Data pushing method, device and server

Abstract

The present disclosure relates to a data pushing method, device and server, comprising: in response to a received access request, detecting whether an access account sending the access request is an account in a target category, wherein the access request is used to request to obtain platform data; if the access account is an account in the target category, determining platform recommended data from candidate platform data based on an adjustment parameter, wherein the adjustment parameter is used to adjust the proportion of target category data in platform recommended data in the process of determining platform recommended data, and the target category data is data that has a newly added access event after performing an interactive operation on the data; and returning platform recommended data containing the target category data to the access account. The present disclosure can improve the pushing effect of data.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of data processing, and particularly relates to a data pushing method and device and a server. BACKGROUND

[0002] With the development of electronic device technology, more and more application programs can be installed on a terminal device, and the same type of application programs can be installed on a terminal device for a user to select. For example, for a user, multiple video application programs can be installed on the terminal device of the user for the user to select. In this application scenario, a server usually pushes data to an access user of an application program.

[0003] However, at present, the server mainly pushes data to a user according to the behavior habit of the user. In this way, in the case of pushing a large amount of data to some users, the application program can not generate additional access users, or only a small number of access users, thereby resulting in that the pushing effect of the data of the server is poor at present. SUMMARY

[0004] The present disclosure provides a data pushing method, device and server to at least solve the problem that the pushing effect of the data of the server is poor in the related art. The technical solutions of the present disclosure are as follows.

[0005] According to a first aspect of an embodiment of the present disclosure, a data pushing method is provided, comprising:

[0006] In response to the received access request, it is detected whether an access account sending the access request is an account in a target category, wherein the access request is used to request to obtain platform data;

[0007] If the access account is the account in the target category, platform recommended data is determined from candidate platform data based on an adjustment parameter, wherein the adjustment parameter is used to adjust the proportion of target category data in the platform recommended data in the process of determining the platform recommended data, and the target category data is data that has a newly added access event after performing an interaction operation on data;

[0008] The platform recommended data containing the target category data is returned to the access account.

[0009] Optionally, the step of determining the platform recommended data from the candidate platform data based on the adjustment parameter comprises:

[0010] Target category data in the candidate platform data is obtained;

[0011] determining first platform recommendation data from the target category data of the candidate platform data, the first platform recommendation data matching a corresponding proportion of the proportion of the platform recommendation data and the adjustment parameter;

[0012] In the case that the platform recommendation data further includes second platform recommendation data, the second platform recommendation data is determined from data of the candidate platform data other than the target category data.

[0013] Optionally, the step of detecting whether the access account sending the access request is an account in the target classification includes:

[0014] obtaining attribute information of the access account;

[0015] comparing the attribute information of the access account with attribute information corresponding to the target classification to obtain a comparison result;

[0016] determining whether the access account is an account in the target classification based on the comparison result;

[0017] wherein, in the case that the comparison result indicates that the comparison is successful, it is determined that the access account is an account in the target classification, and in the case that the comparison result indicates that the comparison fails, it is determined that the access account is not an account in the target classification.

[0018] Optionally, before the step of returning platform recommendation data containing the target category data to the access account, the method further includes:

[0019] sorting each data in the platform recommendation data; wherein the target category data is located before data other than the target category data in the platform recommendation data;

[0020] The step of returning platform recommendation data containing the target category data to the access account includes:

[0021] returning the sorted platform recommendation data to the access account.

[0022] The detection module is configured to detect whether an access account sending an access request is an account in a target classification in response to the received access request, wherein the access request is used to request to obtain platform data;

[0023] The determining module is configured to determine platform recommendation data from the candidate platform data based on an adjustment parameter if the access account is an account in the target category, wherein the adjustment parameter is used to adjust the proportion of target category data in the platform recommendation data in the process of determining the platform recommendation data, and the target category data is data that has a newly added access event after the interactive operation is performed on the data.

[0024] The pushing module is configured to return the platform recommendation data containing the target category data to the access account.

[0025] Optionally, the determining module comprises:

[0026] The first obtaining unit is configured to obtain target category data in the candidate platform data.

[0027] The first determining unit is configured to determine first platform recommendation data from the target category data of the candidate platform data, and the proportion of the first platform recommendation data in the platform recommendation data matches the corresponding proportion of the adjustment parameter.

[0028] The second determining unit is configured to determine second platform recommendation data from data of the candidate platform data other than the target category data in the case that the platform recommendation data further comprises the second platform recommendation data.

[0029] Optionally, the detecting module comprises:

[0030] The second obtaining unit is configured to obtain attribute information of the access account.

[0031] The comparison unit is configured to compare the attribute information of the access account with attribute information corresponding to the target category to obtain a comparison result.

[0032] The detecting unit is configured to determine whether the access account is an account in the target category based on the comparison result.

[0033] In the case that the comparison result indicates a successful comparison, it is determined that the access account is an account in the target category, and in the case that the comparison result indicates a failed comparison, it is determined that the access account is not an account in the target category.

[0034] Optionally, the apparatus further comprises:

[0035] The sorting module is configured to sort each data in the platform recommendation data, and the target category data is located before data other than the target category data in the platform recommendation data.

[0036] The pushing module is specifically configured to return the ranked platform recommendation data to the access account.

[0037] According to a third aspect of the embodiments of the present disclosure, a server is provided, comprising:

[0038] a processor;

[0039] a memory for storing the processor-executable instructions;

[0040] The processor is configured to execute the instructions to implement the data pushing method in any of the first aspect.

[0041] According to a fourth aspect of the embodiments of the present disclosure, a storage medium is provided, when the instructions in the storage medium are executed by a processor of a server, the server can execute the data pushing method in any of the first aspect.

[0042] According to a fifth aspect of the embodiments of the present disclosure, a computer program product is provided, comprising: executable instructions, when the executable instructions are run on a computer, the computer can execute the data pushing method in any of the first aspect.

[0043] The embodiments of the present disclosure provide at least the following beneficial effects:

[0044] In the case of receiving an access request sent by an account in a target category for requesting to obtain platform data, the platform recommendation data containing target category data is returned to the account in the target category. Since the target category data belongs to the data that has occurred a new access event after performing an interactive operation on the data, pushing the target category data to the account in the target category can attract the interest of the account in the target category, so that the probability of the account in the target category performing an interactive operation on the target category data can be improved, thereby the probability of the server occurring a new access event can be improved, and further the pushing effect of the data can be improved.

[0045] Moreover, in the process of determining the platform recommendation data, the proportion of the target category data in the platform recommendation data can be adjusted based on the adjustment parameter, so that different proportions of the target category data can be pushed according to different accounts in the target category if the adjustment parameter is different, thereby realizing personalized pushing of the data for the accounts in the target category, and further the pushing effect of the data can be further improved.

[0046] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0047] The accompanying drawings, which are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure and serve to explain the principles of the present disclosure, and do not constitute an improper limitation on the present disclosure.

[0048] Figure 1 is a flowchart of a data pushing method according to an exemplary embodiment;

[0049] Figure 2 is a user breaking-out schematic diagram of an application;

[0050] Figure 3 is a user and a first breaking-out coefficient corresponding relationship diagram defined by bucketing;

[0051] Figure 4 is a relationship schematic diagram of a first breaking-out coefficient of each user and an inverse of a target parameter;

[0052] Figure 5 is a block diagram of a data pushing device according to an exemplary embodiment;

[0053] Figure 6 is a block diagram of a server according to an exemplary embodiment. DETAILED DESCRIPTION

[0054] In order to make the ordinary person skilled in the art better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be described clearly and completely below with reference to the drawings.

[0055] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The implementation described in the following exemplary embodiments does not represent all implementations consistent with the present disclosure. On the contrary, they are only examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0056] Figure 1 is a flowchart of a data pushing method according to an exemplary embodiment, as Figure 1 shown, the data pushing method is applied in a server, including the following steps:

[0057] In step S101, in response to a received access request, it is detected whether an access account sending the access request is an account in a target category, wherein the access request is used to request to obtain platform data;

[0058] In step S102, if the access account is an account in the target category, platform recommendation data is determined from the candidate platform data based on an adjustment parameter, wherein the adjustment parameter is used to adjust the proportion of target category data in the platform recommendation data in the process of determining the platform recommendation data, and the target category data is data that has a newly added access event after the interactive operation is performed on the data.

[0059] In step S103, the platform recommendation data containing the target category data is returned to the access account.

[0060] In step S101, the application scenario of the embodiment is first introduced, which is applied to a data pushing system and can include a client and a server. A user can log in to an application program on the client to browse data on the application program. The client can send an access request to the server to request platform data for the user to browse when listening to the login event of the user. Correspondingly, the server can receive the access request of the access account and respond to perform data pushing for the access account and return the platform recommendation data to the access account after determining the platform recommendation data.

[0061] The client receives the platform recommendation data sent by the server and displays the platform recommendation data for the user corresponding to the access account to browse. The platform recommendation data includes target category data, which aims to improve the interactive operation of the user corresponding to the access account to expand new users for the application program.

[0062] The platform data can be multimedia data, i.e., video, audio, or e-book data, which is described in detail below taking video as an example.

[0063] In this step, the access account sending the access request can be an in-application user of the application program, which can be understood as a registered user of the application program, which is a concept opposite to an out-application user of the application program. When the server returns the platform recommendation data to the access account, it needs to consider that the data it pushes can improve the interactive operation (which can include a sharing operation) of the in-application user to provide a data entry for the out-application user of the application program and provide conditions for the out-application user of the application program to be converted into an in-application user of the application program.

[0064] The registered user of the application program can be understood as a user who has registered in the application program, a user who has been real-name authenticated in the application program, or an important user (such as a VIP user) who has been upgraded in the application program.

[0065] The off-site users can include two types, the first type is the same nature user as the in-site user, which can be referred to as off-site homogeneity user, such as the same age stage as the in-site user, or the same occupation as the in-site user, or the same city level as the in-site user, and the like. The second type is the different nature user as the in-site user, which can be referred to as off-site difference user, such as the different age stage as the in-site user, or the different occupation as the in-site user, or the different city level as the in-site user, and the like.

[0066] The purpose of the embodiment is to push data to the in-site user of the application program, so that the in-site user can pull new off-site users according to the pushed data, thereby expanding the users of the application program, breaking the inherent user group of the application program, and realizing the user breaking of the application program.

[0067] Referring to Figure 2 , Figure 2 is a user breaking diagram of the application program, as Figure 2 shown, the meaning of breaking has two types, the first type is to expand the existing circle, that is, to pull the off-site homogeneity user of the application program into the circle of the in-site user of the application program, so as to expand the circle of the in-site user of the application program, which can be referred to as homogeneity user breaking. The second type is to break into another circle, that is, to pull the off-site difference user of the application program into the circle of the in-site user of the application program, so as to expand the circle of the in-site user of the application program, which can be referred to as difference user breaking.

[0068] In the case that the server receives the access request, it can query whether the access account is an in-site user or an off-site user, and in the case that it is determined that the access account is an in-site user, it detects whether the access account is an account in the target classification. The account in the target classification can be referred to as breaking user, which can be understood as a user with a relatively high probability of breaking the existing circle. The platform data is usually not in line with its tone, and the similarity of the videos it likes and the videos liked by off-site users is high, so that its new user pulling ability is strong, that is, the ability of the application program to expand new users is strong.

[0069] In step S102, if it is detected that the access account is a breaking user, the platform recommendation data is determined from the candidate platform data based on the adjustment parameter.

[0070] The adjustment parameter can be referred to as a first breaking-out coefficient. The breaking-out user can evaluate the first breaking-out coefficient, which is used to represent the similarity between the video liked by the in-station user and the video liked by the out-station user group, that is, the first breaking-out coefficient can represent the user acquisition ability of the in-station user. The greater the first breaking-out coefficient, the higher the similarity between the video liked by the in-station user and the video liked by the out-station user group corresponding to the in-station user, and the stronger the user acquisition ability of the in-station user, and the more likely the in-station user acquires the out-station user.

[0071] For example, the in-station user likes video A, and a large proportion of the out-station user group corresponding to the in-station user also likes video A, or the in-station user likes video A very much, and the out-station user group corresponding to the in-station user likes video A to a similar degree as the in-station user, that is, the out-station user group also likes video A very much. It indicates that the first breaking-out coefficient of the in-station user is high, and the user acquisition ability is strong.

[0072] In this step, the candidate platform data can be roughly divided into two categories, the first category is target category data, and the second category is data other than the target category data. The target category data belongs to data that occurs a new access event after performing an interactive operation on the data. When the candidate platform data is a video, the target category data can be referred to as a breaking-out video. The breaking-out video can be understood as a video that has a high similarity between the video liked by the in-station user and the video liked by the out-station user.

[0073] For example, the in-station user likes video A, video B and video C, and the out-station user group likes video A, video D and video E, and the breaking-out video is video A. That is, the breaking-out video is a video liked by both the in-station user and the out-station user group. Therefore, after the in-station user performs an interactive operation (which can include a sharing operation) on the breaking-out video, it can usually arouse the interest of the out-station user, thereby guiding the out-station user to register as an in-station user of the application program, and the server can occur a new access event.

[0074] The adjustment parameter is used to adjust the proportion of the target category data in the platform recommendation data in the process of determining the platform recommendation data, that is, the proportion of the target category data in the platform recommendation data corresponds to the first breaking-out coefficient, that is, in the case of a large first breaking-out coefficient, the number of breaking-out videos in the platform recommendation data is correspondingly more, and in the case of a relatively small first breaking-out coefficient, the number of breaking-out videos in the platform recommendation data is correspondingly less.

[0075] For example, if the first break-out coefficient of an access account is 1.04, the number of break-out videos in the platform recommendation data is 8, and the first break-out coefficient of another access account is 1, and the number of break-out videos in the platform recommendation data is 5, the number of break-out videos recommended to the access account is less than that of the access account with the first break-out coefficient of 1.04.

[0076] In step S103, the platform recommendation data containing the target category data is returned to the access account. The platform recommendation data at least includes break-out videos. Of course, if the proportion of break-out videos in the platform recommendation data is not 1, the platform recommendation data also includes second data.

[0077] When returning the platform recommendation data to the access account, the platform recommendation data can be returned in a random order or sorted and returned. Herein, no specific limitation is made.

[0078] In this embodiment, since the content of the platform is usually not in line with the preferences of the break-out users, the push of break-out videos to the break-out users can improve the click intention of the break-out users for the break-out videos and improve the sharing intention. In addition, for the break-out users, the videos they like are highly similar to the videos liked by the off-site user group. Therefore, if the break-out videos are shared to the off-site, the off-site user group will also be interested, thereby guiding the off-site users to register as in-site users of the application program, and realizing the break-out of the users of the application program. In this way, the use of the application program can be expanded from the recommendation strategy of the server, and the promotion cost of the application program is reduced. Therefore, the push effect of the server on the data is good in this embodiment.

[0079] Optionally, based on the first embodiment, the step S102 specifically includes:

[0080] obtaining target category data in the candidate platform data;

[0081] determining first platform recommendation data from the target category data of the candidate platform data, and the proportion of the first platform recommendation data in the platform recommendation data matches the corresponding proportion of the adjustment parameter;

[0082] In the case that the platform recommendation data also includes second platform recommendation data, the second platform recommendation data is determined from the data of the candidate platform data other than the target category data.

[0083] In actual application, the breaking circle video can be evaluated by a second breaking circle coefficient, the second breaking circle coefficient can represent the new user attracting efficiency of the video, that is, the second breaking circle coefficient can be understood as the probability that the video can attract new users outside the station after being shared to the outside of the station. In the case that the shared video is highly similar to the videos liked by the outside user group, the probability of attracting new users outside the station is usually high, and the second breaking circle coefficient is also relatively large. Therefore, the second breaking circle coefficient can also be used to represent the similarity between the video liked by the user inside the station and the video liked by the user outside the station.

[0084] The second breaking circle coefficient of the video can be predicted, specifically, the second breaking circle coefficient of the existing video on the application program can be predicted based on a pre-trained prediction model.

[0085] The prediction model can be trained based on a plurality of training sample videos and label data of the plurality of training sample videos. The plurality of training sample videos are videos shared from the application program within a preset time period. The label data of a first training sample video is data established based on the number of new devices attracted by a single sharing of the first training sample video. The first training sample video is any training sample video in the plurality of training sample videos.

[0086] The preset time period can be set according to actual conditions, for example, three months before the current time. The plurality of training sample videos can be videos shared from the application program within the last three months and having a sharing frequency greater than or equal to a preset frequency, for example, 200 times. The number of new devices attracted by a single sharing of the shared video can be calculated, and the number of new devices attracted by a single sharing = the number of new devices attracted / the sharing frequency.

[0087] The label data can be obtained by establishing labels for the training sample videos based on the number of new devices attracted by a single sharing of the training sample videos. The training sample videos can include positive samples and negative samples. The positive sample can be defined as a training sample video with a number of new devices attracted by a single sharing greater than 0.01, and the negative sample can be defined as a training sample video with a number of new devices attracted by a single sharing less than 0.005. Accordingly, the label data of the training sample video of the positive sample can be set to 1, and the label data of the training sample video of the negative sample can be set to 0.

[0088] After the preset model is trained based on the plurality of training sample videos and the label data of the plurality of training sample videos, the preset model can be used to predict the second breakout coefficient of other videos on the application. Specifically, the data on the application is input into the preset model, and the preset model can analyze the inherent characteristics of the video (including video length and video text area, etc.), author attributes (including number, gender, age, and geographical location distribution, etc.), and video operation indicators (including like rate, follow rate, and long-playing rate, etc.), and output the second breakout coefficient, which can be any value between 0 and 1.

[0089] Here, the single sharing new device number of the video shared by the application in the form of advertisement placement is counted, and the video is labeled based on the single sharing new device number, so that the shared video can be effectively quantified, and then the prediction model can be trained, so that the prediction of the second breakout coefficient of the video on the application based on the prediction model is more interpretable.

[0090] Then, the breakout video in the candidate platform data can be obtained based on the second breakout coefficient of the video. Since the breakout video is a video with high similarity between the video liked by the user in the station and the video liked by the user group outside the station, the breakout video can be a video with a second breakout coefficient greater than a preset threshold on the application, and the preset threshold is usually large, such as any value greater than or equal to 0.5.

[0091] The server can store the second breakout coefficient of each video in the candidate platform data, and mark the video with a second breakout coefficient greater than a preset threshold as a breakout video, and accordingly, obtain the breakout video in the candidate platform data.

[0092] After obtaining the breakout video in the candidate platform data, the first platform recommendation data can be determined from the target category data of the candidate platform data, and the proportion of the first platform recommendation data in the platform recommendation data corresponds to the matching proportion of the adjustment parameter.

[0093] For example, if the number of platform recommendation data is 10, and the adjustment parameter corresponds to a proportion of 0.8, i.e., the number of first platform recommendation data is 8, then 8 are selected from the breakout videos in the candidate platform data, wherein the 8 can be randomly selected, or the breakout videos with the first breakout coefficient can be selected, or other strategies can be used to select 8 breakout videos, which are not limited here.

[0094] In a case where the proportion corresponding to the adjustment parameter is less than 1, that is, in a case where the proportion of the breakout video in the platform recommendation data is less than 1, it is indicated that the platform recommendation data further includes second platform recommendation data, such as the example described above, the platform recommendation data further includes two other videos. At this time, two videos are selected from the videos in the candidate platform data except the breakout video.

[0095] In this embodiment, the breakout video in the candidate platform data is obtained, and the first platform recommendation data and the second platform recommendation data are obtained from the breakout video and the video except the breakout video in the candidate platform data based on the adjustment parameter, so that personalized data pushing can be performed for the access account, and the pushing effect of the video is improved.

[0096] Optionally, based on the first embodiment, the step S101 specifically includes:

[0097] obtaining attribute information of the access account;

[0098] comparing the attribute information of the access account with attribute information corresponding to the target category to obtain a comparison result;

[0099] determining whether the access account is an account in the target category based on the comparison result;

[0100] In a case where the comparison result indicates that the comparison is successful, it is determined that the access account is an account in the target category, and in a case where the comparison result indicates that the comparison fails, it is determined that the access account is not an account in the target category.

[0101] The server can pre-obtain and store the attribute information of the breakout user, and before obtaining the attribute information of the breakout user, it is necessary to first determine which attributes of the user belong to the breakout user, which is described in detail below.

[0102] Specifically, the attribute information of the user can be defined from different levels. For example, it is defined from different dimensions, as shown in Table 1, which is a table of correspondence between users defined from different dimensions and first breakout coefficients. As shown in Table 1 below, the user can be defined from different dimensions, such as the north-south dimension, the city level dimension, the age dimension, and the gender dimension. From the north-south dimension, the user can be divided into northern users and southern users, and from the city level dimension, the user can be divided into users in first-tier and new first-tier cities, users in second- and third-tier cities, and users in fourth- and fifth-tier cities. Each attribute of the user can correspond to a first breakout coefficient.

[0103] Table 1: Correspondence between users defined from different dimensions and first breakout coefficients

[0104]

[0105] For example, the users are defined by buckets, see Figure 3 , Figure 3 is a user corresponding relationship diagram of the first breaking circle coefficient defined by buckets, as shown in Figure 3 Users can be defined from different buckets. If the above four dimensions are considered, the users can be divided into 36 buckets, and each bucket represents a user with a certain attribute, such as a user with attributes of being from the south, a user from a four or five-line city, a user of more than 40 years old, a male user, a user with attributes of being from the north, a user from a first or new first-line city, a female user between the ages of 13 and 22, and so on.

[0106] Each attribute of the user can correspond to a first breaking circle coefficient. The first breaking circle coefficient can be a normalized breaking circle coefficient. The breaking circle coefficient of a user with a certain attribute can be divided by the average value of the breaking circle coefficients of users with various attributes to obtain the first breaking circle coefficient of the user with the certain attribute. According to the size relationship between the first breaking circle coefficient of the user with the certain attribute and the average value, the value can be greater than 1 or less than 1.

[0107] It should be noted that the above attribute division is only an example. For different dimensions, other trends can also be used for division, such as, for the age dimension, the users can be divided into other three attributes, i.e., 12 to 23 years old, 24 to 40 years old, and more than 41 years old. Of course, for different dimensions, the users can also be divided into more attributes or fewer attributes, such as, for the city level dimension, the users can be divided into five attributes, i.e., first or new first-line, second-line, third-line, fourth-line, and fifth-line.

[0108] For each attribute of the user, the first breaking circle coefficient can be calculated in the following manner. Specifically, the first click rate of the user with the certain attribute for the breaking circle video can be obtained, and the second click rate of the user with the certain attribute for the existing video on the platform can be obtained. The first click rate divided by the second click rate can be calculated to obtain the first breaking circle coefficient of the user with the certain attribute.

[0109] In order to verify that the first breaking circle coefficient obtained by calculation conforms to the cognition and can well represent the similarity between the videos liked by the in-site users and the videos liked by the out-site user groups, offline verification can be performed to accurately determine the breaking circle users according to the first breaking circle coefficient. Specifically, the first breaking circle coefficient is negatively correlated with a target parameter corresponding to the user, and the target parameter is calculated based on the retention rate and the penetration rate of the user.

[0110] See Figure 4 , Figure 4 is a relationship diagram of the first breaking circle coefficient of each user and the inverse of the target parameter, as shown in Figure 4As shown, taking 12 buckets of users, i.e., n = 12, as an example, the greater the first breakout coefficient of a user, the greater the reciprocal of the target parameter, and the smaller the first breakout coefficient of a user, the smaller the reciprocal of the target parameter. It can be seen that the rank correlation coefficient of the first breakout coefficient of each user and the reciprocal of the target parameter is relatively high. After calculation, the Spearman rank correlation coefficient of the first breakout coefficient of each user and the reciprocal of the target parameter is 0.9. According to the critical value table of the Spearman rank correlation coefficient test, when n = 12, the critical value of the significant level of 0.01 is 0.7, and therefore, it can be proved that the Spearman rank correlation coefficient is relatively consistent with cognition.

[0111] If the purpose of breaking out homogeneous users and breaking out different users is to be achieved at the same time, the rank correlation of the first breakout coefficient of each user and the reciprocal of the target parameter should be high, i.e., the Spearman rank correlation coefficient is large, that is, the ordering consistency of the first breakout coefficient of each user and the reciprocal of the target parameter should be high. For example, the ordering from high to low of the first breakout coefficient is user A, user B, user C, user D and user E, and the ordering from high to low of the reciprocal of the target parameter is user A, user C, user B, user D and user E, and the ordering consistency is high, and therefore, the first breakout coefficient of each user is relatively reasonable.

[0112] The target parameter is calculated by multiplying the retention rate by the penetration rate, and the first breakout coefficient is negatively correlated with the target parameter. For example, a male user with the attribute of being in the south and over 40 years old has a low retention rate and a low penetration rate for the application program, and the first breakout coefficient of the user should be relatively large; a female user with the attribute of being in the north and between 13 and 22 years old has a high retention rate and a high penetration rate, and the first breakout coefficient of the user should be relatively small.

[0113] After the above verification, it can be concluded that the first breakout coefficient of each user is relatively consistent with cognition, and therefore, the breakout user can be accurately determined according to the first breakout coefficient.

[0114] Further, as described in Table 1 above, in terms of dimensions, the main characteristics of the user with a large first breakout coefficient, i.e., the breakout user, are as follows: biased to the south, biased to the first and new first line, biased to the older age group, and biased to women, that is, these users prefer videos of these attributes to videos preferred by the off-site user group, and are more willing to click on breakout videos. In terms of buckets, the main characteristics of the breakout user are as follows: users in the high age group, the content of the platform is not consistent with their tone, and they have a high willingness to click on breakout videos; female users in the north and in the low age group, the content of the platform is consistent with their tone, and they have a low willingness to click on breakout videos.

[0115] After the server acquires and stores the attribute information of the out-of-circle user, when an access request for an access account is received, the attribute information of the access account can be acquired, and the attribute information of the access account is compared with the attribute information of the out-of-circle user to determine whether the access account is an out-of-circle user. For example, the stored attribute information of the out-of-circle user is high age, and if the attribute information of the access account is high age, it indicates that the access account is an out-of-circle user.

[0116] If the access account is an out-of-circle user, some out-of-circle videos are pushed to the access account, so that the data pushing effect can be improved. If the access account is not an out-of-circle user, the recommendation strategy is normal.

[0117] Optionally, based on embodiment one, before the step S103, the method further includes:

[0118] The data in the platform recommendation data are sorted, and the target category data is located before the data other than the target category data in the platform recommendation data.

[0119] The step S103 specifically includes:

[0120] The sorted platform recommendation data are returned to the access account.

[0121] In this embodiment, in order to improve the probability of the interactive operation of the access account corresponding user on the platform recommendation data, before pushing, the data in the platform recommendation data can be sorted, and the out-of-circle video in the platform recommendation data can be given priority to improve the sorting position of the out-of-circle video in the platform recommendation data, so that the out-of-circle video is sorted before other videos. In addition, different out-of-circle videos can be sorted according to the second out-of-circle coefficient, so that the greater the second out-of-circle coefficient, the higher the sorting. In this way, the out-of-circle video displayed on the client can be more eye-catching, so that the probability of the interactive operation of the access account on the out-of-circle video can be further improved, and the possibility of the user breaking the circle of the application program can be further improved.

[0122] Figure 5 is a block diagram of a data pushing device according to an example embodiment. Referring to Figure 5 The device includes a detection module 501, a determination module 502 and a pushing module 503.

[0123] The detection module 501 is configured to detect whether the access account sending the access request is an account in the target category in response to the received access request, wherein the access request is used to request to acquire platform data;

[0124] The determining module 502 is configured to determine platform recommendation data from the candidate platform data based on an adjustment parameter if the access account is an account in the target category, wherein the adjustment parameter is used to adjust the proportion of target category data in the platform recommendation data in the process of determining the platform recommendation data, and the target category data is data that has a newly added access event after performing an interactive operation on the data.

[0125] The pushing module 503 is configured to return the platform recommendation data containing the target category data to the access account.

[0126] Optionally, the determining module 502 includes:

[0127] The first obtaining unit is configured to obtain target category data in the candidate platform data.

[0128] The first determining unit is configured to determine first platform recommendation data from the target category data of the candidate platform data, and the proportion of the first platform recommendation data in the platform recommendation data matches the corresponding proportion of the adjustment parameter.

[0129] The second determining unit is configured to determine second platform recommendation data from data of the candidate platform data other than the target category data in the case that the platform recommendation data further includes the second platform recommendation data.

[0130] Optionally, the detecting module 501 includes:

[0131] The second obtaining unit is configured to obtain attribute information of the access account.

[0132] The comparison unit is configured to compare the attribute information of the access account with attribute information corresponding to the target category to obtain a comparison result.

[0133] The detecting unit is configured to determine whether the access account is an account in the target category based on the comparison result.

[0134] In the case that the comparison result indicates a successful comparison, it is determined that the access account is an account in the target category, and in the case that the comparison result indicates a failed comparison, it is determined that the access account is not an account in the target category.

[0135] Optionally, the apparatus further includes:

[0136] The sorting module is configured to sort each data in the platform recommendation data, and the target category data is located before data other than the target category data in the platform recommendation data.

[0137] The pushing module 503 is specifically configured to return the ranked platform recommendation data to the access account.

[0138] With regard to the apparatus in the above-mentioned embodiments, the specific manner in which each module performs operations has been described in detail in the embodiments of the method, and thus will not be described in detail here.

[0139] Figure 6 is a block diagram of a server according to an exemplary embodiment, including a processing component 601, which further includes one or more processors, and memory resources represented by memory 602, for storing instructions, such as application programs, executable by the processing component 601. The application programs stored in the memory 602 can include one or more than one module each corresponding to a set of instructions. In addition, the processing component 601 is configured to execute instructions to perform the following processes:

[0140] In response to the received access request, detecting whether the access account sending the access request is an account in the target category, wherein the access request is used to request to obtain platform data;

[0141] If the access account is an account in the target category, determining platform recommendation data from the candidate platform data based on an adjustment parameter, wherein the adjustment parameter is used to adjust the proportion of target category data in the platform recommendation data in the process of determining the platform recommendation data, the target category data being data that has occurred a new access event after performing an interactive operation on the data;

[0142] Returning the platform recommendation data containing the target category data to the access account.

[0143] Optionally, the processing component 601 is further configured to perform:

[0144] Obtaining target category data in the candidate platform data;

[0145] Determining first platform recommendation data from the target category data of the candidate platform data, the proportion of the first platform recommendation data in the platform recommendation data corresponding to the matching proportion of the adjustment parameter;

[0146] In the case where the platform recommendation data further includes second platform recommendation data, determining the second platform recommendation data from the data of the candidate platform data other than the target category data.

[0147] Optionally, the processing component 601 is further configured to perform:

[0148] Obtaining attribute information of the access account;

[0149] comparing the attribute information of the access account with the attribute information corresponding to the target category, to obtain a comparison result;

[0150] determining whether the access account is an account in the target category based on the comparison result;

[0151] wherein, in a case where the comparison result indicates a successful comparison, it is determined that the access account is an account in the target category, and in a case where the comparison result indicates a failed comparison, it is determined that the access account is not an account in the target category.

[0152] Optionally, the processing component 601 is further configured to perform:

[0153] sorting each data in the platform recommendation data; wherein the target category data is located before data other than the target category data in the platform recommendation data;

[0154] returning the sorted platform recommendation data to the access account.

[0155] The server 600 can further include a power supply component 603 configured to perform power management of the server 600, a wired or wireless network interface 604 configured to connect the server 600 to a network, and an input / output (I / O) interface 605. The server 600 can operate based on an operating system stored in the memory 602, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

[0156] In an exemplary embodiment, a storage medium including instructions, such as the memory 602 including instructions, is also provided, and the above-mentioned instructions can be executed by the processing component 601 of the server 600 to complete the above-mentioned method. Optionally, the storage medium can be a non-transitory computer readable storage medium, such as a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, and an optical data storage device, etc.

[0157] Other embodiments of the present disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that are deemed to fall within the general principles of the present disclosure and include commonly known or customary practice in the art. The specification and examples are to be considered exemplary only, with the true scope and spirit of the present disclosure being indicated by the following claims.

[0158] It should be understood that the present disclosure is not limited to the precise construction that has been described above and shown in the accompanying drawings, and that various modifications and changes can be made by those skilled in the art without departing from the scope of the present disclosure. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A data push method, characterized by, The method comprises: In response to the received access request, detecting whether the access account sending the access request is an account in a target classification, wherein the access request is used to request to obtain platform data, and wherein the account in the target classification is a user breaking the circle with a user acquisition capability, i.e., the account can trigger an off-site user to generate a new access event after performing an interactive operation on data; If the access account is an account in the target classification, platform recommendation data is determined from candidate platform data based on an adjustment parameter, wherein the adjustment parameter is a first breaking circle coefficient, and the first breaking circle coefficient represents the user acquisition capability of the access account; the adjustment parameter is used to adjust the proportion of target category data in the platform recommendation data in the process of determining the platform recommendation data, the target category data belongs to data that generates a new access event after performing an interactive operation on data, and the target category data is evaluated by a second breaking circle coefficient, and the second breaking circle coefficient is used to represent the user acquisition efficiency; The platform recommendation data containing the target category data is returned to the access account.

2. The data push method of claim 1, wherein, The step of determining the platform recommendation data from the candidate platform data based on the adjustment parameter comprises: Obtaining target category data in the candidate platform data; Determining first platform recommendation data from the target category data of the candidate platform data, wherein the proportion of the first platform recommendation data in the platform recommendation data matches the corresponding proportion of the adjustment parameter; In the case that the platform recommendation data further comprises second platform recommendation data, the second platform recommendation data is determined from data other than the target category data in the candidate platform data.

3. The data push method of claim 1, wherein, The step of detecting whether the access account sending the access request is an account in a target classification comprises: Obtaining attribute information of the access account; Comparing the attribute information of the access account with attribute information corresponding to the target classification to obtain a comparison result; Based on the comparison result, determining whether the access account is an account in the target classification; If the comparison result indicates that the comparison is successful, it is determined that the access account is an account in the target classification, and if the comparison result indicates that the comparison fails, it is determined that the access account is not an account in the target classification.

4. The data push method of claim 1, wherein, Before the step of returning the platform recommendation data containing the target category data to the access account, the method further comprises: Sorting each data in the platform recommendation data; wherein the target category data is located before data other than the target category data in the platform recommendation data; The step of returning the platform recommendation data containing the target category data to the access account comprises: Returning the sorted platform recommendation data to the access account.

5. A data push apparatus, characterized by comprising: The method comprises: The detection module is configured to detect, in response to the received access request, whether an access account sending the access request is an account in a target classification, where the access request is used to request to obtain platform data, and the account in the target classification is a breakout user with a user acquisition capability, that is, the account can trigger an off-site user to generate a new access event after performing an interactive operation on data. The determination module is configured to determine platform recommendation data from candidate platform data based on an adjustment parameter if the access account is the account in the target classification, where the adjustment parameter is a first breakout coefficient representing the user acquisition capability of the access account, and the adjustment parameter is used to adjust a proportion of target category data in the platform recommendation data in the process of determining the platform recommendation data, the target category data belongs to data that generates a new access event after performing an interactive operation on data, and the target category data is evaluated by a second breakout coefficient representing a user acquisition efficiency. The pushing module is configured to return the platform recommendation data containing the target category data to the access account.

6. The data push apparatus of claim 5, wherein The determination module includes: A first acquisition unit configured to acquire target category data in candidate platform data. A first determination unit configured to determine first platform recommendation data from the target category data of the candidate platform data, where a proportion of the first platform recommendation data in the platform recommendation data corresponds to a matching proportion of the adjustment parameter. A second determination unit configured to determine second platform recommendation data from data of the candidate platform data other than the target category data in the case that the platform recommendation data further includes the second platform recommendation data.

7. The data push apparatus of claim 5, wherein The detection module includes: A second acquisition unit configured to acquire attribute information of the access account. A comparison unit configured to compare the attribute information of the access account with attribute information corresponding to the target classification to obtain a comparison result. A detection unit configured to determine whether the access account is an account in the target classification based on the comparison result. Wherein, in the case that the comparison result indicates a successful comparison, it is determined that the access account is an account in the target classification, and in the case that the comparison result indicates a failed comparison, it is determined that the access account is not an account in the target classification.

8. The data push apparatus of claim 5, wherein, The device further includes: An ordering module configured to order each data in the platform recommendation data, where the target category data is located before data other than the target category data in the platform recommendation data. The pushing module is specifically configured to return the ordered platform recommendation data to the access account.

9. A server, characterized by It includes: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is configured to execute the instructions to implement the data pushing method according to any one of claims 1 to 4.

10. A storage medium, when instructions in the storage medium are executed by a processor of a server, enables the server to perform the data pushing method of any one of claims 1 to 4.

Images