Live broadcast room recommendation method and device, storage medium and program product

Abstract

The invention provides a live broadcast room recommendation method and device, a storage medium and a program product, and belongs to the technical field of computers. The method comprises the following steps: receiving a live broadcast room recommendation request sent by a target account; determining a matching degree between each live broadcast room and the target account; determining at least one target new anchor live broadcast room in the plurality of live broadcast rooms; based on the to-be-exposed quantity of each target new anchor live broadcast room, adjusting a matching degree between each target new anchor live broadcast room and a target account, and obtaining an updated matching degree between each target new anchor live broadcast room and the target account; and determining a live broadcast room recommendation queue based on the updated matching degree between the target new anchor live broadcast room and the target account and the matching degree between other live broadcast rooms except the target new anchor live broadcast room and the target account, and sending the live broadcast room recommendation queue to the target account for recommendation display. According to the invention, the exposure of a new anchor can be improved.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of computer, in particular, to a live room recommendation method and device, a storage medium and a program product. BACKGROUND

[0002] As a popular leisure and entertainment way, live streaming is gradually entering people's lives. In various application programs such as music, video and shopping, live streaming can be used to provide better application experience for users. In addition, a live room recommendation page is displayed in the application program to recommend multiple live rooms that the user may be interested in.

[0003] Currently, the determination method of multiple live rooms displayed for recommendation in the live room recommendation page is as follows: based on the historical live streaming related data of the account to which the user belongs and the historical live streaming related data of each live room, the matching degree between the account and each live room is calculated, and then based on the matching degree, a live room recommendation queue is determined, that is, multiple live rooms recommended for the user are obtained.

[0004] However, since many new anchors have a short broadcasting time and do not have enough historical live streaming related data, the matching degree between the new anchor and the account calculated is not accurate, thereby reducing the exposure rate of the new anchor. SUMMARY

[0005] The present disclosure provides a live room recommendation method, which can solve the technical problems existing in the related art. The technical solution is as follows:

[0006] In a first aspect, a live room recommendation method is provided, and the method comprises:

[0007] receiving a live room recommendation request sent by a target account;

[0008] determining the matching degree between each live room and the target account;

[0009] determining at least one target new anchor live room in multiple live rooms;

[0010] adjusting the matching degree between each target new anchor live room and the target account based on the to-be-exposed amount of each target new anchor live room, to obtain the updated matching degree between each target new anchor live room and the target account;

[0011] determining a live room recommendation queue based on the updated matching degree between the target new anchor live room and the target account and the matching degree between other live rooms except the target new anchor live room and the target account, and sending the live room recommendation queue to the target account for recommendation display.

[0012] In a possible implementation, the determining of the at least one target new anchor live room from the multiple live rooms comprises:

[0013] determining a high-quality new anchor from the multiple new anchors;

[0014] determining a high-quality new anchor live room corresponding to the at least one high-quality new anchor in a live state;

[0015] determining the at least one target new anchor live room from the high-quality new anchor live rooms based on the exposure-to-be-exposed amount and the performance score of each high-quality new anchor live room.

[0016] In a possible implementation, the determining of the high-quality new anchor from the multiple new anchors comprises:

[0017] determining a recommendation score of each new anchor based on historical live data of the new anchor, wherein the historical live data comprises at least one of a click rate, a number of live broadcast days, a daily accumulated live broadcast time, a user effective stay rate, and a fan attention conversion rate;

[0018] if the recommendation score of the new anchor is greater than or equal to a preset score threshold, or if the artificial review of the new anchor is passed, the new anchor is determined as the high-quality new anchor.

[0019] In a possible implementation, after the new anchor is determined as the high-quality new anchor, the method further comprises:

[0020] calculating a mean and a variance of first live data of the multiple high-quality new anchors, and calculating a difference value between the mean and the variance, wherein the first live data comprises at least one of a click rate and a user effective stay rate;

[0021] if the first live data of the high-quality new anchor is less than the difference value in a first latest preset time period, the exposure-to-be-exposed amount of the high-quality new anchor is adjusted downward, and the exposure-to-be-exposed amount of the high-quality new anchor is updated.

[0022] In a possible implementation, after the current exposure-to-be-exposed amount of the high-quality new anchor is adjusted downward to obtain the exposure-to-be-exposed amount of the high-quality new anchor, the method further comprises:

[0023] performing sorting from large to small based on the first live data of each high-quality new anchor to obtain a high-quality ranking;

[0024] for the high-quality new anchor whose exposure-to-be-exposed amount is adjusted downward, a week-on-week decrease value of the first live data of the high-quality new anchor is determined, and high-quality new anchors whose week-on-week decrease values exceed a preset change threshold are returned to the new anchor from the last first preset proportion of high-quality new anchors in the high-quality ranking.

[0025] In a possible implementation, the determining the at least one target new anchor live room based on the to-be-exposed amount and the performance score of each high-quality new anchor live room comprises:

[0026] determining a recommendation probability corresponding to each high-quality new anchor live room based on the to-be-exposed amount of each high-quality new anchor live room, wherein the recommendation probability is in a positive correlation with the to-be-exposed amount;

[0027] performing a descending sorting based on the performance score of each high-quality new anchor live room to obtain a score ranking;

[0028] multiplying the recommendation probability of a second preset proportion of high-quality new anchor live rooms at a front position in the score ranking by a first coefficient, multiplying the recommendation probability of a third preset proportion of high-quality new anchor live rooms at a rear position in the score ranking by a second coefficient, and updating the recommendation probability of the high-quality new anchor live room, wherein the first coefficient is greater than 1, and the second coefficient is greater than 0 and less than 1;

[0029] performing probability sampling on a plurality of high-quality new anchor live rooms based on the recommendation probability of each high-quality new anchor live room to obtain the at least one target new anchor live room.

[0030] In a possible implementation, the method further comprises:

[0031] for each high-quality new anchor, determining other anchors matched with the high-quality new anchor based on anchor portrait data of the high-quality new anchor and anchor portrait data of other anchors except the new anchor;

[0032] determining a to-be-exposed amount of a high-quality new anchor live room corresponding to the high-quality new anchor based on a historical exposed amount of a live room corresponding to the other anchor matched with the high-quality new anchor and a current exposed amount of the high-quality new anchor live room corresponding to the high-quality new anchor.

[0033] In a possible implementation, the method further comprises:

[0034] determining a performance score of each high-quality new anchor live room based on second live data of each high-quality new anchor live room in a second latest preset time period, wherein the second live data comprises at least one of a click rate, an average stay time after entering a room, and a second-entrance-then-quit rate.

[0035] In a possible implementation, the matching degree between each target new anchor live room and the target account is adjusted based on the to-be-exposed amount of each target new anchor live room, respectively, to obtain an updated matching degree between each target new anchor live room and the target account, including:

[0036] The matching degree between each target new anchor live room and the target account is adjusted based on the to-be-exposed amount and the current on-air duration of each target new anchor live room, respectively, to obtain an updated matching degree between each target new anchor live room and the target account.

[0037] In a possible implementation, the matching degree between each target new anchor live room and the target account is adjusted based on the to-be-exposed amount of each target new anchor live room and the current on-air duration of the target new anchor live room, respectively, to obtain an updated matching degree between each target new anchor live room and the target account, including:

[0038] For each target new anchor live room, the to-be-exposed amount, the time derivative of the to-be-exposed amount, and the current on-air duration of the target new anchor live room are obtained.

[0039] The proportional gain, the derivative gain, and the time gain are determined, where the proportional gain is in a positive correlation with the to-be-exposed amount, the derivative gain is in a negative correlation with the time derivative of the to-be-exposed amount, and the time gain is in a positive correlation with the current on-air duration.

[0040] The to-be-exposed amount, the time derivative of the to-be-exposed amount, and the current on-air duration are weighted based on the proportional gain, the derivative gain, and the time gain, to obtain a weight value corresponding to the target new anchor live room.

[0041] The weight value is multiplied by the matching degree between the target new anchor live room and the target account, to obtain an updated matching degree between the target new anchor live room and the target account.

[0042] In a possible implementation, the matching degree between each live room and the target account is determined, including:

[0043] The user behavior sequence feature and the user portrait feature corresponding to the target account are obtained.

[0044] For each live room, the anchor offline real-time index feature and the anchor portrait feature corresponding to the live room, the user-anchor cross feature and the context information feature between the live room and the target account are obtained.

[0045] determine a matching degree between the live room price and the target account based on the user behavior sequence features and the user portrait features corresponding to the target account, the anchor offline real-time index features and the anchor portrait features corresponding to the live room, the user-anchor cross features and the context information features between the live room and the target account, and a matching degree prediction model.

[0046] In a second aspect, a live room recommendation device is provided, and the device includes:

[0047] a receiving module configured to receive a live room recommendation request sent by a target account;

[0048] a matching degree determination module configured to determine a matching degree between each live room and the target account;

[0049] a target determination module configured to determine at least one target new anchor live room from a plurality of live rooms;

[0050] an updating module configured to adjust the matching degree between each target new anchor live room and the target account based on a to-be-exposed amount of the target new anchor live room, to obtain an updated matching degree between the target new anchor live room and the target account;

[0051] a queue determination module configured to determine a live room recommendation queue based on the updated matching degree between the target new anchor live room and the target account and the matching degrees between other live rooms except the target new anchor live room and the target account, and send the live room recommendation queue to the target account for recommendation display.

[0052] In a possible implementation, the target determination module is configured to:

[0053] determine a high-quality new anchor from a plurality of new anchors;

[0054] determine a high-quality new anchor live room corresponding to at least one high-quality new anchor in a live state;

[0055] determine the at least one target new anchor live room from the high-quality new anchor live rooms based on a to-be-exposed amount and a performance score of each high-quality new anchor live room.

[0056] In a possible implementation, the target determination module is configured to:

[0057] determine a recommendation score of each new anchor based on historical live data of the new anchor, wherein the historical live data includes at least one of a click rate, a number of live broadcast days, a daily cumulative live broadcast time, a user effective stay rate, and a fan attention conversion rate;

[0058] If the recommendation score of the new anchor is greater than or equal to a preset score threshold, or the artificial review determination of the new anchor is passed, the new anchor is determined as the high-quality new anchor.

[0059] In a possible implementation, the target determination module is further configured to:

[0060] calculate a mean and a variance of first live broadcast data of the plurality of high-quality new anchors, and calculate a difference value between the mean and the variance, wherein the first live broadcast data includes at least one of a click rate and a user effective stay rate;

[0061] If the first live broadcast data of the high-quality new anchor is less than the difference value in a first latest preset time period, the exposure-to-be-exposed amount of the high-quality new anchor is adjusted downward, and the exposure-to-be-exposed amount of the high-quality new anchor is updated.

[0062] In a possible implementation, the target determination module is further configured to:

[0063] sort the high-quality new anchors in a descending order based on the first live broadcast data of each high-quality new anchor, to obtain a high-quality ranking;

[0064] For the high-quality new anchor whose exposure-to-be-exposed amount is adjusted downward, a week-on-week decrease value of the first live broadcast data of the high-quality new anchor is determined, and high-quality new anchors whose week-on-week decrease values exceed a preset change threshold are returned to the new anchor from the last first preset proportion of high-quality new anchors in the high-quality ranking.

[0065] In a possible implementation, the target determination module is configured to:

[0066] determine a recommendation probability corresponding to each high-quality new anchor live broadcast room based on the exposure-to-be-exposed amount of the high-quality new anchor live broadcast room, wherein the recommendation probability and the exposure-to-be-exposed amount are in a positive correlation;

[0067] sort the high-quality new anchor live broadcast rooms in a descending order based on the performance scores of the high-quality new anchor live broadcast rooms, to obtain a score ranking;

[0068] multiply the recommendation probability of the high-quality new anchor live broadcast room in the last first preset proportion in the score ranking by a first coefficient, multiply the recommendation probability of the high-quality new anchor live broadcast room in the last third preset proportion in the score ranking by a second coefficient, and update the recommendation probability of the high-quality new anchor live broadcast room, wherein the first coefficient is greater than 1, and the second coefficient is greater than 0 and less than 1;

[0069] perform probability sampling on the plurality of high-quality new anchor live broadcast rooms based on the recommendation probability of each high-quality new anchor live broadcast room, to obtain the at least one target new anchor live broadcast room.

[0070] In a possible implementation, the target determining module is further configured to:

[0071] For each high-quality new anchor, based on anchor portrait data of the high-quality new anchor and anchor portrait data of other anchors except the new anchor, determine other anchors matched with the high-quality new anchor;

[0072] Based on historical exposure amounts of live broadcast rooms corresponding to the other anchors matched with the high-quality new anchor and a current exposure amount of a high-quality new anchor live broadcast room corresponding to the high-quality new anchor, determine a to-be-exposed amount of the high-quality new anchor live broadcast room corresponding to the high-quality new anchor.

[0073] In a possible implementation, the target determining module is further configured to:

[0074] Based on second live broadcast data of each high-quality new anchor live broadcast room in a second latest preset time period, determine a performance score of each high-quality new anchor live broadcast room, wherein the second live broadcast data includes at least one of a click rate, a per-person stay time after entering a room, and a second-entrance-then-leave rate.

[0075] In a possible implementation, the updating module is configured to:

[0076] Based on the to-be-exposed amount and the current on-air length of each target new anchor live broadcast room, respectively adjust a matching degree between each target new anchor live broadcast room and the target account to obtain an updated matching degree between each target new anchor live broadcast room and the target account.

[0077] In a possible implementation, the updating module is configured to:

[0078] For each target new anchor live broadcast room, obtain a to-be-exposed amount of the target new anchor live broadcast room, a time derivative of the to-be-exposed amount, and a current on-air length;

[0079] Determine a proportional gain, a derivative gain, and a time gain, wherein the proportional gain is in a positive correlation with the to-be-exposed amount, the derivative gain is in a negative correlation with the time derivative of the to-be-exposed amount, and the time gain is in a positive correlation with the current on-air length;

[0080] Based on the proportional gain, the derivative gain, and the time gain, perform weighted processing on the to-be-exposed amount, the time derivative of the to-be-exposed amount, and the current on-air length to obtain a weight value corresponding to the target new anchor live broadcast room;

[0081] The weight value is multiplied by the matching degree between the target new anchor live room and the target account, to obtain an updated matching degree between the target new anchor live room and the target account.

[0082] In a possible implementation, the matching degree determination module is configured to:

[0083] obtain user behavior sequence features and user portrait features corresponding to the target account;

[0084] For each live room, obtain anchor offline real-time indicator features and anchor portrait features corresponding to the live room, user-anchor cross features and context information features between the live room and the target account;

[0085] determine the matching degree between the live room and the target account based on the user behavior sequence features and user portrait features corresponding to the target account, the anchor offline real-time indicator features and anchor portrait features corresponding to the live room, the user-anchor cross features and context information features between the live room and the target account, and a matching degree prediction model.

[0086] In a third aspect, a computer device is provided, which includes a processor and a memory. The memory stores at least one instruction, which is loaded and executed by the processor to implement operations performed by the live room recommendation method.

[0087] In a fourth aspect, a computer readable storage medium is provided, which stores at least one instruction. The instruction is loaded and executed by a processor to implement operations performed by the live room recommendation method.

[0088] In a fifth aspect, a computer program product is provided, which includes at least one instruction. The at least one instruction is loaded and executed by a processor to implement operations performed by the live room recommendation method.

[0089] The technical solution provided by the present disclosure has the beneficial effects that: the solution mentioned in the present disclosure can adjust the matching degree between the target new anchor live room and the target account based on the to-be-exposed quantity of each target new anchor live room, so that the updated matching degree can meet the exposure demand of the target new anchor live room while meeting the effective matching between the target account and the target new anchor live room. Thus, the target new anchor live room with high exposure demand that is interesting to the target account can be pushed to the target account based on the updated matching degree, so as to improve the exposure quantity of new anchors. BRIEF DESCRIPTION OF DRAWINGS

[0090] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings needed to be used in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present disclosure, and for those skilled in the art, other drawings can be obtained based on these drawings without creative labor.

[0091] Figure 1 is a flow diagram of a live room recommendation method provided by an embodiment of the present disclosure;

[0092] Figure 2 is a flow diagram of a live room recommendation method provided by an embodiment of the present disclosure;

[0093] Figure 3 is a schematic diagram of a live recommendation page provided by an embodiment of the present disclosure;

[0094] Figure 4 is a flow diagram of a method for determining a target new anchor live room provided by an embodiment of the present disclosure;

[0095] Figure 5 is a structural schematic diagram of a live room recommendation device provided by an embodiment of the present disclosure;

[0096] Figure 6 is a structural block diagram of a terminal provided by an embodiment of the present disclosure;

[0097] Figure 7 is a structural block diagram of a server provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0098] In order to make the objects, technical solutions and advantages of the present disclosure clearer, the embodiments of the present disclosure will be further described in detail below with reference to the drawings.

[0099] The present disclosure provides a live room recommendation method, which can be implemented by a computer device. The computer device can be a terminal and a server, etc. The terminal can be a desktop computer, a notebook computer, a tablet computer, a mobile phone, etc. The server can be a single server or a server cluster, etc.

[0100] The computer device can include a processor, a memory, a communication component, etc.

[0101] The processor can be a central processing unit (CPU). The processor can be used to read instructions and process data, for example, determining a matching degree between each live room and a target account, determining a target new anchor live room, adjusting the matching degree between the target new anchor live room and the target account based on an exposure amount of each target anchor live room, determining a live room recommendation queue, etc.

[0102] The memory can be various volatile memory or non-volatile memory, such as a solid state disk (SSD), a dynamic random access memory (DRAM) memory, and the like. The memory can be used for data storage, for example, storing the matching degree between the live room and the target account, storing the determined target new anchor live room, storing the obtained updated matching degree, storing the determined live room recommendation queue, and the like.

[0103] The communication component can be a wired network connector, a wireless fidelity (WiFi) module, a Bluetooth module, a cellular network communication module, and the like. The communication component can be used for data transmission with other devices, for example, receiving a live room recommendation request sent by a target account, sending a live room recommendation queue to a target account for recommendation display, and the like.

[0104] The live room recommendation method provided by the present disclosure can be applied to various application programs with live function, for example, music application program, video application program, shopping application program, and the like. When a user opens the application program and clicks on the live room recommendation page, the background server of the application program can send the determined live room recommendation queue to the user terminal based on the live room recommendation method provided by the present disclosure, to display the recommended live rooms to the user.

[0105] Figure 1 And Figure 2 is a flowchart of a live room recommendation method provided by an embodiment of the present disclosure. Referring to Figure 1 And Figure 2 The embodiment includes:

[0106] 101. Receive a live room recommendation request sent by a target account.

[0107] In implementation, when a user wants to watch live in an application program, the user can click to open a live room recommendation page, thereby triggering the user terminal of the user to send a live room recommendation request to the background server of the application program, and the live room recommendation request carries the related account information of the target account logged in on the application program.

[0108] 102. Determine the matching degree between each live room and the target account.

[0109] In implementation, when the background server receives the live room recommendation request sent by the target account, the matching degree between the target account and each live room currently in live state in the application program can be determined. The higher the matching degree, the more interested the target account is in the live room corresponding to the matching degree. If the live room is recommended to the target account, the probability of the target account entering the live room is higher.

[0110] In a possible implementation, the method of determining the matching degree between each live room and the target account can be as follows: obtaining the user behavior sequence features and user portrait features corresponding to the target account; for each live room, obtaining the anchor offline real-time index features and anchor portrait features corresponding to the live room, the user-anchor cross features and context information features between the live room and the target account; determining the matching degree between the live room and the target account based on the user behavior sequence features and user portrait features corresponding to the target account, the anchor offline real-time index features and anchor portrait features corresponding to the live room, the user-anchor cross features and context information features between the live room and the target account, and a matching degree prediction model.

[0111] In implementation, the user behavior sequence features and user portrait features corresponding to the target account can be obtained, wherein the user behavior sequence features can include the ID of the live room previously entered by the target account, the ID of the live room followed by the target account, the ID of the live room paid by the target account, the ID of the live room chatted by the target account, and the like, and the user portrait features can include the age, gender, education, and the like of the target account.

[0112] Then, for each live room, the following processing can be performed respectively: obtaining the anchor offline real-time index features and anchor portrait features corresponding to the live room, wherein the anchor offline real-time index features can include the click rate, the average stay time after entering the room, the number of followers, and the like of the live room, and the anchor portrait features can include the gender, age, live type, and the like of the anchor corresponding to the live room.

[0113] The interaction features between the live room and the target account can also be obtained, including the user-anchor cross features and context information features between the live room and the target account, wherein the user-anchor cross features can include the stay time of the target account in the live room, the payment amount of the target account in the live room, and the like, and the context information features can include the application program in which the target account enters the live room, the recommended position of the live room when the target account enters the live room, and the like.

[0114] After the above-mentioned multiple features are acquired, the user behavior sequence features and the user portrait features corresponding to the target account, the anchor offline real-time index features and the anchor portrait features corresponding to the live broadcast room, the user-anchor cross features and the context information features between the live broadcast room and the target account are input into the matching degree prediction model to obtain the output matching degree between the live broadcast room and the target account.

[0115] In this way, the matching degree between each live broadcast room and the target account can be obtained through the above-mentioned method.

[0116] 103. Determine at least one target new anchor live broadcast room in multiple live broadcast rooms.

[0117] Since many new anchors have a short on-air time, the feature data corresponding to the new anchor acquired is less, which leads to an inaccurate matching degree between the new anchor and the target account, so that many new anchors that many users may be interested in cannot be pushed to the users, and the exposure of the new anchor is reduced.

[0118] Therefore, in the embodiments of the present disclosure, at least one target new anchor live broadcast room can be determined in multiple live broadcast rooms, and the anchor corresponding to the target new anchor live broadcast room is a new anchor.

[0119] In implementation, the live broadcast room corresponding to the new anchor can be determined in multiple live broadcast rooms first, and then at least one target new anchor live broadcast room can be selected from the multiple live broadcast rooms corresponding to the new anchor.

[0120] There can be multiple methods for determining the live broadcast room corresponding to the new anchor from multiple live broadcast rooms.

[0121] For example, whether the anchor is a new anchor can be determined based on the duration between the anchor signing time and the current time. When the duration between the anchor signing time and the current time is less than a first preset number of days, the anchor can be determined as a new anchor.

[0122] For another example, whether the anchor is a new anchor can be determined based on the on-air days of the anchor. When the on-air days of the anchor are less than a second preset number of days, the anchor can be determined as a new anchor.

[0123] The above-mentioned are only two kinds of methods for determining the new anchor, and the determination method can also be any other reasonable method. The first preset number of days and the second preset number of days can also be any reasonable value, and the embodiments of the present disclosure are not limited.

[0124] The method for determining the target new anchor live broadcast room in the multiple live broadcast rooms corresponding to the new anchor can also have multiple methods, and the specific method is introduced below, which is not described here.

[0125] 104、based on the to-be-exposed quantity of each target new anchor live room, respectively adjust the matching degree between each target new anchor live room and the target account, to obtain an updated matching degree between each target new anchor live room and the target account.

[0126] In implementation, after one or more target new anchor live rooms are determined, the matching degree between the target new anchor live room and the target account can be adjusted based on the to-be-exposed quantity of the target new anchor live room to obtain an updated matching degree, which is a value obtained after the matching degree and the to-be-exposed quantity are comprehensively considered, that is, the updated matching degree can not only reflect the matching degree between the target new anchor live room and the target account, but also reflect the exposure demand of the target new anchor live room.

[0127] The to-be-exposed quantity of the target new anchor live room is the difference between the target exposure quantity and the current actual exposure quantity of the target new anchor live room. The greater the to-be-exposed quantity, the greater the exposure demand of the target new anchor live room.

[0128] When the to-be-exposed quantity is large, the value of the matching degree can be adjusted to be increased, so that the updated matching degree is greater than the matching degree, and thus the recommendation probability of the target new anchor live room is increased. When the to-be-exposed quantity is small, the value of the matching degree can be adjusted to be decreased, so that the updated matching degree is less than the matching degree, and thus the recommendation probability of the target new anchor live room is reduced, and then the target new anchor live room with greater exposure demand is pushed to the user.

[0129] Alternatively, when the to-be-exposed quantity is large, the value of the matching degree can be adjusted to be increased greatly, so that the updated matching degree is greater than the matching degree, and the difference between them is relatively large. When the to-be-exposed quantity is small, the value of the matching degree can be adjusted to be increased slightly, so that the updated matching degree is greater than the matching degree, and the difference between them is relatively small. In this way, different target new anchor live rooms with different exposure demands can be respectively allocated with different supporting traffic to improve the exposure quantity of new anchors.

[0130] In this way, the updated matching degree can enable the live room recommendation in the subsequent step 105 to consider not only the matching degree between the live room and the target account, but also the exposure demand of the target new anchor live room, thereby improving the exposure quantity of new anchors to a certain extent.

[0131] In a possible implementation, when adjusting the matching degree, the to-be-exposed quantity mentioned above can be considered, and the current on-air duration of the target new anchor live room can also be considered. The processing method can be: based on the to-be-exposed quantity and the current on-air duration of each target new anchor live room, respectively adjust the matching degree between each target new anchor live room and the target account, to obtain an updated matching degree between each target new anchor live room and the target account.

[0132] In implementation, for each target new anchor live room, the matching degree between the target new anchor live room and the target account can be adjusted based on the to-be-exposed quantity and the current on-air duration of the target new anchor live room, so as to obtain an updated matching degree. The current on-air duration is the cumulative on-air duration of the target new anchor live room on the current day.

[0133] On the one hand, as described above, the updated matching degree takes into account the exposure demand of the target new anchor live room. On the other hand, the longer the current on-air duration of the target new anchor live room, the faster it will be off-air, so it is necessary to timely allocate support traffic to the target new anchor live room to make its to-be-exposed quantity close to the target exposure quantity.

[0134] Therefore, the to-be-exposed quantity and the current on-air duration can be comprehensively considered to adjust the matching degree and obtain an updated matching degree. In this way, the live room corresponding to the new anchor can be allocated support traffic more accurately and timely through the updated matching degree, so as to further improve the exposure quantity of the new anchor.

[0135] In a possible implementation, the specific method of adjusting the matching degree based on the to-be-exposed quantity and the current on-air duration can be as follows:

[0136] For each target new anchor live room, the to-be-exposed quantity, the time derivative of the to-be-exposed quantity, and the current on-air duration of the target new anchor live room are obtained. The proportional gain, the derivative gain, and the time gain are determined, wherein the proportional gain is positively correlated with the to-be-exposed quantity, the derivative gain is negatively correlated with the time derivative of the to-be-exposed quantity, and the time gain is positively correlated with the current on-air duration. The to-be-exposed quantity, the time derivative of the to-be-exposed quantity, and the current on-air duration are weighted based on the proportional gain, the derivative gain, and the time gain to obtain a weight value corresponding to the target new anchor live room. The weight value is multiplied by the matching degree between the target new anchor live room and the target account to obtain an updated matching degree between the target new anchor live room and the target account.

[0137] In implementation, the to-be-exposed quantity, the time derivative of the to-be-exposed quantity, and the current on-air duration of the target new anchor live room can be obtained first, wherein the time derivative of the to-be-exposed quantity is used to evaluate the traffic growth rate or the exposure quantity growth rate of the target new anchor live room.

[0138] The corresponding proportional gain can be determined or adjusted based on the to-be-exposed quantity of the target new anchor live room. The proportional gain is positively correlated with the to-be-exposed quantity, that is, the greater the to-be-exposed quantity, the greater the exposure demand, at which time the proportional gain can be increased to increase the weight of the to-be-exposed quantity.

[0139] The derivative gain corresponding to the target new anchor live room can also be determined or adjusted based on the time derivative of the to-be-exposed amount of the target new anchor live room. The derivative gain and the time derivative of the to-be-exposed amount are in a negative correlation relationship, that is, the greater the time derivative of the to-be-exposed amount, the greater the flow rate, at which time the derivative gain can be reduced to reduce the weight of the time derivative of the to-be-exposed amount.

[0140] The time gain can also be determined or adjusted based on the current on-air duration of the target new anchor live room. The time gain and the current on-air duration are in a positive correlation relationship, that is, the greater the current on-air duration, the greater the possibility of off-air, at which time the time gain can be increased to increase the weight of the current on-air duration.

[0141] After the proportional gain, the derivative gain, and the time gain are determined, the weighted sum of the to-be-exposed amount, the time derivative of the to-be-exposed amount, and the current on-air duration can be calculated based on the proportional gain, the derivative gain, and the time gain to obtain the weight value corresponding to the target new anchor live room. Then, the weight value is multiplied by the matching degree between the target new anchor live room and the target account to obtain the updated matching degree.

[0142] Referring to the following formula (1) and formula (2), the weight value corresponding to the target new anchor live room can be calculated based on the formula (1) and formula (2):

[0143] e(t) = r(t) - y(t) (1)

[0144]

[0145] Wherein, e(t) is the to-be-exposed amount of the target new anchor live room, r(t) is the target exposure amount of the target new anchor live room, y(t) is the current actual exposure amount of the target new anchor live room, u(t) is the weight value corresponding to the target new anchor live room, K p is the proportional gain, K d is the derivative gain, K t is the time gain, and T is the current on-air duration of the target new anchor live room.

[0146] In the embodiments of the present disclosure, the proportional gain, the derivative gain, and the time gain can be respectively provided with corresponding upper and lower limits to prevent the overall fluctuation from being too large.

[0147] 105、Based on the updated matching degree between the target new anchor live room and the target account, and the matching degree between other live rooms except the target new anchor live room and the target account, a live room recommendation queue is determined and sent to the target account for recommendation display.

[0148] In implementation, the multiple live broadcast rooms can be sorted based on the updated matching degree corresponding to the target new anchor live broadcast room and the matching degrees corresponding to the other live broadcast rooms, the first preset number of live broadcast rooms in the sorting form a live broadcast room recommendation queue, and the background server can send the live broadcast room recommendation queue to the target account. After the target account receives the live broadcast room recommendation queue, the target account displays the live broadcast room recommendation queue on the corresponding recommendation position in the live broadcast recommendation page according to the sorting (see Figure 3 ).

[0149] The method of sorting the multiple live broadcast rooms based on the updated matching degree and the matching degree can be to sort directly according to the order from large to small of the numerical values of the updated matching degree and the matching degree, or can be to sort by comprehensively considering the updated matching degree, the matching degree, and other parameters (for example, whether it is currently during a holiday, whether it is during a certain activity, and the like), and the like. The embodiments of the present disclosure do not limit this.

[0150] In this way, the matching degree between the target new anchor live broadcast room and the target account is adjusted by the to-be-exposed amount of each target new anchor live broadcast room, so that the obtained updated matching degree can meet the exposure demand of the target new anchor live broadcast room on the premise of meeting the effective matching between the target account and the target new anchor live broadcast room. Therefore, the target new anchor live broadcast room is recommended based on the updated matching degree, which can further push the new anchor live broadcast room that the target account is interested in and has a high exposure demand to the target account, thereby improving the exposure amount of the new anchor.

[0151] Moreover, according to experiments, the number of new anchors with a monthly income of 3,000 yuan is 387 using the live broadcast room recommendation method mentioned in the related art, and the number of new anchors with a monthly income of 3,000 yuan is 645 using the live broadcast room recommendation method provided in the embodiments of the present disclosure. The number is obviously increased, which can effectively improve the exposure amount of the new anchor.

[0152] In the embodiments of the present disclosure, the method of determining at least one target new anchor live broadcast room in the multiple live broadcast rooms can be various. Next, see Figure 4 , the possible implementation methods are introduced:

[0153] 401. Determine a high-quality new anchor in the multiple new anchors.

[0154] In implementation, the new anchor can be screened when each new anchor enters the application program, and the new anchor that meets the screening condition is determined as a high-quality new anchor. The high-quality new anchor is a new anchor that has a better live broadcast effect or more user clicks, collections, and the like in the new anchors.

[0155] In a possible implementation, the screening method can be: determining a recommendation score of each new anchor based on historical live data of each new anchor, wherein the historical live data includes at least one of a click rate, a number of live broadcast days, a daily accumulated live broadcast time, a user effective stay rate, and a fan attention conversion rate; if the recommendation score of the new anchor is greater than or equal to a preset score threshold, or if the artificial review determination of the new anchor is passed, the new anchor is determined to be a high-quality new anchor.

[0156] In implementation, in one aspect, for each new anchor, historical live data of the new anchor can be obtained, a recommendation score of the new anchor can be calculated based on the historical live data, and if the recommendation score is greater than or equal to a preset score threshold, the new anchor can be determined to be a high-quality new anchor.

[0157] In another aspect, the artificial review determination can also be used to form a multi-role evaluation group (operation, user representative, algorithm expert, etc.), a double-blind scoring mechanism can be used, and scoring can be performed from multiple dimensions (lens performance, image recognition, scene arrangement quality, etc.). A dynamic threshold can be set, that is, a preset number of new anchors in the same period before artificial scoring can be determined to be passed, that is, determined to be a high-quality new anchor.

[0158] Through the above two methods, high-quality new anchors among new anchors can be screened.

[0159] In the above method of screening new anchors based on historical live data of the new anchors, the historical live data can include at least one of a click rate, a number of live broadcast days, a daily accumulated live broadcast time, a user effective stay rate, and a fan attention conversion rate.

[0160] When there are multiple historical live data, the multiple historical live data lines can be normalized to obtain a normalized value corresponding to each type of historical live data. Then, the normalized values corresponding to the multiple historical live data can be weighted according to a preset weight corresponding to each type of historical live data, that is, a recommendation score of the new anchor is obtained.

[0161] For example, the historical live data includes the number of days of live broadcast, the cumulative live broadcast time per day, the user effective stay rate, and the fan attention conversion rate. When calculating the recommendation score of the new host, the number of days of live broadcast, the cumulative live broadcast time per day, the user effective stay rate, and the fan attention conversion rate of the new host can be obtained first, and then the four data are normalized to obtain the corresponding normalized values. Then, the number of days of live broadcast, the cumulative live broadcast time per day, the user effective stay rate, and the fan attention conversion rate are calculated based on the first preset weight corresponding to the number of days of live broadcast, the second preset weight corresponding to the cumulative live broadcast time per day, the third preset weight corresponding to the user effective stay rate, and the fourth preset weight corresponding to the fan attention conversion rate. The weighted sum of the normalized values of the number of days of live broadcast, the cumulative live broadcast time per day, the user effective stay rate, and the fan attention conversion rate is obtained to obtain the recommendation score of the new host.

[0162] Alternatively, when filtering based on the multiple historical live data, XGBoost (eXtreme Gradient Boosting, optimized distributed gradient boosting library) or LR (Logistic Regression, logistic regression) model, etc. can also be used, and the present disclosure does not make specific limitations.

[0163] Through the above method, each new host can be filtered. The hosts that pass the filtering are determined as high-quality new hosts, and the hosts that do not pass the filtering are eliminated. For the eliminated new hosts, a recovery mechanism can be set. The eliminated new hosts can be filtered again every week. The filtering method can be based on the historical live data of the new host and the preset weight of each historical live data to perform weighted calculation to obtain the recommendation score of the new host. If the recommendation score is greater than or equal to a preset score threshold, the eliminated new host can be recovered as a high-quality new host.

[0164] In the above filtering method, a time decay function can also be used to correct the early data sparsity for the same kind of historical live data. That is, if the historical live data is recent data, the preset weight of the historical live data can be multiplied by a preset coefficient (greater than 1) to obtain a new preset weight. When calculating the recommendation score, the new preset weight can be used for weighted calculation, so that the obtained recommendation score is more consistent with the latest live performance of the new host.

[0165] In one possible implementation, for the hosts determined as high-quality new hosts, a real-time supervision mechanism is also needed. The specific method is as follows:

[0166] The mean and variance of the first live data of the multiple high-quality new hosts are calculated, and the difference between the mean and the variance is calculated. If the first live data of the high-quality new host is always less than the difference in the first latest preset time period, the to-be-exposed amount of the high-quality new host is adjusted downward, and the to-be-exposed amount of the high-quality new host is updated.

[0167] In implementation, the mean and variance of the first live data of the plurality of high-quality new anchors can be periodically calculated in the first latest preset period, and the difference between the two can be calculated. Then, for each high-quality new anchor, the first live data of the high-quality new anchor can be periodically obtained, and if the first live data of the high-quality new anchor is lower than the corresponding difference in the first latest preset period, it indicates that the live performance of the high-quality new anchor in the first latest preset period is poor. At this time, a first-level warning can be issued, and the real-time traffic of the high-quality new anchor is adjusted downward, that is, the to-be-exposed amount of the high-quality new anchor is adjusted downward to obtain a new to-be-exposed amount of the high-quality new anchor, and the to-be-exposed amount of the high-quality new anchor is updated. In this way, in the processing of the above step 104, if the new anchor corresponding to the target new anchor live room is the anchor for which the first-level warning is issued, the to-be-exposed amount of the target new anchor live room used is the updated to-be-exposed amount.

[0168] The first latest preset period can be three days, and the period can be one day, that is, the mean and variance of the first live data are calculated every day, and the difference between the two is calculated.

[0169] Then, for each high-quality new anchor, the following processing is performed: the first live data of the high-quality new anchor is obtained every day, and if the first live data of the high-quality new anchor is lower than the difference of the day in the last three days, a first-level warning is issued to the high-quality new anchor, and the real-time traffic of the high-quality new anchor is adjusted downward.

[0170] In this way, more support traffic can be allocated to new anchors with good live performance when the matching degree is updated for recommendation, thereby improving the exposure amount of new anchors.

[0171] In the present embodiment, the first live data includes at least one of the click rate and the user effective stay rate.

[0172] When there are multiple first live data, the above judgment can be performed for each first live data, and if each first live data of the high-quality new anchor is less than the difference in the first latest preset period, the to-be-exposed amount of the high-quality new anchor is adjusted downward.

[0173] Further, after the to-be-exposed amount of the high-quality new anchor with poor live performance is adjusted downward, the following supervision mechanism can be performed on it:

[0174] Based on the first live data of each high-quality new anchor, a high-quality ranking is obtained by sorting from large to small. For the high-quality new anchor whose to-be-exposed amount is adjusted downward, a week-on-week decrease value of the first live data of the high-quality new anchor is determined, and the high-quality new anchor whose week-on-week decrease value exceeds a preset change threshold among the first preset proportion of high-quality new anchors at the back of the high-quality ranking is returned to the new anchor.

[0175] In implementation, the plurality of high-quality new anchors can be sorted in descending order of the first live data every week to obtain a high-quality ranking.

[0176] In addition, for each high-quality new anchor that has undergone the decrease adjustment of the to-be-exposed amount, the following judgment can also be performed: a week-over-week decrease value of the first live data of the high-quality new anchor is calculated, i.e., a decrease ratio of the first live data of the latest week compared with that of the last week, if the week-over-week decrease value of the first live data of the high-quality new anchor exceeds a preset change threshold, and the high-quality new anchor is one of the first preset percentage of high-quality new anchors at the rear position in the high-quality ranking, it is determined that the live performance of the high-quality new anchor is extremely poor, and the high-quality new anchor is removed from the set of high-quality new anchors, i.e., is no longer a high-quality new anchor, and is returned to the category of ordinary new anchors. Alternatively, when it is determined that the live performance of the high-quality new anchor is extremely poor, the high-quality new anchor is not immediately removed, but is flexibly removed after waiting for a preset time period.

[0177] The preset change threshold and the first preset percentage can be any reasonable values, for example, the preset change threshold can be 20%, and the first preset percentage can be 10%, so that the high-quality new anchor whose week-over-week decrease value exceeds 20% and that is located at the rear 10% in the high-quality ranking is removed from the set of high-quality new anchors and returned to the category of ordinary new anchors.

[0178] It can be understood that when there are multiple first live data, the above judgment can be performed for each type of first live data, and when the week-over-week decrease value of each type of first live data and the position of the high-quality new anchor in the high-quality ranking corresponding to each type of first live data both satisfy the above conditions, the high-quality new anchor can be removed from the set of high-quality new anchors.

[0179] In addition to the above removal method, when the high-quality new anchor is no longer a new anchor as the live time length is prolonged, the high-quality new anchor becomes an old anchor and is naturally removed from the set of high-quality new anchors.

[0180] In a possible implementation, for an anchor that has been determined to be a high-quality new anchor, the disclosure embodiments further set a blacklist mechanism: if the monthly average live time length of the high-quality new anchor is lower than a preset live time length threshold, or the high-quality new anchor performs data brushing or other improper behaviors, the high-quality new anchor is pulled into the blacklist, the high-quality new anchor is removed from the set of high-quality new anchors and returned to the category of ordinary new anchors, and will not be determined to be a high-quality new anchor again in the future.

[0181] 402、determine the high-quality new anchor live room corresponding to at least one high-quality new anchor in a live state.

[0182] In implementation, when it is needed to determine a high-quality new anchor live room among multiple live rooms, it can be judged in sequence whether the anchor corresponding to the live room in a live state is a high-quality new anchor, and if so, the live room is a high-quality new anchor live room.

[0183] Alternatively, a live room that starts broadcasting can also be determined among multiple high-quality new anchors, that is, a high-quality new anchor live room is obtained.

[0184] 403. Based on the to-be-exposed amount and the performance score of each high-quality new anchor live room, at least one target new anchor live room is determined among the high-quality new anchor live rooms.

[0185] In implementation, at least one target new anchor live room can be selected among multiple high-quality new anchor live rooms based on the to-be-exposed amount and the performance score.

[0186] In a possible implementation, the method for determining a target new anchor live room among high-quality new anchor live rooms can be: determining a high-quality new anchor live room whose to-be-exposed amount is greater than a preset to-be-exposed amount threshold and whose performance score is greater than a preset performance score threshold as the target new anchor live room.

[0187] Alternatively, in another possible implementation, the method for determining a target new anchor live room among high-quality new anchor live rooms can also be: determining a recommendation probability corresponding to each high-quality new anchor live room based on the to-be-exposed amount of each high-quality new anchor live room, wherein the recommendation probability and the to-be-exposed amount are in a positive correlation; performing sorting from large to small based on the performance score of each high-quality new anchor live room to obtain a score ranking; multiplying the recommendation probability of a second preset proportion of high-quality new anchor live rooms in the front position in the score ranking by a first coefficient, multiplying the recommendation probability of a third preset proportion of high-quality new anchor live rooms in the rear position in the score ranking by a second coefficient, and updating the recommendation probability of the high-quality new anchor live room, wherein the first coefficient is greater than 1, and the second coefficient is greater than 0 and less than 1; performing probability sampling on multiple high-quality new anchor live rooms based on the recommendation probability of each high-quality new anchor live room to obtain at least one target new anchor live room.

[0188] In implementation, the recommendation probability corresponding to each high-quality new anchor live room can be determined based on the to-be-exposed amount of each high-quality new anchor live room first. The greater the to-be-exposed amount, the greater the exposure demand of the high-quality new anchor live room, and the greater the corresponding recommendation probability, so that the high-quality new anchor is more likely to be recommended.

[0189] In addition, the performance score of each high-quality new anchor live room can also be determined. The higher the performance score, the better the current live performance of the high-quality new anchor live room. After the performance score of each high-quality new anchor live room is determined, the multiple high-quality new anchor live rooms can be sorted in order of performance score from large to small to obtain a score ranking.

[0190] Then, the recommendation probability of the second preset proportion of high-quality new anchor live rooms in the front of the score ranking is multiplied by the first coefficient, and the recommendation probability of the second preset proportion of high-quality new anchor live rooms is updated, that is, the recommendation probability of the second preset proportion of high-quality new anchor live rooms is replaced by the value after the corresponding recommendation probability is multiplied by the first coefficient. Since the first coefficient is greater than 1, the updated recommendation probability of the high-quality new anchor live room is greater, which can improve the recommendation probability of the high-quality new anchor live room with good performance.

[0191] The recommendation probability of the third preset proportion of high-quality new anchor live rooms in the back of the score ranking is multiplied by the second coefficient, and the recommendation probability of the third preset proportion of high-quality new anchor live rooms is updated, that is, the recommendation probability of the third preset proportion of high-quality new anchor live rooms is replaced by the value after the corresponding recommendation probability is multiplied by the second coefficient. Since the second coefficient is greater than 0 and less than 1, the updated recommendation probability of the high-quality new anchor live room is smaller, which can reduce the recommendation probability of the high-quality new anchor live room with general performance.

[0192] After the recommendation probability is updated by the above method, the multiple high-quality new anchor live rooms can be probabilistically sampled based on the recommendation probability of each high-quality new anchor live room, so as to obtain at least one target new anchor live room. The high-quality new anchor live room with a larger recommendation probability is more likely to be selected as a target new anchor live room, and the high-quality new anchor live room with a smaller recommendation probability is less likely to be selected.

[0193] In the above probabilistic sampling, the number of target new anchor live rooms can be limited first, that is, when the multiple high-quality new anchor live rooms are probabilistically sampled, a preset number of target new anchor live rooms are selected according to the recommendation probability of each high-quality new anchor live room.

[0194] Alternatively, the total to-be-exposed amount of the target new anchor live room can also be limited first, which can be limited to the current exposure amount that can be allocated to the new anchor. In this way, when the multiple high-quality new anchor live rooms are probabilistically sampled, one or more target new anchor live rooms are selected according to the recommendation probability of each high-quality new anchor live room, so that the total to-be-exposed amount of the selected target new anchor live room is equal to or slightly greater than the current exposure amount that can be allocated to the new anchor.

[0195] In a possible implementation, in the process of determining the target new anchor live room, the quantity to be exposed of each high-quality new anchor live room needs to be obtained, and the method can be as follows: for each high-quality new anchor, based on the anchor portrait data of the high-quality new anchor and the anchor portrait data of other anchors except the new anchor, other anchors matching the high-quality new anchor are determined; based on the historical exposure quantity of the live room corresponding to the other anchor matching the high-quality new anchor and the current exposure quantity of the high-quality new anchor live room corresponding to the high-quality new anchor, the quantity to be exposed of the high-quality new anchor live room corresponding to the high-quality new anchor is determined.

[0196] In implementation, the anchor portrait data of the high-quality new anchor can be obtained, and the anchor portrait data of other anchors (for example, old anchors) except the new anchor in the application program is obtained, the anchor portrait data of the high-quality new anchor is matched with similar other anchors, and one or more other anchors matching the high-quality new anchor are determined.

[0197] Then, based on the historical exposure quantity of the live room corresponding to the other anchor matching the high-quality new anchor, the target exposure quantity of the high-quality new anchor live room lost by the high-quality new anchor can be predicted.

[0198] The difference between the predicted target exposure quantity of the high-quality new anchor live room and the current exposure quantity of the high-quality new anchor live room is calculated to obtain the quantity to be exposed of the high-quality new anchor live room.

[0199] In the process of determining the other anchor matching the high-quality new anchor, the method can be as follows: the other anchor whose anchor portrait data is the same as or belongs to the same preset numerical range as the anchor portrait data of the high-quality new anchor is determined as the other anchor matching the high-quality new anchor.

[0200] For example, the anchor portrait data can include the talent type and the historical index performance (for example, the amount of received gifts), and the other anchor whose talent type is the same as that of the high-quality new anchor and whose historical index performance belongs to the same preset numerical range as that of the high-quality new anchor is determined as the other anchor matching the high-quality new anchor.

[0201] It can be understood that the compared historical index performance of the high-quality new anchor and the historical index performance of the other anchor refer to the data of the same broadcasting day or the same broadcasting period, for example, both are the Nth day of broadcasting or both are the first month of broadcasting.

[0202] Alternatively, in addition to the above method, the Prophet model (a time series prediction model) can also be used to predict the quantity of high-quality new anchor live rooms to be exposed. Prophet is an open-source time series prediction model suitable for data with seasonal and trend changes, and can handle holiday effects and missing values. These features are useful in live streaming traffic prediction, as live streaming traffic may have periodic fluctuations, such as daily peak hours or specific times of the week.

[0203] In a possible implementation, in the process of determining the target new anchor live room, the performance score of each high-quality new anchor live room needs to be obtained, and the method can be: based on the second live data of each high-quality new anchor live room in the second latest preset time period, determining the performance score of each high-quality new anchor live room.

[0204] In implementation, the second live data can include at least one of a click rate, an average stay time after entering the room, and a second-off rate after entering the room. Among them, the click rate and the average stay time after entering the room are positive feedback factors, and the second-off rate after entering the room is a negative feedback factor.

[0205] If the second live data has only one positive feedback factor, the second live data can be directly used as the performance score of the high-quality new anchor live room.

[0206] If the second live data has only one negative feedback factor, the negative of the second live data can be used as the performance score of the high-quality new anchor live room.

[0207] If the second live data includes positive feedback factors and negative feedback factors, the positive feedback factors and the negative feedback factors can be multiplied by their respective weights to obtain weighted positive feedback factors and weighted negative feedback factors, and then the difference between the weighted positive feedback factors and the weighted negative feedback factors is calculated to obtain the performance score of the high-quality new anchor live room.

[0208] For example, the second live data can include a click rate, an average stay time after entering the room, and a second-off rate after entering the room. In calculating the performance score of the high-quality new anchor live room, the click rate can be multiplied by the corresponding fifth preset weight to obtain a weighted click rate; the average stay time after entering the room can be multiplied by the corresponding sixth preset weight to obtain a weighted average stay time after entering the room; and the second-off rate after entering the room can be multiplied by the corresponding seventh preset weight to obtain a weighted second-off rate after entering the room. Then, the sum of the weighted click rate and the weighted average stay time after entering the room can be calculated to obtain a first value, and the difference between the first value and the weighted second-off rate after entering the room is calculated to obtain the performance score of the high-quality new anchor live room.

[0209] It can be understood that the above method is only a few possible methods for determining the target new anchor live room process, and other any reasonable methods can also be used, and the embodiments of the present disclosure do not limit this.

[0210] All the optional technical solutions described above can be combined to form optional embodiments of the present disclosure, which will not be repeated here.

[0211] The scheme mentioned in the embodiments of the present disclosure can adjust the matching degree between the target new anchor live room and the target account by the to-be-exposed amount of each target new anchor live room, so that the updated matching degree meets the effective matching between the target account and the target new anchor live room, and also meets the exposure demand of the target new anchor live room. Therefore, based on the updated matching degree, the target new anchor live room is recommended, which can better push the new anchor live room that the target account is interested in and has a high exposure demand to the target account, thereby improving the exposure amount of the new anchor.

[0212] The embodiments of the present disclosure provide a live room recommendation device, which can be a computer device in the above embodiments, as shown in Figure 5 , the device comprises:

[0213] The receiving module 510 is configured to receive a live room recommendation request sent by a target account.

[0214] The matching degree determination module 520 is configured to determine the matching degree between each live room and the target account.

[0215] The target determination module 530 is configured to determine at least one target new anchor live room in a plurality of live rooms.

[0216] The updating module 540 is configured to adjust the matching degree between each target new anchor live room and the target account based on the to-be-exposed amount of each target new anchor live room, to obtain the updated matching degree between each target new anchor live room and the target account.

[0217] The queue determination module 550 is configured to determine a live room recommendation queue based on the updated matching degree between the target new anchor live room and the target account, and the matching degree between other live rooms except the target new anchor live room and the target account, and send the live room recommendation queue to the target account for recommendation display.

[0218] In a possible implementation, the target determination module 530 is configured to:

[0219] determine a high-quality new anchor in a plurality of new anchors;

[0220] determine at least one high-quality new anchor live room corresponding to a high-quality new anchor in a live state;

[0221] determine the at least one target new anchor live room in the high-quality new anchor live room based on the exposure amount to be exposed and the performance score of each high-quality new anchor live room.

[0222] In a possible implementation, the target determination module 530 is configured to:

[0223] determine a recommendation score of each new anchor based on historical live data of the new anchor, wherein the historical live data comprises at least one of a click rate, a number of live broadcast days, a daily cumulative live broadcast time, a user effective stay rate, and a fan attention conversion rate;

[0224] if the recommendation score of the new anchor is greater than or equal to a preset score threshold, or if the artificial review of the new anchor is passed, the new anchor is determined as the high-quality new anchor.

[0225] In a possible implementation, the target determination module 530 is further configured to:

[0226] calculate a mean and a variance of first live data of a plurality of high-quality new anchors, and calculate a difference value between the mean and the variance, wherein the first live data comprises at least one of a click rate and a user effective stay rate;

[0227] if the first live data of the high-quality new anchor is less than the difference value all the time in a first latest preset period, the exposure amount to be exposed of the high-quality new anchor is adjusted downward, and the exposure amount to be exposed of the high-quality new anchor is updated.

[0228] In a possible implementation, the target determination module 530 is further configured to:

[0229] sort each high-quality new anchor based on the first live data from large to small to obtain a high-quality ranking;

[0230] for the high-quality new anchor whose exposure amount to be exposed is adjusted downward, determine a week-on-week decrease value of the first live data of the high-quality new anchor, and return the high-quality new anchor whose week-on-week decrease value exceeds a preset change threshold to the new anchor from the first preset proportion of high-quality new anchors at the back of the high-quality ranking.

[0231] In a possible implementation, the target determination module 530 is configured to:

[0232] determine a recommendation probability corresponding to each high-quality new anchor live room based on the exposure amount to be exposed of the high-quality new anchor live room, wherein the recommendation probability and the exposure amount to be exposed are in a positive correlation.

[0233] performing a descending ordering based on the performance scores of each high-quality new anchor live room, to obtain a score ordering;

[0234] multiplying the recommendation probability of a second preset proportion of high-quality new anchor live rooms at the front of the score ordering by a first coefficient, multiplying the recommendation probability of a third preset proportion of high-quality new anchor live rooms at the back of the score ordering by a second coefficient, and updating the recommendation probability of the high-quality new anchor live room, wherein the first coefficient is greater than 1, and the second coefficient is greater than 0 and less than 1;

[0235] performing probability sampling on the plurality of high-quality new anchor live rooms based on the recommendation probability of each high-quality new anchor live room, to obtain the at least one target new anchor live room.

[0236] In a possible implementation, the target determination module 530 is further configured to:

[0237] for each high-quality new anchor, determining other anchors matched with the high-quality new anchor based on anchor portrait data of the high-quality new anchor and anchor portrait data of other anchors except the new anchor;

[0238] determining a to-be-exposed amount of the high-quality new anchor live room corresponding to the high-quality new anchor based on a historical exposure amount of a live room corresponding to the other anchor matched with the high-quality new anchor and a current exposure amount of the high-quality new anchor live room corresponding to the high-quality new anchor.

[0239] In a possible implementation, the target determination module 530 is further configured to:

[0240] determining a performance score of each high-quality new anchor live room based on second live data of the high-quality new anchor live room in a second latest preset time period, wherein the second live data includes at least one of a click rate, an average stay time after entering a room, and a second-leave rate after entering a room.

[0241] In a possible implementation, the updating module 540 is configured to:

[0242] adjusting a matching degree between each target new anchor live room and the target account based on the to-be-exposed amount and the current on-air time length of the target new anchor live room, to obtain an updated matching degree between each target new anchor live room and the target account.

[0243] In a possible implementation, the updating module 540 is configured to:

[0244] For each target new anchor live room, obtain a to-be-exposed amount of the target new anchor live room, a time derivative of the to-be-exposed amount, and a current on-air duration;

[0245] determine a proportional gain, a derivative gain, and a time gain, wherein the proportional gain is in a positive correlation with the to-be-exposed amount, the derivative gain is in a negative correlation with the time derivative of the to-be-exposed amount, and the time gain is in a positive correlation with the current on-air duration;

[0246] based on the proportional gain, the derivative gain, and the time gain, perform weighted processing on the to-be-exposed amount, the time derivative of the to-be-exposed amount, and the current on-air duration, to obtain a weight value corresponding to the target new anchor live room;

[0247] multiply the weight value by a matching degree between the target new anchor live room and the target account, to obtain an updated matching degree between the target new anchor live room and the target account.

[0248] In a possible implementation, the matching degree determination module 520 is configured to:

[0249] obtain user behavior sequence features and user portrait features corresponding to the target account;

[0250] for each live room, obtain anchor offline real-time indicator features and anchor portrait features corresponding to the live room, user-anchor cross features and context information features between the live room and the target account;

[0251] based on the user behavior sequence features and the user portrait features corresponding to the target account, the anchor offline real-time indicator features and the anchor portrait features corresponding to the live room, the user-anchor cross features and the context information features between the live room and the target account, and a matching degree prediction model, determine the matching degree between the live room and the target account.

[0252] It should be noted that the live room recommendation apparatus provided in the above embodiments is only used as an example for dividing the above functions in the process of recommending a live room, and in actual applications, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the apparatus is divided into different functional modules to complete all or part of the functions described above. In addition, the live room recommendation apparatus and the live room recommendation method provided in the above embodiments belong to the same concept, and the specific implementation process is described in detail in the method embodiments, which will not be repeated here.

[0253] Figure 6A structural block diagram of a terminal 600 provided by one exemplary embodiment of the present disclosure is shown. The terminal can be the computer device in the above embodiments. The terminal 600 can be a smartphone, a tablet computer, an MP3 player, an MP4 player, a notebook computer, or a desktop computer. The terminal 600 can also be referred to as a user equipment, a portable terminal, a laptop terminal, a desktop terminal, or other names.

[0254] Generally, the terminal 600 includes a processor 601 and a memory 602.

[0255] The processor 601 can include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 601 can be implemented in at least one of a hardware form of a DSP (digital signal processing), an FPGA (field-programmable gate array), a PLA (programmable logic array). The processor 601 can also include a main processor and a coprocessor. The main processor is a processor for processing data in an awake state, also referred to as a CPU (central processing unit). The coprocessor is a low-power processor for processing data in a standby state. In some embodiments, the processor 601 can be integrated with a GPU (graphics processing unit) for rendering and drawing content required to be displayed by a display screen. In some embodiments, the processor 601 can further include an AI (artificial intelligence) processor for processing computing operations related to machine learning.

[0256] The memory 602 can include one or more computer-readable storage media, which can be non-transitory. The memory 602 can also include a high-speed random access memory, and a non-volatile memory such as one or more disk storage devices, flash storage devices. In some embodiments, the non-transitory computer-readable storage medium in the memory 602 is used to store at least one instruction for being executed by the processor 601 to implement the live room recommendation method provided by the method embodiments in the present disclosure.

[0257] In some embodiments, terminal 600 can further include a peripheral device interface 603 and at least one peripheral device. The processor 601, the memory 602, and the peripheral device interface 603 can be connected through a bus or a signal line. Each peripheral device can be connected to the peripheral device interface 603 through a bus, a signal line, or a circuit board. Specifically, the peripheral device includes at least one of a radio frequency circuit 604, a display screen 605, a camera 606, an audio circuit 607, a positioning component 608, and a power supply 609.

[0258] The peripheral device interface 603 can be used to connect at least one peripheral device related to input / output to the processor 601 and the memory 602. In some embodiments, the processor 601, the memory 602, and the peripheral device interface 603 are integrated on the same chip or circuit board; in some other embodiments, any one or two of the processor 601, the memory 602, and the peripheral device interface 603 can be implemented on a separate chip or circuit board, and the present embodiment is not limited in this regard.

[0259] The radio frequency circuit 604 is used to receive and transmit RF (radio frequency) signals, also known as electromagnetic signals. The radio frequency circuit 604 communicates with a communication network and other communication devices through electromagnetic signals. The radio frequency circuit 604 converts electrical signals into electromagnetic signals for transmission, or converts received electromagnetic signals into electrical signals. Optionally, the radio frequency circuit 604 includes an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a subscriber identity module card, and the like. The radio frequency circuit 604 can communicate with other terminals through at least one wireless communication protocol. The wireless communication protocol includes but is not limited to metropolitan area networks, various generations of mobile communication networks (2G, 3G, 4G, and 5G), wireless local area networks, and / or WiFi (wireless fidelity) networks. In some embodiments, the radio frequency circuit 604 can also include NFC (near field communication) related circuitry, and the present disclosure is not limited in this regard.

[0260] The display screen 605 is configured to display a UI (user interface). The UI can include graphics, text, icons, video, and any combination thereof. When the display screen 605 is a touch display screen, the display screen 605 is further configured to capture touch signals on or above the surface of the display screen 605. The touch signals can be input to the processor 601 as control signals for processing. In this case, the display screen 605 can also be configured to provide virtual buttons and / or virtual keyboard, also known as soft buttons and / or soft keyboard. In some embodiments, the display screen 605 can be one, arranged on the front panel of the terminal 600; in other embodiments, the display screen 605 can be at least two, arranged on different surfaces of the terminal 600 or in a folding design; in still other embodiments, the display screen 605 can be a flexible display screen, arranged on a curved surface or a folding surface of the terminal 600. Even, the display screen 605 can also be arranged in an irregular shape other than a rectangle, i.e., a special-shaped screen. The display screen 605 can be made of LCD (liquid crystal display), OLED (organic light-emitting diode), etc.

[0261] The camera assembly 606 is configured to capture images or videos. Optionally, the camera assembly 606 includes a front-facing camera and a rear-facing camera. Typically, the front-facing camera is arranged on the front panel of the terminal, and the rear-facing camera is arranged on the back of the terminal. In some embodiments, the rear-facing camera is at least two, which are any one of a main camera, a depth-of-field camera, a wide-angle camera, and a telephoto camera, to realize the background blur function by fusing the main camera and the depth-of-field camera, the panoramic shooting and VR (virtual reality) shooting function by fusing the main camera and the wide-angle camera, or other fusion shooting functions. In some embodiments, the camera assembly 606 can further include a flash. The flash can be a single-color-temperature flash or a dual-color-temperature flash. The dual-color-temperature flash refers to a combination of a warm light flash and a cold light flash, which can be used for light compensation under different color temperatures.

[0262] The audio circuit 607 can include a microphone and a speaker. The microphone is used to collect sound waves of the user and the environment, and convert the sound waves into an electrical signal input to the processor 601 for processing, or input to the radio frequency circuit 604 to realize voice communication. For the purpose of stereo sound collection or noise reduction, the microphone can be multiple, respectively arranged at different parts of the terminal 600. The microphone can also be an array microphone or an omnidirectional collection type microphone. The speaker is used to convert the electrical signal from the processor 601 or the radio frequency circuit 604 into sound waves. The speaker can be a conventional diaphragm speaker, or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, not only can the electrical signal be converted into a sound wave that humans can hear, but also can be converted into a sound wave that humans cannot hear for ranging purposes. In some embodiments, the audio circuit 607 can also include a headphone jack.

[0263] The positioning component 608 is used to position the current geographic location of the terminal 600 to realize navigation or LBS (location based service). The positioning component 608 can be a positioning component based on GPS (global positioning system), Beidou system, Glonass system or Galileo system.

[0264] The power supply 609 is used to supply power to each component in the terminal 600. The power supply 609 can be alternating current, direct current, disposable battery or rechargeable battery. When the power supply 609 includes a rechargeable battery, the rechargeable battery can support wired charging or wireless charging. The rechargeable battery can also be used to support fast charging technology.

[0265] In some embodiments, the terminal 600 also includes one or more sensors 610. The one or more sensors 610 include but are not limited to: an acceleration sensor 611, a gyroscope sensor 612, a pressure sensor 613, a fingerprint sensor 614, an optical sensor 615 and a proximity sensor 616.

[0266] The acceleration sensor 611 can detect the acceleration in three coordinate axes of the coordinate system established by the terminal 600. For example, the acceleration sensor 611 can be used to detect the components of gravitational acceleration in three coordinate axes. The processor 601 can control the display screen 605 to display the user interface in a landscape view or a portrait view according to the gravitational acceleration signal collected by the acceleration sensor 611. The acceleration sensor 611 can also be used for game or user motion data collection.

[0267] The gyroscope sensor 612 can detect the body direction and rotation angle of the terminal 600, and can collect 3D motions of the user on the terminal 600 in cooperation with the acceleration sensor 611. The processor 601 can implement the following functions according to the data collected by the gyroscope sensor 612: motion sensing (such as changing the UI according to the tilt operation of the user), image stabilization when shooting, game control, and inertial navigation.

[0268] The pressure sensor 613 can be arranged on the side frame of the terminal 600 and / or the lower layer of the display screen 605. When the pressure sensor 613 is arranged on the side frame of the terminal 600, the holding signal of the user on the terminal 600 can be detected, and the left-hand or right-hand recognition or shortcut operation can be performed by the processor 601 according to the holding signal collected by the pressure sensor 613. When the pressure sensor 613 is arranged on the lower layer of the display screen 605, the controllable control on the UI interface can be controlled by the processor 601 according to the pressure operation of the user on the display screen 605. The controllable control includes at least one of the button control, the scroll bar control, the icon control, and the menu control.

[0269] The fingerprint sensor 614 is used to collect the fingerprint of the user, and the identity of the user can be recognized by the processor 601 according to the fingerprint collected by the fingerprint sensor 614, or by the fingerprint sensor 614 according to the collected fingerprint. When the identity of the user is recognized as a trusted identity, the processor 601 authorizes the user to perform related sensitive operations, including unlocking the screen, viewing encrypted information, downloading software, payment, and changing settings. The fingerprint sensor 614 can be arranged on the front, back, or side of the terminal 600. When the physical button or the manufacturer's logo is arranged on the terminal 600, the fingerprint sensor 614 can be integrated with the physical button or the manufacturer's logo.

[0270] The optical sensor 615 is used to collect the ambient light intensity. In one embodiment, the processor 601 can control the display brightness of the display screen 605 according to the ambient light intensity collected by the optical sensor 615. Specifically, when the ambient light intensity is high, the display brightness of the display screen 605 is increased; when the ambient light intensity is low, the display brightness of the display screen 605 is decreased. In another embodiment, the processor 601 can also dynamically adjust the shooting parameters of the camera assembly 606 according to the ambient light intensity collected by the optical sensor 615.

[0271] The proximity sensor 616, also referred to as a distance sensor, is usually arranged on the front panel of the terminal 600. The proximity sensor 616 is used to collect the distance between the user and the front of the terminal 600. In an embodiment, when the proximity sensor 616 detects that the distance between the user and the front of the terminal 600 gradually decreases, the display screen 605 is switched from the bright screen state to the screen-off state under the control of the processor 601; when the proximity sensor 616 detects that the distance between the user and the front of the terminal 600 gradually increases, the display screen 605 is switched from the screen-off state to the bright screen state under the control of the processor 601.

[0272] Those skilled in the art can understand that the structure shown in the foregoing embodiments is not a limitation on the terminal 600, and the terminal 600 can include more or fewer components than those shown in the drawings, or some components can be combined, or different component arrangements can be adopted. Figure 6

[0273] Figure 7 FIG. 7 is a structural schematic diagram of a server according to an embodiment of the present disclosure. The server 700 can have great differences due to different configurations or performances, and can include one or more central processing units (CPUs) 701 and one or more memories 702. The memory 702 stores at least one instruction, which is loaded and executed by the processor 701 to implement the method provided by the above-mentioned various method embodiments. Of course, the server can also have a wired or wireless network interface, a keyboard, an input and output interface, and other components for implementing device functions, which are not described here.

[0274] In an exemplary embodiment, a computer readable storage medium, such as a memory including instructions, is also provided, which can be executed by a processor in a terminal to complete the live room recommendation method in the above-mentioned embodiments. The computer readable storage medium can be non-transitory. For example, the computer readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.

[0275] Those of ordinary skill in the art can understand that all or part of the steps of the above-mentioned embodiments can be completed by hardware, or by a program instructing related hardware, and the program can be stored in a computer readable storage medium, such as a read-only memory, a magnetic disk or an optical disk.

[0276] ​It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data for analysis, stored data, displayed data, etc.) and signals (including but not limited to signals transmitted between user terminals and other devices, etc.) involved in the present disclosure are all authorized by the user or fully authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions. For example, the "live room recommendation request", "to be exposed amount", "historical live data" and the like involved in the present disclosure are obtained under full authorization.

[0277] The above only describes optional embodiments of the present disclosure and is not intended to limit the present disclosure. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for recommending live streaming rooms, characterized in that, The method includes: Receive live stream recommendation requests sent by the target account; Determine the match between each live stream room and the target account; Identify at least one target new streamer's live stream among multiple live stream rooms; Based on the amount of exposure to be gained for each target new streamer's live stream room, the matching degree between each target new streamer's live stream room and the target account is adjusted to obtain the updated matching degree between each target new streamer's live stream room and the target account. Based on the update matching degree between the target new streamer's live room and the target account, and the matching degree between other live rooms besides the target new streamer's live room and the target account, a live room recommendation queue is determined, and the live room recommendation queue is sent to the target account for recommendation display.

2. The method according to claim 1, characterized in that, The step of identifying at least one target new streamer's live stream room among multiple live stream rooms includes: Identify high-quality new streamers from among multiple new streamers; Identify at least one high-quality new streamer's live stream room that is currently live. Based on the expected exposure and performance score of each high-quality new streamer's live stream, at least one target new streamer's live stream is determined from among the high-quality new streamer's live streams.

3. The method according to claim 2, characterized in that, The process of identifying high-quality new streamers from among multiple new streamers includes: Based on the historical live streaming data of each new streamer, a recommendation score is determined for each new streamer, wherein the historical live streaming data includes at least one of the following: click-through rate, number of days of live streaming, cumulative daily live streaming duration, effective user dwell time, and fan conversion rate. If the new streamer's recommendation score is greater than or equal to the preset score threshold, or if the new streamer passes the manual review, then the new streamer is determined to be a high-quality new streamer.

4. The method according to claim 2, characterized in that, After determining that the new streamer is the high-quality new streamer, the method further includes: Calculate the mean and variance of the first live streaming data of multiple high-quality new anchors, and calculate the difference between the mean and variance, wherein the first live streaming data includes at least one of click-through rate and effective user dwell rate; If the first live broadcast data of the high-quality new anchor is consistently less than the difference within the first latest preset time period, then the exposure amount of the high-quality new anchor will be adjusted downwards and the exposure amount of the high-quality new anchor will be updated.

5. The method according to claim 4, characterized in that, After adjusting the current exposure amount of the high-quality new anchor to decrease to obtain the exposure amount of the high-quality new anchor, the method further includes: Based on the first live broadcast data of each high-quality new anchor, sort them from largest to smallest to obtain the high-quality ranking; For high-quality new anchors whose exposure volume has been reduced, determine the week-on-week decrease value of the first live broadcast data of the high-quality new anchors, and return the high-quality new anchors with a week-on-week decrease value exceeding a preset change threshold from the first preset percentage of high-quality new anchors in the high-quality ranking to the new anchors.

6. The method according to claim 2, characterized in that, The step of determining at least one target new streamer livestream room from among the high-quality new streamer livestream rooms based on the expected exposure and performance score of each high-quality new streamer livestream room includes: Based on the amount of exposure to be gained for each high-quality new streamer's live stream, the recommendation probability corresponding to each high-quality new streamer's live stream is determined, wherein the recommendation probability is positively correlated with the amount of exposure to be gained; The performance scores of each high-quality new streamer's live broadcast room are sorted from largest to smallest to obtain the score ranking. The recommendation probability of the second preset percentage of high-quality new anchor live stream rooms ranked first in the rating ranking is multiplied by the first coefficient, and the recommendation probability of the third preset percentage of high-quality new anchor live stream rooms ranked last in the rating ranking is multiplied by the second coefficient to update the recommendation probability of the high-quality new anchor live stream rooms. Wherein, the first coefficient is greater than 1, and the second coefficient is greater than 0 and less than 1. Based on the recommendation probability of each high-quality new streamer's live stream, probability sampling is performed on multiple high-quality new streamer's live streams to obtain at least one target new streamer's live stream.

7. The method according to claim 2, characterized in that, The method further includes: For each high-quality new streamer, based on the streamer profile data of the high-quality new streamer and the streamer profile data of other streamers besides the new streamer, other streamers that match the high-quality new streamer are determined; Based on the historical exposure of other live streams that match the high-quality new streamer, and the current exposure of the high-quality new streamer's live stream, the amount of exposure to be made for the high-quality new streamer's live stream is determined.

8. The method according to claim 2, characterized in that, The method further includes: Based on the second live streaming data of each high-quality new anchor's live streaming room within the second latest preset time period, a performance score is determined for each high-quality new anchor's live streaming room, wherein the second live streaming data includes at least one of click rate, average stay time per person after entering the room, and instant exit rate after entering the room.

9. The method according to claim 1, characterized in that, The process involves adjusting the matching degree between each target new streamer's live stream room and the target account based on the expected exposure volume of each target new streamer's live stream room, to obtain an updated matching degree between each target new streamer's live stream room and the target account, including: Based on the expected exposure and current broadcast duration of each target new streamer's live stream room, the matching degree between each target new streamer's live stream room and the target account is adjusted to obtain the updated matching degree between each target new streamer's live stream room and the target account.

10. The method according to claim 9, characterized in that, The matching degree between each target new streamer's live stream room and the target account is adjusted based on the expected exposure and current broadcast duration of each target new streamer's live stream room, to obtain an updated matching degree between each target new streamer's live stream room and the target account, including: For each target new streamer's live stream room, obtain the expected exposure amount of the target new streamer's live stream room, the time derivative of the expected exposure amount, and the current broadcast duration; Determine the proportional gain, derivative gain, and time gain, wherein the proportional gain is positively correlated with the amount to be exposed, the derivative gain is negatively correlated with the time derivative of the amount to be exposed, and the time gain is positively correlated with the current broadcast duration; Based on the proportional gain, the derivative gain, and the time gain, the amount to be exposed, the time derivative of the amount to be exposed, and the current broadcast duration are weighted to obtain the weight value corresponding to the target new anchor's live broadcast room. The weight value is multiplied by the matching degree between the target new streamer's live room and the target account to obtain the updated matching degree between the target new streamer's live room and the target account.

11. The method according to claim 1, characterized in that, Determining the matching degree between each live streaming room and the target account includes: Obtain the user behavior sequence features and user profile features corresponding to the target account; For each live streaming room, obtain the offline real-time indicator features and streamer profile features of the corresponding live streaming room, as well as the user-streamer cross features and contextual information features between the live streaming room and the target account; Based on the user behavior sequence features and user profile features corresponding to the target account, the offline real-time indicator features and streamer profile features corresponding to the live room, the user-streamer cross features and contextual information features between the live room and the target account, and the matching degree prediction model, the matching degree between the live price and the target account is determined.

12. A computer device, characterized in that, The computer device includes a processor and a memory, the memory storing at least one instruction, which is loaded and executed by the processor to perform the operations performed by the live streaming recommendation method as described in any one of claims 1 to 11.

13. A computer-readable storage medium, characterized in that, The storage medium stores at least one instruction, which is loaded and executed by a processor to perform the operation of the live streaming recommendation method as described in any one of claims 1 to 11.

14. A computer program product, characterized in that, The computer program product includes at least one instruction, which is loaded and executed by a processor to perform the operation of the live streaming recommendation method as described in any one of claims 1 to 11.

Images