the current value-added service and the current time.

[0146] Each value-added service has a cool-down period, which is to avoid the same value-added service being triggered to the user too frequently. If the cool-down period of the value-added service has not yet ended (that is, the value-added service has been triggered recently and the recommended time interval has not yet been reached), the value-added service cannot be displayed to the user immediately. At this time, the system will skip the value-added service and continue to select the next value-added service in the queue for cool-down period judgment.

[0147] When the value-added service is triggered, the trigger time is recorded. When judging the cooling time, the time difference between the trigger time and the current time can be used for judgment. For example, if the cooling time is 5 minutes, the trigger time is 12:50, and the current time is 13:00, the time difference is 10 minutes. If the cooling time is exceeded, the cooling time has ended.

[0148] The second judgment submodule 4003 is used to determine whether to trigger the current value-added service and update the trigger time after the trigger if the time difference is not less than the cooling time and the cooling time ends; otherwise, the cooling time has not ended and the next value-added service is selected for cooling time judgment.

[0149] When the time difference is greater than or equal to the cooling time, the cooling time is considered to have ended, and the trigger probability of the current value-added service is used to determine whether to trigger the value-added service. After the value-added service is triggered, the value-added service is displayed and the trigger time is updated using the current time. If it is not triggered, the next value-added service is selected for cooling time determination. On the contrary, if the time difference is less than the cooling time, the cooling time is considered not to have ended, and the next value-added service is also selected for cooling time determination.

[0150] For example, if the cooling time is 5 minutes, the trigger time is 12:50, the current time is 13:00, and the time difference is 10 minutes, which exceeds the cooling time, then the cooling time has ended.

[0151] After the cooling-off period, the system will determine whether to display the value-added service based on the triggering probability of the value-added service.

[0152] As an optional implementation, a random number in the range of [0,1] can be randomly generated during probability judgment, and the value-added service can be determined by comparing the random number with the trigger probability. Specifically, if the random number is greater than the trigger probability, the trigger is not performed; otherwise, the trigger is performed. For example, if the trigger probability is 80% and the random number is 0.9, then 0.9>0.8, that is, the value-added service cannot be triggered.

[0153] As an optional implementation, if the value-added service is triggered, the system will display the value-added service and update the triggering time, that is, record that the value-added service has been triggered and start the next cooling cycle.

[0154] As an optional implementation, Fig. 9 is a schematic diagram of the structure of a module for adaptive adjustment provided by an embodiment of the present invention, such as Fig. 9 As shown, the system also includes:

[0155] The effect data calculation module 500 is used to calculate user feedback information, value-added service trigger rate and value-added service conversion rate.

[0156] User feedback information is the probability that the user turns off the value-added service;

[0157] The value-added service trigger rate is the ratio of the number of value-added service triggers to the number of times selected into the value-added service queue within a certain period of time;

[0158] The value-added service conversion rate is the ratio of the number of value-added service conversions to the number of triggers within a certain period of time.

[0159] The above values ​​can be calculated based on analysis of data over a certain period of time.

[0160] The first adjustment module 600 is used to adjust the cooling time and / or preset priority of the corresponding value-added service according to user feedback information.

[0161] If the probability of a user closing a value-added service is high, the cooling time of the corresponding value-added service can be increased, or its preset priority can be lowered; conversely, if the probability of a user closing a value-added service is low, the cooling time of the corresponding value-added service can be reduced, or its preset priority can be increased.

[0162] The second adjustment module 700 is used to adjust the preset trigger threshold and / or weight coefficient according to the value-added service trigger rate.

[0163] If the trigger rate of a value-added service is too low, the weight coefficient of each item in the priority calculation process can be increased, or the preset trigger threshold of the corresponding value-added service can be lowered; conversely, if the trigger rate of a value-added service is too high, the corresponding weight can be lowered, or the preset trigger threshold of the corresponding value-added service can be increased.

[0164] The third adjustment module 800 is used to adjust the content of the value-added service according to the value-added service conversion rate.

[0165] If the conversion rate of value-added services is too low, the content of the value-added services can be adjusted to make them more in line with user needs and with more reasonable conditions.

[0166] As an optional implementation, Fig.10is a schematic diagram of the structure of the value-added service matching module provided in an embodiment of the present invention. Fig.10 As shown, the value-added service matching module 200 includes:

[0167] The feature combination submodule 2001 is used to combine the features in the user portrait with the features in the song information according to preset rules to obtain matching conditions.

[0168] Preset rules are a set of logical conditions used to combine user profile features and song information features. The rules define how user features and song features are associated, such as the combination of user identity and song clarity, the combination of user language preference and song language, and the combination of the number of times a user has requested a song and the difficulty of singing a song. User profile features and song information features can be a logical combination of multiple features, and multiple conditions can be connected using logical operators such as AND, OR, and NOT.

[0169] The matching submodule 2002 is used to match the value-added services in the value-added service database according to the matching conditions. When the matching conditions meet the matching rules corresponding to the value-added services, the value-added services are added to the value-added service set; wherein the matching rules are pre-set and are a combination of the characteristics of the user portrait and the characteristics of the song information according to the preset rules.

[0170] The value-added service database stores all the value-added services that can be recommended and their corresponding matching rules. Each value-added service has a unique ID in the database, and the matching process is the process of traversing all the value-added services in the value-added service database. Each value-added service corresponds to a preset matching rule, which is also a feature combination based on the preset rule. For example, when the value-added service is high-definition accompaniment download, its corresponding matching rule is that the user identity is VIP and the song clarity is high-definition. If the matching condition meets the matching rule during the matching process, the high-definition accompaniment download value-added service is added to the value-added service set.

[0171] The above technical solution has the following beneficial effects: through comprehensive analysis of user portraits and song information, the recommended song value-added service types are more in line with user needs; by setting a cooling-off time, it is ensured that the same value-added service will not be recommended to users frequently, thereby ensuring the diversification of value-added service recommendations and improving user experience; by adaptively adjusting various parameters, the accuracy of value-added service recommendations is ensured, thereby improving user experience.

[0172] The specific implementation methods of the above invention further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above content is only the specific implementation methods of the present invention and is not intended to limit the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for recommending value-added services, characterized in that: include: Obtain the user profile of the user and the song information of the target song; Matching a relevant value-added service set according to the user portrait and song information, and using the value-added service set as an initial value-added service queue; Calculate the priority and trigger probability of each value-added service according to the user portrait, song information and value-added services, and rearrange the initial value-added service queue in descending order of priority to obtain a current value-added service queue; The value-added service is displayed according to the current value-added service queue, trigger probability and cooling time; wherein the cooling time is preset.

2. The value-added service recommendation method according to claim 1, characterized in that: The calculating the priority and triggering probability of each value-added service according to the user portrait, song information and value-added services includes: Determine the preset priorities of value-added services; Calculate user information score based on user portrait; Calculate a song information score according to the song information; Calculate the content relevance score based on the song request preference in the user portrait and the preference type in the song information; The preset priority, user information score, song information score and content relevance score are respectively set with weight coefficients and then weighted summed to obtain the priority of the value-added service; The trigger probability is calculated based on the priority of the value-added service and the preset trigger threshold.

3. The value-added service recommendation method according to claim 2, characterized in that: The displaying of the value-added service according to the current value-added service queue, trigger probability and cooling time includes: Selecting a value-added service from the current value-added service queue in order of priority as the current value-added service; Calculate the time difference between the trigger time corresponding to the current value-added service and the current time; When the time difference is not less than the cooling time, the cooling time ends, and whether the current value-added service is triggered is determined based on the trigger probability. If triggered, the current value-added service is displayed and the trigger time is updated. If not triggered, the next value-added service is selected for cooling time judgment; otherwise, the cooling time has not ended, and the next value-added service is selected for cooling time judgment.

4. The value-added service recommendation method according to claim 3, characterized in that: After displaying the value-added service according to the current value-added service queue, the trigger probability and the cooling time, the method further includes: Calculate user feedback information, value-added service trigger rate and value-added service conversion rate; Adjusting the cooling time and / or preset priority of the corresponding value-added service according to the user feedback information; Adjusting the preset trigger threshold and / or the weight coefficient according to the value-added service trigger rate; The content of the value-added service is adjusted according to the value-added service conversion rate.

5. The value-added service recommendation method according to claim 4, characterized in that: The value-added service set related to matching the user portrait and song information includes: Combining the features in the user portrait with the features in the song information according to a preset rule to obtain a matching condition; The value-added services are matched in the value-added service database according to the matching conditions, and when the matching conditions meet the matching rules corresponding to any value-added service, the any value-added service is added to the value-added service set; wherein the matching rules are pre-set and are a combination of the characteristics of the user portrait and the characteristics of the song information according to preset rules.

6. A value-added service recommendation system, characterized in that: include: An initial data acquisition module is used to obtain the user portrait of the user and the song information of the target song; A value-added service matching module, used for matching a relevant value-added service set according to the user portrait and song information, and using the value-added service set as an initial value-added service queue; A sorting module, used to calculate the priority and triggering probability of each value-added service according to the user portrait, song information and value-added services, and rearrange the initial value-added service queue in descending order of priority to obtain a current value-added service queue; The display module is used to display the value-added service according to the current value-added service queue, trigger probability and cooling time; wherein the cooling time is preset.

7. The value-added service recommendation system according to claim 6, characterized in that: The sorting module comprises: A preset priority determination submodule, used to determine the preset priority of the value-added service; The user information score calculation submodule is used to calculate the user information score based on the user portrait; A song information score calculation submodule, used to calculate a song information score based on the song information; A relevance score calculation submodule is used to calculate the content relevance score based on the song request preference in the user portrait and the preference type in the song information; A weighted summation submodule, used to set weight coefficients for the preset priority, user information score, song information score and content relevance score, and then perform weighted summation to obtain the priority of the value-added service; The trigger probability calculation submodule is used to calculate the trigger probability according to the priority of the value-added service and the preset trigger threshold.

8. The value-added service recommendation system according to claim 7, characterized in that: The display module comprises: A value-added service selection submodule, configured to select a value-added service from the current value-added service queue in order of priority as the current value-added service; The time difference calculation submodule is used to calculate the time difference between the trigger time corresponding to the current value-added service and the current time; The judgment submodule is used to judge whether the current value-added service is triggered according to the trigger probability when the time difference is not less than the cooling time and the cooling time ends. If triggered, the current value-added service is displayed and the trigger time is updated. If not triggered, the next value-added service is selected for cooling time judgment; otherwise, the cooling time has not ended and the next value-added service is selected for cooling time judgment.

9. The value-added service recommendation system according to claim 8, characterized in that: Also includes: Effect data calculation module, used to calculate user feedback information, value-added service trigger rate and value-added service conversion rate; A first adjustment module, configured to adjust a cooling time and / or a preset priority of a corresponding value-added service according to the user feedback information; A second adjustment module, configured to adjust the preset trigger threshold and / or the weight coefficient according to the value-added service trigger rate; The third adjustment module is used to adjust the content of the value-added service according to the value-added service conversion rate.

10. The value-added service recommendation system according to claim 9, characterized in that: The value-added service matching module includes: A feature combination submodule, used to combine the features in the user portrait with the features in the song information according to a preset rule to obtain a matching condition; The matching submodule is used to match the value-added services in the value-added service database according to the matching conditions. When the matching conditions meet the matching rules corresponding to any value-added service, the any value-added service is added to the value-added service set; wherein the matching rules are pre-set and are a combination of the characteristics of the user portrait and the characteristics of the song information according to preset rules.

Images