CN115002691A

CN115002691A - Message sending method, device, equipment and computer readable storage medium

Info

Publication number: CN115002691A
Application number: CN202110224724.4A
Authority: CN
Inventors: 刘少金; 黄开莉; 罗小琼
Original assignee: China Mobile Communications Group Co Ltd; China Mobile Group Sichuan Co Ltd
Current assignee: China Mobile Communications Group Co Ltd; China Mobile Group Sichuan Co Ltd
Priority date: 2021-03-01
Filing date: 2021-03-01
Publication date: 2022-09-02
Anticipated expiration: 2041-03-01
Also published as: CN115002691B

Abstract

One embodiment of the present specification provides a message sending method, a message sending device, a message sending apparatus, and a computer-readable storage medium, where the method includes: in the user set, generating a user portrait of each user according to the behavior data of each user; mining a frequent item set of a user tag according to a user portrait of each user, and selecting a target frequent item set containing a target user tag from the frequent item set of the user tag; searching a target user in a user set according to the target frequent item set, wherein the user portrait of the target user comprises any user tag in the target frequent item set; and screening core users from the target users according to the behavior data of the target message sharing behavior among the target users, and sending notification messages to the core users. Through the embodiment, the problems that notification short messages are sent to all users, the number of the sent short messages is too large, resources are wasted, the users are easily disturbed, and the satisfaction degree of the users is reduced can be solved.

Description

Message sending method, device, equipment and computer readable storage medium

Technical Field

The present document relates to the field of mobile communications, and in particular, to a message sending method, apparatus, device, and computer-readable storage medium.

Background

Currently, communication operators often send notification short messages to communication subscribers (hereinafter, simply referred to as subscribers), such as sending a care short message during holidays and sending a notification short message during network failures. In the prior art, a communication operator sends a notification-type short message to each user, however, not every user browses the notification-type short message, and some users may choose to delete or ignore the notification-type short message directly. Therefore, the problem that the number of sent short messages is too large and resources are wasted exists when the notification short messages are sent to all users, and the user satisfaction is reduced because the notification short messages are sent to all users easily to disturb the users.

Disclosure of Invention

An object of one embodiment of the present specification is to provide a message sending method, an apparatus, a device, and a computer-readable storage medium, so as to solve the problems that notification-like short messages are sent to each user, the number of sent short messages is excessive, resources are wasted, and user satisfaction is reduced due to easy disturbance to the user.

To solve the above technical problem, one embodiment of the present specification is implemented as follows:

in a first aspect, an embodiment of the present specification provides a message sending method, including:

generating a user portrait of each user according to the behavior data of each user in a user set; wherein the user representation comprises at least one user tag, each user tag corresponding to a user action;

mining a frequent item set of the user tags according to the user portrait of each user, and selecting a target frequent item set containing target user tags from the frequent item set of the user tags; the target user tag is a user tag corresponding to a target message sharing behavior;

searching a target user in the user set according to the target frequent item set, wherein the user representation of the target user comprises any user tag in the target frequent item set;

and according to the behavior data of the target message sharing behavior among the target users, screening core users relative to the target message sharing behavior among the target users, and sending notification messages aiming at the user set to the core users.

In a second aspect, another embodiment of the present specification provides a message sending apparatus, including:

the portrait generation unit is used for generating a user portrait of each user according to the behavior data of each user in the user set; wherein the user representation comprises at least one user tag, each user tag corresponding to a user action;

the set determining unit is used for mining the frequent item set of the user tags according to the user portrait of each user, and selecting a target frequent item set containing target user tags from the frequent item set of the user tags; the target user tag is a user tag corresponding to a target message sharing behavior;

the user searching unit is used for searching a target user in the user set according to the target frequent item set, and the user portrait of the target user comprises any user tag in the target frequent item set;

and the message sending unit is used for screening core users corresponding to the target message sharing behaviors from the target users according to the behavior data of the target message sharing behaviors among the target users, and sending notification messages aiming at the user set to the core users.

In a third aspect, a further embodiment of the present specification provides a message sending device, which includes a memory and a processor, wherein the memory stores computer-executable instructions, and when the computer-executable instructions are executed on the processor, the steps of the method according to the first aspect can be implemented.

In a fourth aspect, a further embodiment of the present specification provides a computer-readable storage medium having stored thereon computer-executable instructions that, when executed by a processor, are capable of implementing the steps of the method of the first aspect.

In an embodiment of the present specification, in the frequent item set of the user tags, a target frequent item set including a target user tag is selected, where the target user tag is a user tag corresponding to a target message sharing behavior, and therefore, the target frequent item set can represent other user behaviors frequently occurring in the target message sharing behavior. The user portrait of the target user comprises any user tag in the target frequent item set, so that the target user has a target message sharing behavior, or has other behaviors frequently appearing in the target message sharing behavior, and the target user has strong message spreading capacity. The core users are obtained by screening the target users according to the behavior data of the target message sharing behavior among the target users, so that the core users also have strong message transmission capacity, and the number of the core users is less than that of the target users. By sending the notification message to the core user, the effect of sending the notification message to a small part of users with strong message propagation capacity can be achieved, so that the message is propagated to most of the users, and the problems that notification type short messages are sent to all the users, the number of the sent short messages is too large, resources are wasted, and the user satisfaction is reduced due to the fact that the users are easily disturbed are solved.

Drawings

In order to more clearly illustrate the technical solutions in one or more embodiments of the present disclosure, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present disclosure, and for those skilled in the art, other drawings can be obtained according to these drawings without any creative effort.

Fig. 1 is a schematic flowchart of a message sending method according to an embodiment of the present disclosure;

fig. 2 is a schematic combination diagram of user tags a, b, c, and d provided in an embodiment of the present disclosure;

fig. 3 is a schematic block diagram of a message sending apparatus according to an embodiment of the present disclosure;

fig. 4 is a schematic structural diagram of a message sending device according to an embodiment of the present specification.

Detailed Description

In order to make those skilled in the art better understand the technical solutions in one or more embodiments of the present disclosure, the technical solutions in one or more embodiments of the present disclosure will be clearly and completely described below with reference to the drawings in one or more embodiments of the present disclosure, and it is obvious that the described embodiments are only a part of the embodiments of the present disclosure, and not all embodiments. All other embodiments that can be derived by a person skilled in the art from one or more of the embodiments described herein without making any inventive step shall fall within the scope of protection of this document.

Fig. 1 is a flowchart of a message sending method provided in an embodiment of this specification, where the method may be applied to a background server side of a communication carrier and executed by a server of the communication carrier, as shown in fig. 1, the method includes the following steps:

step S102, in a user set, generating a user portrait of each user according to the behavior data of each user; the user portrait comprises at least one user label, and each user label corresponds to one user behavior;

step S104, mining a frequent item set of a user tag according to the user portrait of each user, and selecting a target frequent item set containing a target user tag from the frequent item set of the user tag; the target user tag is a user tag corresponding to the target message sharing behavior;

step S106, searching a target user in a user set according to the target frequent item set, wherein the user portrait of the target user comprises any user tag in the target frequent item set;

step S108, according to behavior data of target message sharing behaviors among the target users, core users corresponding to the target message sharing behaviors are screened from the target users, and notification messages aiming at the user set are sent to the core users.

In an embodiment of the present specification, in the frequent item set of the user tags, a target frequent item set including a target user tag is selected, where the target user tag is a user tag corresponding to a target message sharing behavior, and therefore the target frequent item set can represent other user behaviors frequently occurring with the target message sharing behavior. Because the user portrait of the target user comprises any user tag in the target frequent item set, the target user has a target message sharing behavior, or has other behaviors frequently appearing with the target message sharing behavior, and therefore the target user has a strong message spreading capability. The core users are obtained by screening the target users according to the behavior data of the target message sharing behavior among the target users, so that the core users also have strong message transmission capacity, and the number of the core users is less than that of the target users. By sending the notification message to the core user, the effect of sending the notification message to a small part of users with strong message propagation capacity can be achieved, so that the message is propagated to most of the users, and the problems that notification type short messages are sent to all the users, the number of the sent short messages is too large, resources are wasted, the users are easily disturbed, and the satisfaction degree of the users is reduced are solved.

In this embodiment, when a communication operator needs to send a notification type message, such as a notification type short message, to some users, the users may be combined into a user set. These users may be users located in a geographic area, such as a province, or users handling a communication package, such as a family province.

In step S102, a user figure for each user is generated from the behavior data for each user in the user group. The user representation includes at least one user tag, each user tag corresponding to a user behavior. The method specifically comprises the following actions:

(a1) determining a behavior weight value of each user relative to each preset user behavior according to the behavior data of each user;

(a2) acquiring a first user tag corresponding to each preset user behavior, and determining a tag weight value of each user relative to each first user tag according to a behavior weight value of each user relative to each preset user behavior;

(a3) selecting a corresponding second user label for each user in each first user label according to the label weight value of each user relative to each first user label;

(a4) and generating a user portrait of each user according to the second user label corresponding to each user.

In this embodiment, the process of generating a user representation for each user is the same, and therefore, a user is explained in detail below as an example.

In this embodiment, the preset user behavior includes a message sharing behavior based on a communication package and a message sharing behavior based on an application. The message sharing behavior based on the communication package can comprise a call behavior and a message sending behavior based on the communication package, the communication package can comprise one or more preset communication packages, the call behavior comprises at least one of a voice call behavior and a video call behavior, and the message sending behavior comprises at least one of a short message sending behavior and a multimedia message sending behavior. The application-based message sharing behavior may include application-based voice call behavior, video call behavior, and message sending behavior, the application may include one or more preset applications, and the message sending behavior may send at least one of text, pictures, and voice.

Taking the communication package as a family, and the application includes a WeChat and a QQ as an example, the preset user behavior includes a voice call behavior based on the family, a text messaging behavior based on the family, a voice call behavior based on the WeChat, a video call behavior and a message sending behavior, a voice call behavior based on the QQ, a video call behavior and a message sending behavior. Therefore, the types of the preset user behaviors are increased along with the increase of the number of the preset communication packages and the number of the preset application programs, and in practical application, the preset communication packages and the preset application programs can be determined according to actual requirements. For example, when a core user who receives a message is determined among all users of a communication package, the communication package may be set as a communication package corresponding to a preset user behavior.

In the action (a1), determining a behavior weight value of each user with respect to each preset user behavior according to the behavior data of each user specifically includes:

(a11) determining the behavior frequency and/or the behavior duration of each user relative to each preset user behavior according to the behavior data of each user;

(a12) and determining a behavior weight value of each user relative to each preset user behavior according to the determined behavior frequency and/or behavior duration.

Specifically, in the action (a11), taking the preset user behaviors including the voice call behavior based on the family-happy, the texting behavior based on the family-happy, the voice call behavior based on the WeChat, the video call behavior and the messaging behavior, the voice call behavior based on the QQ, the video call behavior and the messaging behavior as examples, taking a user as an example, according to the behavior data of the user, the behavior frequency and/or the behavior duration of the user relative to each user preset user behavior is determined, that is, the frequency and/or the duration of the voice call based on the family-happy, the frequency and/or the duration of the texting based on the family-happy, the frequency and/or the duration of the voice call based on the WeChat, the video call based on the WeChat, the frequency and/or the duration of the messaging based on the WeChat, Frequency and/or duration of a QQ-based voice call, frequency and/or duration of a QQ-based video call, frequency and/or duration of a QQ-based messaging. The action frequency and/or the action duration referred to in this embodiment may include an action frequency, may also include an action duration, and may also include both an action frequency and an action duration.

In act (a12), a behavior weight value for each user with respect to each of the pre-set user behaviors is determined based on the determined behavior frequency and/or behavior duration.

In the above example, according to the frequency and/or duration of the voice call of the user based on the jukebox, the behavior weight value of the user relative to the voice call based on the jukebox is determined, according to the frequency and/or duration of the short message sent by the user based on the jukebox, the behavior weight value of the user relative to the short message sent by the user based on the jukebox is determined, according to the frequency and/or duration of the voice call of the user based on the micro-letter, the behavior weight value of the user relative to the voice call based on the micro-letter is determined, according to the frequency and/or duration of the video call of the user based on the micro-letter, the behavior weight value of the user relative to the video call based on the micro-letter is determined, and other preset user behaviors are not repeated here.

When the behavior weight value is determined according to the behavior frequency, the result of multiplying the behavior frequency by 100 can be used as the behavior weight value, when the behavior weight value is determined according to the behavior duration, the behavior duration can be used as the behavior weight value, when the behavior weight value is determined according to the behavior frequency and the behavior duration, the behavior frequency can be multiplied by 100 and divided by the first preset value to obtain a first result, the behavior duration is divided by the second preset value to obtain a second result, and the sum of the first result and the second result is used as the behavior weight value. The process of dividing the action frequency by 100 by the first preset value and the process of dividing the action duration by the second preset value can be called normalization.

Returning to the process of generating the user representation, taking a user as an example, after determining a behavior weight value of the user relative to each preset user behavior according to the behavior data of the user, an action (a2) is performed to obtain a first user label corresponding to each preset user behavior, and a label weight value of the user relative to each first user label is determined according to the behavior weight value of the user relative to each preset user behavior.

In this embodiment, a corresponding first user tag is set in advance for each preset user behavior, and table 1 below is a schematic table of the first user tags.

TABLE 1

In this action, after the first user tag corresponding to each preset user behavior is acquired, the behavior weight value of the user relative to each preset user behavior is determined as the tag weight value of the user relative to each first user tag. For example, if the behavior weight value of the user with respect to the voice call behavior based on the jukebox is 5, the tag weight value of the user with respect to the voice of the jukebox is 5, and if the behavior weight value of the user with respect to the text messaging behavior based on the jukebox is 3, the tag weight value of the user with respect to the text messaging behavior of the jukebox is 3.

Next, in the above action (a3), according to the label weight value of each user relative to each first user label, a corresponding second user label is selected for each user in each first user label. Taking a user as an example, according to the label weight value of the user relative to each first user label, selecting a label with the label weight value larger than a preset value from each first user label as a second user label corresponding to the user.

For example, the first user tag corresponding to the user has a family-happy voice, a family-happy short message, a WeChat voice, a WeChat video, and a WeChat message, where the weight values of the tags of the family-happy voice and the family-happy short message are greater than preset values, and then the family-happy voice and the family-happy short message are used as the second user tag corresponding to the user.

It should be noted here that, on the basis of the preset communication package and the preset application program, the preset user behavior is also set, and the first user tag corresponding to each preset user behavior is also set, so that each user has the same first user tag, and the difference is that the tag weight value of each first user tag of each user may be different. Since the label weight value of each first user label of each user may be different, when the second user label is selected according to the preset value, the second user label of each user may be different. Due to the existence of the preset value, there may be a user without the second user tag.

Next, in the above operation (a4), a user image for each user is generated based on the second user tag corresponding to each user. For example, a user may generate a user representation of the user according to the second user tags corresponding to the user, where each second user tag corresponding to the user may be combined into the user representation of the user. For example, a user representation of a user may be "happy-home voice, happy-home short message, WeChat video, WeChat voice, WeChat message".

Returning to the flow in fig. 1, after step S102 is executed, step S104 is executed to mine a frequent set of user tags from the user representation of each user.

Mining a frequent item set of user tags according to a user portrait of each user, which specifically comprises:

(b1) traversing the user portrait of each user, and generating a plurality of user label combinations according to each user label obtained by traversing, wherein each user label combination at least comprises two user labels;

(b2) selecting user label combinations to be calculated from the user label combinations according to the sequence of the number of the user labels included in each user label combination from small to large;

(b3) calculating the support degree and the confidence degree of the selected user label combination by adopting a preset frequent item set mining algorithm according to the user labels included in the user portrait of each user;

(b4) in the calculation process, judging whether to calculate the support degree and the confidence degree of other user label combinations related to the user label combination according to the support degree or the confidence degree of any user label combination; the other user label combination associated with the user label combination is a combination comprising the user labels in the user label combination;

(b5) and according to the calculation result, determining the user label combination with the support degree and the confidence degree meeting the corresponding preset requirements as a frequent item set of the user labels.

First, in act (b1), the user image of each user is traversed to obtain individual user tags, where the traversed user tags refer to the user tags remaining after the duplicate user tags are removed. And then, combining the user tags obtained by traversing to obtain a user tag combination. The combination mode comprises two-two combination, three-three combination, four-four combination and up-to-NN combination, wherein N is the number of the user labels obtained through traversal.

In act (b2), the user tag combinations to be calculated are selected from the user tag combinations in the order of the number of user tags included in each user tag combination from small to large. For example, in the action (b1), three user tags a, b, and c are obtained through traversal, and the user tag combination generated in the action (b1) includes (a, b), (a, c), (b, c), (a, b, c), then in the action (b2), the user tag combinations to be calculated are selected from the user tag combinations in sequence of (a, b), (a, c), (b, c), (a, b, c), where the three user tag combinations (a, b), (a, c), (b, c) all include two user tags, so the selection order of the three user tag combinations may not be fixed as long as the user tag combinations (a, b, c) are preceded by.

In act (b3), a predetermined frequent itemset mining algorithm is used to calculate the support and confidence of the selected user-tag combinations based on the user tags included in the user image of each user. For example, the support and confidence of the user tag combination (a, b) are calculated by Apriori algorithm based on the user tags included in the user representation of each user. When the support degree is calculated, for example, there are 100 users in total and 100 user images in total, and the number of occurrences of the user tag combination (a, b) in the 100 user images is 30, the support degree of the user tag combination (a, b) is 30/100 ═ 0.3. When the confidence is calculated, if the number of occurrences of the user tag a is 60 and the number of occurrences of the user tag b is 90 in the 100 user images, the confidence of the user tag combination (a, b) with respect to the user tag a is 30/60 ═ 0.5, and the confidence of the user tag combination (a, b) with respect to the user tag b is 30/90 ═ 0.33. In the 100 user representations, each user representation includes user tags such as (a, b, c), (b, c, d), (a, b, c, d), etc., where a, b, c, d are user tags.

In the act (b4), in the calculation process, it is determined whether to calculate the support degree and the confidence degree of other user label combinations associated with the user label combination according to the support degree or the confidence degree of any user label combination; the other user tag combinations associated with the user tag combination are combinations that include all of the user tags in the user tag combination. Judging whether to calculate the support degree and the confidence degree of other user label combinations related to any user label combination according to the support degree or the confidence degree of the user label combination, wherein the method comprises the following steps:

(b41) if the support degree or the confidence degree of any user label combination is determined not to meet the corresponding preset requirements, giving up to calculating the support degree and the confidence degree of other user label combinations associated with the user label combination;

(b42) and if the support degree and the confidence degree of any user label combination are determined to meet the corresponding preset requirements, calculating the support degree and the confidence degree of other user label combinations associated with the user label combination.

Specifically, assuming that there are four user tags abcd in a user tag combination (a, b), other user tag combinations associated with the user tag combination (a, b) include (a, b, c), (a, b, d), (a, b, c, d). If the user tag combination (a, b) is a non-frequent item set, then the other user tag combinations (a, b, c), (a, b, d), (a, b, c, d) associated with the user tag combination (a, b) must also be non-frequent item sets, and their support and confidence levels need not be calculated.

Fig. 2 is a combination schematic diagram of user tags a, b, c, and d provided in an embodiment of the present specification, and as shown in fig. 2, the user tags a, b, c, and d are combined two by two, three by three, and four by four to obtain multiple combination situations in fig. 2, if the user tag group (a, b) is a non-frequent item set, then other user tag groups (a, b, c), (a, b, d), (a, b, c, and d) associated with the user tag group (a, b) are also necessarily non-frequent item sets, and their support degrees and confidence degrees do not need to be calculated.

Therefore, in this step, if it is determined that the support degree or the confidence degree of any user tag combination does not satisfy the corresponding preset requirement, for example, the support degree or the confidence degree is not greater than the corresponding preset threshold, it is determined that the any user tag combination is a non-frequent item set, the calculation of the support degree and the confidence degree of the other user tag combinations associated with the user tag combination is abandoned, and if it is determined that the support degree and the confidence degree of any user tag combination both satisfy the corresponding preset requirement, for example, the support degree and the confidence degree are both greater than the corresponding preset threshold, it is determined that the any user tag combination is a frequent item set, and the support degree and the confidence degree of the other user tag combinations associated with the user tag combination are calculated.

In the act (b5), according to the calculation results of the support degree and the confidence degree, the user tags whose support degree and confidence degree both meet the corresponding preset requirements are combined to be determined as a frequent item set of the user tags. For example, the user label combination with the support degree and the confidence degree both greater than the corresponding preset threshold is determined as a frequent item set of the user labels.

Through the above-mentioned action (b4), when the support degree or the confidence degree of any user tag combination does not meet the corresponding preset requirement, the support degree and the confidence degree of other user tag combinations associated with the user tag combination can be abandoned, so that the calculation amount in the process of determining the frequent item set of the user tags is reduced, and the efficiency of determining the frequent item set is improved.

Returning to the flow in fig. 1, in step S104, after mining the frequent item set of the user tags, in the frequent item set of the user tags, a target frequent item set including a target user tag is selected. The target user tag is a user tag corresponding to the target message sharing behavior.

Since the purpose of this embodiment is to propagate the notification message to other users through fewer users, in consideration of the propagation path of the notification message, the target message sharing behavior may be determined according to the propagation path of the notification message, for example, the notification message needs to be propagated in a calling manner, so that the target message sharing behavior may be set as the voice call behavior based on the family happy in the above example. It can be understood that the target message sharing behavior is necessarily located in the preset user behavior described above. Certainly, the target message sharing behavior may include multiple behaviors, and in this case, the process may be executed once for each target message sharing behavior, so as to determine a batch of core users for each target message sharing behavior, and use a set of the batches of core users as the finally determined core users.

After the target message sharing behavior is determined, the target message sharing behavior is located in the preset user behaviors, so that the target message sharing behavior has a corresponding first user tag, and the corresponding first user tag is the target user tag corresponding to the target message sharing behavior. And further, selecting a target frequent item set containing a target user tag from the frequent item sets of the user tags.

Next, in step S106, a target user is searched in the user set according to the target frequent item set, where the user representation of the target user includes any user tag in the target frequent item set. Specifically, all user tags included in the target frequent item set are analyzed, and in the user set, users whose user figures include any one user tag in the target frequent item set are all combined to serve as target users.

Because the target frequent item set is a frequent item set including a target user tag corresponding to the target message sharing behavior, the target frequent item set can represent other user behaviors frequently appearing with the target message sharing behavior. Because the user portrait of the target user comprises any user tag in the target frequent item set, the target user has a target message sharing behavior, or has other behaviors frequently appearing with the target message sharing behavior, and therefore the target user has a strong message spreading capability. Therefore, through the steps S104 and S106, the target user with a strong message dissemination ability can be found in the user set around the target message sharing behavior as a core.

In one embodiment, the notification message may be sent directly to the target user. However, in consideration of the large number of users of the target users, in this embodiment, next, step S108 is executed, and according to the behavior data of the target message sharing behavior among the target users, a core user corresponding to the target message sharing behavior is screened among the target users, and a notification message for the user set is sent to the core user.

In this step, according to behavior data of a target message sharing behavior between target users, a core user corresponding to the target message sharing behavior is screened from the target users, including:

(c1) in the target users, calculating a behavior weight value between every two users according to behavior frequency and/or behavior duration of a target message sharing behavior between every two users;

(c2) clustering target users according to the behavior weight value between every two users, and determining a clustering center user of each type;

(c3) and taking the cluster center user of each class as a core user of the target user relative to the target message sharing behavior.

Taking the target message sharing behavior as the voice call behavior based on happy family as an example in the above example, in the action (c1), among the target users, a behavior weight value between two users is calculated according to a behavior frequency and/or a behavior duration of the target message sharing behavior between the two users, for example, a behavior weight value between two users is calculated according to a behavior frequency based on the voice call behavior based on happy family between the two users, the behavior weight value is equal to the behavior frequency multiplied by 100, as another example, a behavior weight value between two users is calculated according to a behavior duration based on the voice call behavior based on happy family between the two users, the behavior weight value is equal to the behavior duration, as an example, a behavior weight value between two users is calculated according to a behavior frequency and a behavior duration based on the voice call behavior based on happy family between the two users, the behavior weight value is equal to the behavior duration + the behavior frequency multiplied by 100.

In the action (c2), clustering target users according to the behavior weight value between every two users, and determining a clustering center user of each class, wherein the step specifically comprises the following steps:

(c21) determining the distance between every two users according to the behavior weight value between every two users; distance is inversely related to the behavior weight value; the behavior weight value is positively correlated with the behavior frequency and is positively correlated with the behavior duration;

(c22) and clustering the target users by adopting a preset clustering algorithm according to the distance between every two users.

In the act (c21), a distance between two users is determined according to a behavior weight value between the two users, where the distance is negatively correlated with the behavior weight value, and the distance is greater, the behavior weight value is smaller, and the distance may be a reciprocal of the behavior weight value. The behavior weight value is positively correlated with the behavior frequency, and is positively correlated with the behavior duration, the larger the behavior frequency is, the larger the behavior weight value is, the larger the behavior duration is, and the larger the behavior weight value is.

In act (c22), the target users are clustered using a predetermined clustering algorithm, such as knn clustering algorithm, based on the distance between each two users.

After clustering, in acts (c2) and (c3), cluster-center users of each class are determined, and the cluster-center users of each class are taken as core users of the target users who share behaviors with respect to the target message.

Based on the principle of a clustering algorithm, in each class of users obtained by clustering, the distance between the class of users and the clustering center user of the class is shorter, and the distance between the class of users and the clustering center users of other classes is shorter. And taking each type of clustering center user as a core user relative to the target message sharing behavior in the target users, thereby obtaining the core user. In step S108, a notification message for the user set is also sent to the core user. The notification message may be a care type short message such as a festival and holiday greeting short message.

The core users are obtained by screening the target users according to the behavior data of the target message sharing behavior among the target users, so that the core users also have strong message transmission capacity, and the number of the core users is less than that of the target users. By the method in the embodiment, the notification message is sent to the core user, so that the effect of sending the notification message to a small part of users with strong message propagation capacity can be achieved, the message is propagated to most of the users, and the problems that notification type short messages are sent to all the users, the number of the sent short messages is too large, resources are wasted, disturbance is easily caused to the users, and the satisfaction degree of the users is reduced are solved.

Fig. 3 is a schematic diagram of module components of a message sending apparatus according to an embodiment of the present disclosure, and as shown in fig. 3, the apparatus includes:

a representation generating unit 31 for generating a user representation of each user in a user set according to the behavior data of each user; wherein the user representation comprises at least one user tag, each user tag corresponding to a user action;

a set determining unit 32, configured to mine a frequent item set of the user tag according to the user portrait of each user, and select a target frequent item set including a target user tag from the frequent item set of the user tag; the target user tag is a user tag corresponding to a target message sharing behavior;

a user searching unit 33, configured to search a target user from the user set according to the target frequent item set, where a user representation of the target user includes any user tag in the target frequent item set;

a message sending unit 34, configured to screen, from the target users, core users corresponding to the target message sharing behavior according to behavior data of the target message sharing behavior among the target users, and send a notification message for the user set to the core users.

Optionally, the portrait generation unit 31 is specifically configured to:

determining a behavior weight value of each user relative to each preset user behavior according to the behavior data of each user;

acquiring a first user tag corresponding to each preset user behavior, and determining a tag weight value of each user relative to each first user tag according to a behavior weight value of each user relative to each preset user behavior;

selecting a corresponding second user label for each user in each first user label according to the label weight value of each user relative to each first user label;

and generating a user portrait of each user according to the second user label corresponding to each user.

Optionally, the preset user behavior includes a message sharing behavior based on a communication package and a message sharing behavior based on an application program; the portrait generation unit 31 is further specifically configured to:

determining the behavior frequency and/or the behavior duration of each user relative to each preset user behavior according to the behavior data of each user;

and determining a behavior weight value of each user relative to each preset user behavior according to the behavior frequency and/or the behavior duration.

Optionally, the set determining unit 32 is specifically configured to:

traversing the user portrait of each user, and generating a plurality of user label combinations according to each user label obtained by traversing, wherein each user label combination at least comprises two user labels;

selecting user label combinations to be calculated from the user label combinations according to the sequence of the number of the user labels included in each user label combination from small to large;

according to the user tags included in the user portrait of each user, calculating the support degree and the confidence degree of the selected user tag combination by adopting a preset frequent item set mining algorithm;

in the calculation process, judging whether to calculate the support degree and the confidence degree of other user label combinations associated with the user label combination according to the support degree or the confidence degree of any user label combination; the other user label combination associated with the user label combination is a combination comprising the user labels in the user label combination;

and according to the calculation result, determining the user label combination with the support degree and the confidence degree meeting the corresponding preset requirements as a frequent item set of the user labels.

Optionally, the set determining unit 32 is further specifically configured to:

if the support degree or the confidence degree of any user label combination is determined not to meet the corresponding preset requirements, giving up to calculating the support degree and the confidence degree of other user label combinations associated with the user label combination;

and if the support degree and the confidence degree of any user label combination are determined to meet the corresponding preset requirements, calculating the support degree and the confidence degree of other user label combinations associated with the user label combination.

Optionally, the message sending unit 34 is specifically configured to:

in the target users, calculating a behavior weight value between every two users according to behavior frequency and/or behavior duration of a target message sharing behavior between every two users;

clustering the target users according to the behavior weight values between every two users, and determining a clustering center user of each type;

and taking each type of cluster center users as core users of the target users relative to the target message sharing behavior.

Optionally, the message sending unit 34 is further specifically configured to:

determining the distance between every two users according to the behavior weight value between every two users; the distance is inversely related to the behavior weight value; the behavior weight value is positively correlated with the behavior frequency and positively correlated with the behavior duration;

and clustering the target users by adopting a preset clustering algorithm according to the distance between every two users.

The message sending apparatus provided in an embodiment of the present specification can implement each process in the foregoing message sending method embodiment, and achieve the same function and effect, which is not repeated here.

Further, an embodiment of this specification further provides a message sending device, fig. 4 is a schematic structural diagram of the message sending device provided in an embodiment of this specification, and as shown in fig. 4, the device includes: memory 601, processor 602, bus 603, and communication interface 604. The memory 601, processor 602, and communication interface 604 communicate via bus 603. the communication interface 604 may include input and output interfaces including, but not limited to, a keyboard, mouse, display, microphone, and the like.

In fig. 4, the memory 601 stores thereon computer-executable instructions executable on the processor 602, and when executed by the processor 602, the computer-executable instructions implement the following processes:

Optionally, the computer executable instructions, when executed by the processor, generate a user representation for each of the users from the behavioral data for each of the users, including:

Optionally, when the computer-executable instructions are executed by the processor, the preset user behavior includes a message sharing behavior based on a communication package and a message sharing behavior based on an application program;

determining a behavior weight value of each user relative to each preset user behavior according to the behavior data of each user, wherein the behavior weight value comprises the following steps:

Optionally, the computer executable instructions, when executed by the processor, mine a frequent set of items of the user tags from the user representation of each of the users, comprising:

in the calculation process, judging whether to calculate the support degree and the confidence degree of other user label combinations related to the user label combination according to the support degree or the confidence degree of any user label combination; the other user label combination associated with the user label combination is a combination comprising the user labels in the user label combination;

Optionally, when executed by the processor, the determining whether to calculate the support degree and the confidence degree of other user tag combinations associated with any user tag combination according to the support degree or the confidence degree of the user tag combination includes:

Optionally, when executed by the processor, the computer-executable instructions screen, from the target users, core users corresponding to the target message sharing behavior according to behavior data of the target message sharing behavior among the target users, including:

Optionally, when executed by the processor, the computer-executable instructions perform clustering on the target users according to a behavior weight value between two users, including:

In an embodiment of the present specification, in the frequent item set of the user tags, a target frequent item set including a target user tag is selected, where the target user tag is a user tag corresponding to a target message sharing behavior, and therefore the target frequent item set can represent other user behaviors frequently occurring with the target message sharing behavior. The user portrait of the target user comprises any user tag in the target frequent item set, so that the target user has a target message sharing behavior, or has other behaviors frequently appearing in the target message sharing behavior, and the target user has strong message spreading capacity. The core users are obtained by screening the target users according to the behavior data of the target message sharing behavior among the target users, so that the core users also have strong message transmission capacity, and the number of the core users is less than that of the target users. By sending the notification message to the core user, the effect of sending the notification message to a small part of users with strong message propagation capacity can be achieved, so that the message is propagated to most of the users, and the problems that notification type short messages are sent to all the users, the number of the sent short messages is too large, resources are wasted, and the user satisfaction is reduced due to the fact that the users are easily disturbed are solved.

The message sending device provided in an embodiment of this specification can implement each process in the foregoing message sending method embodiment, and achieve the same function and effect, which is not repeated here.

Further, another embodiment of the present specification also provides a computer-readable storage medium for storing computer-executable instructions, which when executed by a processor implement the following process:

and according to behavior data of the target message sharing behavior among the target users, screening core users corresponding to the target message sharing behavior from the target users, and sending notification messages aiming at the user set to the core users.

Optionally, the computer executable instructions, when executed by the processor, generate a user representation for each of the users based on the behavioral data for each of the users, comprising:

Optionally, when the computer-executable instructions are executed by a processor, the preset user behavior includes a message sharing behavior based on a communication package and a message sharing behavior based on an application;

Optionally, the computer executable instructions, when executed by a processor, mine a frequent set of items of the user tags from the user representation of each of the users, comprising:

Optionally, when executed by the processor, the computer-executable instructions determine whether to calculate the support and confidence of other user tag combinations associated with any user tag combination according to the support and confidence of the user tag combination, including:

Optionally, when executed by a processor, the computer-executable instructions screen, among the target users, core users corresponding to a target message sharing behavior according to behavior data of the target message sharing behavior among the target users, and the method includes:

In an embodiment of the present specification, in the frequent item set of the user tags, a target frequent item set including a target user tag is selected, where the target user tag is a user tag corresponding to a target message sharing behavior, and therefore the target frequent item set can represent other user behaviors frequently occurring with the target message sharing behavior. Because the user portrait of the target user comprises any user tag in the target frequent item set, the target user has a target message sharing behavior, or has other behaviors frequently appearing with the target message sharing behavior, and therefore the target user has a strong message spreading capability. The core users are obtained by screening the target users according to the behavior data of the target message sharing behavior among the target users, so that the core users also have strong message transmission capacity, and the number of the core users is less than that of the target users. By sending the notification message to the core user, the effect of sending the notification message to a small part of users with strong message propagation capacity can be achieved, so that the message is propagated to most of the users, and the problems that notification type short messages are sent to all the users, the number of the sent short messages is too large, resources are wasted, and the user satisfaction is reduced due to the fact that the users are easily disturbed are solved.

An embodiment of the present specification provides a computer-readable storage medium, which can implement the processes in the foregoing message sending method embodiment, and achieve the same functions and effects, and will not be repeated here.

The computer-readable storage medium includes a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk.

As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Disks (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

As will be appreciated by one skilled in the art, embodiments of the present description may be provided as a method, system, or computer program product. Accordingly, the description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the description may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The above description is only an example of the present specification, and is not intended to limit the present specification. Various modifications and alterations to this description will become apparent to those skilled in the art. Any modification, equivalent replacement, improvement or the like made within the spirit and principle of the present specification should be included in the scope of the claims.

Claims

1. A method for sending a message, comprising:

generating a user portrait of each user according to the behavior data of each user in a user set; wherein the user representation comprises at least one user tag, each user tag corresponding to a user behavior;

mining a frequent item set of the user tag according to the user portrait of each user, and selecting a target frequent item set containing a target user tag from the frequent item set of the user tag; the target user tag is a user tag corresponding to a target message sharing behavior;

2. The method of claim 1, wherein generating a user representation of each of the users from the behavioral data of each of the users comprises:

3. The method of claim 2, wherein the preset user behavior comprises communication package-based message sharing behavior and application-based message sharing behavior;

4. The method of claim 1, wherein mining a frequent set of items of the user tags from the user representation of each of the users comprises:

according to user tags included in the user portrait of each user, a preset frequent item set mining algorithm is adopted to calculate the support degree and the confidence degree of the selected user tag combination;

5. The method of claim 4, wherein determining whether to calculate the support and confidence of other user tag combinations associated with any user tag combination according to the support or confidence of the user tag combination comprises:

6. The method according to claim 1, wherein the screening, among the target users, core users relative to the target message sharing behavior according to behavior data of the target message sharing behavior among the target users comprises:

7. The method of claim 6, wherein clustering the target users according to behavior weight values between two users comprises:

8. A message transmission apparatus, comprising:

the portrait generation unit is used for generating a user portrait of each user according to the behavior data of each user in a user set; wherein the user representation comprises at least one user tag, each user tag corresponding to a user behavior;

9. A messaging device comprising a memory and a processor, the memory having stored thereon computer-executable instructions that, when executed on the processor, are capable of performing the steps of the method of any of claims 1 to 7.

10. A computer-readable storage medium having computer-executable instructions stored thereon which, when executed by a processor, are capable of performing the steps of the method of any one of claims 1 to 7.