WO2019019711A1

WO2019019711A1 - Method and apparatus for publishing behaviour pattern data, terminal device and medium

Info

Publication number: WO2019019711A1
Application number: PCT/CN2018/083551
Authority: WO
Inventors: 王健宗; 吴天博; 黄章成; 肖京
Original assignee: 平安科技（深圳）有限公司
Priority date: 2017-07-24
Filing date: 2018-04-18
Publication date: 2019-01-31
Also published as: CN107798249B; CN107798249A

Abstract

Provided are a method and apparatus for publishing behaviour pattern data, a terminal device and a medium, which are suitable for the field of data processing. The method comprises: acquiring social media data published by a plurality of users; establishing a behaviour pattern distribution function corresponding to the social media data; integrating a pre-set random noise function and the behaviour pattern distribution function, so as to obtain a noise-added behaviour pattern distribution function; inputting behaviour pattern data query parameters into the noise-added behaviour pattern distribution function, so as to obtain noise-added behaviour pattern data; and publishing the noise-added behaviour pattern data. According to the solution, by integrating a random noise function with a behaviour pattern distribution function, all behaviour pattern data published on the basis of a noise-added behaviour pattern distribution function have noise, and thus the privacy security of users is enhanced. At the same time, the noise-added behaviour pattern data still retain the reference value of original behaviour pattern data, and thus it is ensured that scientific research personnel can carry out effective analysis and mining based on the noise-added behaviour pattern data.

Description

Method, device, terminal device and medium for distributing behavior pattern data

This application claims the priority of the Chinese Patent Application filed on July 24, 2017, the Chinese Patent Office, the application number is 201710605631.X, and the invention name is "the release method of the behavioral model data and the terminal device", the entire contents of which are incorporated by reference. In this application.

Technical field

The present application belongs to the field of data processing, and in particular, to a method, device, terminal device and medium for distributing behavior mode data.

Background technique

The development of social media makes the user's personal privacy more and more accessible to others. Specifically, since the social media data includes not only the effective information published by the user but also the relationship between the user and the user, the attacker can mine the effective information posted by the user and the mutual users. The relationship matches the user's identity in the social network, and thus can obtain the user's private information in all directions. However, if these social media data are provided to the right researchers, they can use data mining analysis methods to extract information that has a driving value for social progress.

In the prior art, in order to protect the privacy of the user, it is generally required to encrypt the user behavior mode data corresponding to the social media data. In this way, it is difficult for researchers to analyze user behavior patterns based on encrypted data. Therefore, there is a mutually exclusive relationship between the protection of personal privacy in social media data and the analysis and utilization of social media data to promote social progress. In the prior art, it is impossible to release the behavior mode data with mining value while protecting the privacy of the user.

technical problem

In view of this, the embodiment of the present application provides a method, an apparatus, a terminal device, and a medium for distributing behavior pattern data, so as to solve the problem that the behavior mode data of the mining value cannot be released while protecting the privacy of the user in the prior art. The problem.

Technical solution

A first aspect of the embodiments of the present application provides a method for publishing behavior mode data, including:

Get social media data published by multiple users;

Establishing a behavior pattern distribution function corresponding to the social media data;

Integrating the preset random noise function with the behavior pattern distribution function to obtain a noise-added behavior pattern distribution function;

Entering the behavior mode data query parameter into the noise-added behavior pattern distribution function to obtain the noise-added behavior pattern data;

The noise-added behavior pattern data is released.

A second aspect of the embodiments of the present application provides a device for distributing behavior pattern data, the apparatus for distributing behavior pattern data comprising means for performing a method for distributing behavior pattern data according to the above first aspect.

A third aspect of the embodiments of the present application provides a terminal device, including a memory and a processor, where the computer stores computer readable instructions executable on the processor, the processor executing the computer The step of the method of issuing the behavior pattern data as described in the first aspect is implemented when the instruction is read.

A fourth aspect of the embodiments of the present application provides a computer readable storage medium storing computer readable instructions, the computer readable instructions being executed by a processor to implement the first aspect as described in the first aspect The steps of the method of publishing behavioral data.

Beneficial effect

In the embodiment of the present application, by establishing a behavior pattern distribution function corresponding to the social media data, the behavior pattern data that are not related to each other can be summarized; by integrating the random noise function into the behavior pattern distribution function, the pattern distribution function based on the noise enhancement behavior is obtained. The published behavioral pattern data is noisy, so even if the attacker steals the released noisy behavior pattern data, it still cannot accurately match the noisy behavior pattern data to each user, thus strengthening the user's individual. Privacy security. At the same time, the behavior mode data after the noise preservation still retains the reference value of the original behavior pattern data, thus ensuring that the scientific research personnel can perform effective analysis and mining based on the noise-added behavior pattern data.

DRAWINGS

1 is a flowchart of an implementation of a method for distributing behavior pattern data provided by an embodiment of the present application;

2 is a specific implementation flowchart of a method for publishing behavior mode data provided by an embodiment of the present application;

FIG. 3 is a specific implementation flowchart of a method for publishing behavior pattern data according to an embodiment of the present application;

4 is a structural block diagram of an apparatus for distributing behavior pattern data according to an embodiment of the present application;

FIG. 5 is a schematic structural diagram of a terminal device according to an embodiment of the present application.

Embodiments of the invention

In order to explain the technical solutions described in the present application, the following description will be made by way of specific embodiments.

FIG. 1 is a flowchart showing an implementation process of a method for distributing behavior pattern data provided by an embodiment of the present application, which is described in detail as follows:

S101: Acquire social media data published by multiple users.

In the embodiment of the present application, all data published by the user in a social platform such as a forum, a blog, a microblog, a chat room, or a circle of friends is social media data. Since the social platform collects and stores the social media data published by the user, the social media data is collected and stored, so that the social media data stored by each social platform can be read by calling a preset platform docking interface.

For the same user, it often has social accounts of multiple social platforms, and social accounts often use the same registered mailboxes. Therefore, the social media data published by the same social accounts with the same registered mailboxes are summarized and counted. The social media data corresponding to each real user is obtained by summarizing.

S102: Establish a behavior pattern distribution function corresponding to the social media data.

The social media data published by each user on each social platform is analyzed and processed to determine the user's behavior pattern distribution. The behavior pattern distribution indicates the probability of publishing various types of social media data. The above categories may be, for example, sports, music, food, travel, and the like.

In order to generalize the overall state of behavior pattern distribution between social media data of multiple users, the behavior pattern distribution of the user is modeled. Specifically, each user's behavior pattern distribution is traversed first, and then processed by an existing behavior pattern modeling algorithm to generate a behavior pattern distribution function that can be used to describe multiple user behavior patterns.

As an embodiment of the present application, as shown in FIG. 2, the foregoing S102 specifically includes:

S201: Divide the multiple users into multiple user sets.

In this embodiment of the present application, each user has a corresponding one of the user tags. User tags are used to identify different users. The user tag can be, for example, a user's social account number, user name, and the like.

If the social media data obtained in the above S101 is N (N is greater than zero and N is an integer) social media data published by the user, the N users constitute one user set. A user tag is recorded in the user set.

To obtain each subset of the user set, the subset is a set of users obtained by dividing a plurality of users, that is, each subset is also a set of users including one or more user tags.

For example, if the social media data published by 8 users is obtained, and the user labels of the 8 users are “Tom”, “Susan”, “Sue”, “Jack”, “Bob”, “John”, “Zoe”, respectively. "And "Lily", then a set of users composed of these 8 users is {Tom; Susan; Sue; Jack; Bob; John; Zoe; Lily}. A subset of the set of users is, for example, {Tom; Susan; Sue} and {Jack; Bob; John} and the like.

S202: Perform parsing processing on the social media data corresponding to each of the user sets to obtain behavior pattern data of the user set.

For one of the user sets, the social media data corresponding to the user set is social media data published by each user in the user set. Therefore, since the user tags in different user sets are different, the social media data corresponding to different user sets is also different.

The social media data corresponding to each user set is processed, and the behavior pattern data of each user in the user set can be determined. The behavior pattern data indicates the frequency of occurrence and weight of various categories of words in the social media data published by the user.

S203: Input behavior pattern data of each of the user sets into a preset behavior pattern distribution model.

As an embodiment of the present application, the above behavior pattern distribution model is:

Wherein the x is a set of users, the y is a set containing respective words in the social media data, and the f _j (x, y) is a proportion of the j-th feature of the user in the y, λ _j,y is the attribute parameter corresponding to the j-th feature of the user, the Z(x) is a normalization factor, and F is the total number of features of the user set x.

In the embodiment of the present application, the social media data published by each user in the user set x is subjected to word segmentation processing to filter out a plurality of words corresponding to each user. Social media data often contains words that do not have actual semantics. These words do not have the data mining reference value. Therefore, among the selected words, the stop words are removed, and the keywords corresponding to each user are obtained. The first vector formed by each keyword corresponding to each user is referred to as the information state space of the user, and each element in the information state space is different from each other. For example, the information state space corresponding to the user Tom is [apple, London, Messi, football, rain].

For each keyword in the information state space of each user, the frequency of occurrence of the keyword in the social media data corresponding to the user is obtained. The appearance frequencies corresponding to the respective keywords are collectively output as the second vector [a ₁ , a ₂ , a ₃ , a ₄ ..., a _n ]. Where n is the total number of elements in the user's information state space.

In the embodiment of the present application, each feature corresponding to the user set is obtained by a Linear Discriminant Analysis algorithm, and the feature is used to represent the category of the keyword. Each feature has a preset one vocabulary, and each feature corresponds to one attribute parameter λ.

It is determined whether each keyword in the information state space corresponding to the user exists in the thesaurus. If one or more keywords in the information state space corresponding to the user exist in the thesaurus, the respective element values corresponding to the one or more keywords in the second vector are obtained. After the respective element values are added to obtain a sum value, the product of the sum value and the attribute parameter λ of the feature corresponding to the lexicon is output as a behavior feature distribution value of the user.

For example, if f ₁ is one of the features representing “sports” and the lexicon corresponding to the feature is [football, Messi, Ronaldo, manchester], and the attribute parameter is λ ₁ , then the information state corresponding to the user Tom In the space [apple, London, Messi, football, rain], both "Messi" and "football" in the information state space exist in the thesaurus. Assuming that the second vector corresponding to the information state space is [a ₁ , a ₂ , a ₃ , a ₄ ..., a _n ], the element values corresponding to the "Messi" and "football" in the second vector are respectively obtained. , ie a ₃ and a ₄ . Finally, the behavioral attribute distribution value corresponding to the user's "sports" feature is (a ₃ + a ₄ ) λ ₁ .

By analogy, after calculating the distribution of the behavior characteristics of each user in the user set, the behavior pattern distribution model of the attribute parameter λ can be obtained by substituting the above behavior pattern distribution model.

Further, the value corresponding to the normalization factor Z(x) in the behavior pattern distribution model is

S204: Maximizing an entropy of the behavior pattern distribution model to determine an attribute parameter in the behavior pattern distribution model.

In the embodiment of the present application, each attribute parameter in the behavior pattern distribution model is determined based on the maximum entropy principle, wherein each attribute parameter corresponds to one feature. Since the principle of maximum entropy is a common technical means in the art, it will not be repeated one by one.

As an implementation example of the present application, S204 may be implemented based on the following principles: adjusting each attribute parameter in the behavior pattern distribution model multiple times; after each adjustment, calculating the entropy of the current behavior pattern distribution model and recording the attribute parameters at this time; Among the calculated entropies, the largest entropy value is determined, and the attribute parameters corresponding to the entropy value are selected.

S205: Output the behavior pattern distribution model determined by the attribute parameter as a behavior pattern distribution function corresponding to the social media data.

After the attribute parameter λ _{j,y is} determined, the attribute parameter corresponding to each feature is a constant value. Therefore, by re-substituting the constant value into the above formula (1), the behavior of the user set x and the set y as variables can be obtained. Mode distribution function.

In the embodiment of the present application, the behavior pattern data corresponding to each user set is fitted through a preset behavior pattern distribution model, and the behavior pattern distribution function after the attribute parameter determination can be generated. Since the normalization factor is added to the behavior pattern distribution model, the over-fitting of the behavior pattern data is avoided, the accuracy of the behavior pattern distribution function is improved, and the behavior pattern data obtained after the subsequent noise-adding processing can be ensured. It reflects the overall distribution of behavioral attributes and maintains the value of data mining.

S103: Integrate the preset random noise function with the behavior pattern distribution function to obtain a noise-added behavior pattern distribution function.

After adding and integrating the random noise function Q(x) and the behavior pattern distribution function F(x), the obtained noise-added behavior pattern distribution function A(x) is as follows:

A(x)=F(x)+Q(x)

Where x represents the above set of users.

In the embodiment of the present application, the preset random noise function may be a function based on a mechanism such as a Laplace mechanism or an exponential mechanism.

As an embodiment of the present application, the preset random noise function obeys the scale parameter as

Laplace distribution, at this time, the random noise function can be expressed as

Wherein Δf is a global sensitivity of the behavior pattern distribution function, and the ε is a preset privacy protection budget parameter.

In the embodiment of the present application, the privacy protection budget parameter ε is used to control the probability ratio of the behavior pattern distribution function to obtain the same output on two adjacent user sets, which embodies the privacy protection level that the behavior pattern distribution function can provide. The smaller the value of the privacy protection budget parameter ε, the higher the level of privacy protection. When ε=0, the protection level reaches the highest. At this time, for any two adjacent user sets, the behavior pattern distribution function will output two probability distributions completely. The same result. However, these results do not reflect any useful information about the data set. Therefore, in order to achieve a balance between security and usability of the output, the default privacy protection budget parameter ε has a value of 0.1.

Specifically, when the preset random noise function is a Laplacian-based function, the global sensitivity Δf of the user set x is first calculated, and then combined with the preset privacy protection budget parameter ε to determine that the Laplace is consistent. Distributed random noise function.

The global sensitivity Δf of the user set x is calculated by the following formula:

Where F(x) is the above behavior pattern distribution function. ‖F(x)-F(x')‖ is the first-order norm distance between F(x) and F(x'). F(x) and F(x') are adjacent data sets, that is, the maximum difference between the total number of set elements corresponding to F(x) and F(x') is 1.

In the embodiment of the present application, the random user function is added to the behavior pattern distribution function to protect the potential user privacy information in the behavior pattern data. When the attacker needs to query the behavior pattern data, the attacking process is not for the attacker. Visible, thus realizing that even if the attacker has mastered any of the noisy behavior pattern data, the attacker can still not guess the user's original behavior pattern data, thereby being able to completely remove the possibility of leakage of private information from the data source.

S104: Input the behavior mode data query parameter into the noise-added behavior pattern distribution function to obtain the noise-added behavior pattern data.

In the embodiment of the present application, the behavior mode query parameter may be automatically generated, or may be obtained from a data query request sent by the querier.

For example, each user set corresponding to the social media data and each feature appearing in the user set may be traversed, and each of the user sets and any corresponding one of the features are behavior mode query parameters that may appear in the data query request. . Therefore, one of the user sets and features can be read, and the user set and features are used as behavior mode data query parameters, and input into the noisy behavior mode distribution function A(x)=F(x)+Q(x).

Based on each attribute parameter in the noise-adding behavior pattern distribution function and the acquired behavior pattern data query parameter, the noise-added behavior pattern data corresponding to the behavior pattern data query parameter can be directly calculated.

As an embodiment of the present application, FIG. 3 shows a specific implementation process of a method for publishing behavior mode data provided by an embodiment of the present application, which is described in detail below;

S301: Acquire a data query request carrying a behavior mode data query parameter, where the behavior mode data query parameter includes a user set and a feature.

S302: Generate a random parameter corresponding to the feature.

S303: Input the random parameter and the behavior mode data query parameter into the noise-adding behavior pattern distribution function to obtain the noise-added behavior pattern data.

Since the user set and characteristics have been determined, the user's behavior pattern data can be calculated by the behavior pattern distribution function F(x). For each feature in the user set, a random parameter a is generated, and the random parameter a is input into the random noise function, and the random noise value can be calculated based on the global sensitivity Δf of the user set x and the preset privacy protection budget parameter ε. . After the random noise value and the user behavior pattern data are added, the noise-added behavior pattern data is obtained.

S105: Publish the noise-added behavior mode data.

The process of releasing the noise-added behavior pattern data is a process of sharing the noise-added behavior pattern data to any external user under the condition of satisfying the privacy protection, including a process of responding to the behavior pattern data query request issued by the queryer, and All the noise-added behavior pattern data is actively released on a preset data sharing platform.

It should be understood that the size of the sequence of the steps in the above embodiments does not mean that the order of execution is performed. The order of execution of each process should be determined by its function and internal logic, and should not be construed as limiting the implementation process of the embodiments of the present application.

Figure 4 is a block diagram showing the structure of the behavior mode data providing apparatus provided in the embodiment of the present application. For the convenience of description, only the embodiment of the present application is shown. Related parts.

Referring to Figure 4, the apparatus includes:

The obtaining unit 41 is configured to acquire social media data published by a plurality of users.

The establishing unit 42 is configured to establish a behavior pattern distribution function corresponding to the social media data.

The noise adding unit 43 is configured to integrate the preset random noise function with the behavior pattern distribution function to obtain a noise added behavior pattern distribution function.

The input unit 44 is configured to input the behavior mode data query parameter into the noise-added behavior pattern distribution function to obtain the noise-added behavior pattern data.

The issuing unit 45 is configured to release the noise-added behavior mode data.

Optionally, the establishing unit 42 includes:

Dividing a subunit for dividing the plurality of users into a plurality of user sets.

The parsing subunit is configured to parse the social media data corresponding to each of the user sets to obtain behavior pattern data of the user set.

The first input subunit is configured to input behavior mode data of each of the user sets into a preset behavior pattern distribution model.

Determining a subunit for maximizing an entropy of the behavioral pattern distribution model to determine an attribute parameter in the behavioral pattern distribution model.

And an output subunit, configured to output the behavior pattern distribution model determined by the attribute parameter as a behavior pattern distribution function corresponding to the social media data.

Optionally, the behavior pattern distribution model is:

Wherein the x is a set of users, the y is a set containing respective words in the social media data, and the f _j (x, y) is a proportion of the j-th feature of the user in the y, λ _j,y is the attribute parameter corresponding to the j-th feature of the user, and the Z(x) is a normalization factor.

Optionally, the input unit 44 includes:

The obtaining subunit is configured to obtain a data query request carrying a behavior mode data query parameter, where the behavior mode data query parameter includes a user set and a feature.

Generating a subunit for generating a random parameter corresponding to the feature.

And a second input subunit, configured to input the random parameter and the behavior mode data query parameter into the noise added behavior mode distribution function to obtain the noise added behavior pattern data.

Optionally, the preset random noise function obeys a scale parameter

Laplace distribution.

FIG. 5 is a schematic diagram of a terminal device according to an embodiment of the present application. As shown in FIG. 5, the terminal device 5 of this embodiment includes a processor 50 and a memory 51 in which computer readable instructions 52, such as behavioral mode data, executable on the processor 50 are stored. Publish the program. The processor 50 executes the computer readable instructions 52 to implement the steps in the foregoing method for distributing the respective behavior pattern data, such as steps S101 to S105 shown in FIG. 1. Alternatively, the processor 50, when executing the computer readable instructions 52, implements the functions of the various modules/units in the various apparatus embodiments described above, such as the functions of the units 41 through 45 shown in FIG.

Illustratively, the computer readable instructions 52 may be partitioned into one or more modules/units that are stored in the memory 51 and executed by the processor 50, To complete this application. The one or more modules/units may be a series of computer readable instruction segments capable of performing a particular function, the instruction segments being used to describe the execution of the computer readable instructions 52 in the terminal device 5.

The terminal device 5 may be a computing device such as a desktop computer, a notebook, a palmtop computer, and a cloud server. The terminal device may include, but is not limited to, the processor 50 and the memory 51. It will be understood by those skilled in the art that FIG. 5 is only an example of the terminal device 5, does not constitute a limitation of the terminal device 5, may include more or less components than the illustrated, or combine some components, or different components. For example, the terminal device may further include an input/output device, a network access device, a bus, and the like.

The processor 50 may be a central processing unit (CPU), or may be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, etc. The general purpose processor may be a microprocessor or the processor or any conventional processor or the like.

The memory 51 may be an internal storage unit of the terminal device 5, such as a hard disk or a memory of the terminal device 5. The memory 51 may also be an external storage device of the terminal device 5, such as a plug-in hard disk equipped on the terminal device 5, a smart memory card (SMC), and a secure digital (SD). Card, flash card, etc. Further, the memory 51 may also include both an internal storage unit of the terminal device 5 and an external storage device. The memory 51 is configured to store the computer readable instructions and other programs and data required by the terminal device. The memory 51 can also be used to temporarily store data that has been output or is about to be output.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application, in essence or the contribution to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium. A number of instructions are included to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and the like, which can store program codes. .

The above embodiments are only used to explain the technical solutions of the present application, and are not limited thereto; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still The technical solutions described in the embodiments are modified, or the equivalents of the technical features are replaced by the equivalents. The modifications and substitutions of the embodiments do not depart from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

A method for distributing behavior pattern data, comprising:

Get social media data published by multiple users;

Establishing a behavior pattern distribution function corresponding to the social media data;

Integrating the preset random noise function with the behavior pattern distribution function to obtain a noise-added behavior pattern distribution function;

Entering the behavior mode data query parameter into the noise-added behavior pattern distribution function to obtain the noise-added behavior pattern data;

The noise-added behavior pattern data is released.
The method for distributing behavior pattern data according to claim 1, wherein the establishing a behavior pattern distribution function corresponding to the social media data comprises:

Dividing the plurality of users into a plurality of user sets;

Performing parsing processing on the social media data corresponding to each of the user sets to obtain behavior pattern data of the user set;

Inputting behavior pattern data of each of the user sets into a preset behavior pattern distribution model;

Maximizing an entropy of the behavior pattern distribution model to determine an attribute parameter in the behavior pattern distribution model;

Outputting the behavior pattern distribution model determined by the attribute parameter as a behavior pattern distribution function corresponding to the social media data.
The method for distributing behavior pattern data according to claim 2, wherein the behavior pattern distribution model is:

Wherein the x is a set of users, the y is a set containing respective words in the social media data, and the f j (x, y) is a proportion of the j-th feature of the user in the y, λ j,y is the attribute parameter corresponding to the j-th feature of the user, and the Z(x) is a normalization factor.
The method for distributing behavioral pattern data according to claim 2 or 3, wherein the inputting the behavioral pattern data query parameter into the noise-added behavior pattern distribution function to obtain the noise-added behavior pattern data comprises:

Obtaining a data query request carrying a behavior mode data query parameter, where the behavior pattern data query parameter includes a user set and a feature;

Generating a random parameter corresponding to the feature;

And inputting the random parameter and the behavior pattern data query parameter into the noise-added behavior pattern distribution function to obtain the noise-added behavior pattern data.
The method for distributing behavior pattern data according to claim 1, wherein the preset random noise function obeys a scale parameter
Laplace distribution;

Wherein Δf is a global sensitivity of the behavior pattern distribution function, and the ε is a preset privacy protection budget parameter.
A device for distributing behavior pattern data, comprising:

An obtaining unit, configured to acquire social media data published by multiple users;

Establishing a unit, configured to establish a behavior pattern distribution function corresponding to the social media data;

a noise adding unit, configured to integrate the preset random noise function with the behavior pattern distribution function to obtain a noise-added behavior pattern distribution function;

An input unit, configured to input a behavior mode data query parameter into the noise-added behavior mode distribution function to obtain a noise-added behavior mode data;

a publishing unit, configured to release the noise-added behavior mode data.
The apparatus for distributing behavior pattern data according to claim 6, wherein the establishing unit 42 comprises:

Dividing a subunit for dividing the plurality of users into a plurality of user sets;

a parsing subunit, configured to parse the social media data corresponding to each of the user sets to obtain behavior pattern data of the user set;

a first input subunit, configured to input behavior mode data of each of the user sets into a preset behavior pattern distribution model;

Determining a subunit for maximizing an entropy of the behavior pattern distribution model to determine an attribute parameter in the behavior pattern distribution model;

And an output subunit, configured to output the behavior pattern distribution model determined by the attribute parameter as a behavior pattern distribution function corresponding to the social media data.
The apparatus for distributing behavior pattern data according to claim 7, wherein the behavior pattern distribution model is:

Wherein the x is a set of users, the y is a set containing respective words in the social media data, and the f j (x, y) is a proportion of the j-th feature of the user in the y, λ j,y is the attribute parameter corresponding to the j-th feature of the user, and the Z(x) is a normalization factor.
The apparatus for distributing behavior pattern data according to claim 7 or 8, wherein the input unit comprises:

Obtaining a sub-unit, configured to acquire a data query request carrying a behavior mode data query parameter, where the behavior mode data query parameter includes a user set and a feature;

Generating a subunit for generating a random parameter corresponding to the feature;

And a second input subunit, configured to input the random parameter and the behavior mode data query parameter into the noise added behavior mode distribution function to obtain the noise added behavior pattern data.
The apparatus for distributing behavior pattern data according to claim 6, wherein the preset random noise function obeys a scale parameter
Laplace distribution.

Wherein Δf is a global sensitivity of the behavior pattern distribution function, and the ε is a preset privacy protection budget parameter.
A terminal device, comprising: a memory and a processor, wherein the memory stores computer readable instructions executable on the processor, and the processor implements the following steps when the computer readable instructions are executed :

Get social media data published by multiple users;

Establishing a behavior pattern distribution function corresponding to the social media data;

Integrating the preset random noise function with the behavior pattern distribution function to obtain a noise-added behavior pattern distribution function;

Entering the behavior mode data query parameter into the noise-added behavior pattern distribution function to obtain the noise-added behavior pattern data;

The noise-added behavior pattern data is released.
The terminal device according to claim 11, wherein the establishing a behavior pattern distribution function corresponding to the social media data comprises:

Dividing the plurality of users into a plurality of user sets;

Performing parsing processing on the social media data corresponding to each of the user sets to obtain behavior pattern data of the user set;

Inputting behavior pattern data of each of the user sets into a preset behavior pattern distribution model;

Maximizing an entropy of the behavior pattern distribution model to determine an attribute parameter in the behavior pattern distribution model;

Outputting the behavior pattern distribution model determined by the attribute parameter as a behavior pattern distribution function corresponding to the social media data.
The terminal device according to claim 12, wherein the behavior pattern distribution model is:

Wherein the x is a set of users, the y is a set containing respective words in the social media data, and the f j (x, y) is a proportion of the j-th feature of the user in the y, λ j,y is the attribute parameter corresponding to the j-th feature of the user, and the Z(x) is a normalization factor.
The terminal device according to claim 12 or 13, wherein the inputting the behavior mode data query parameter into the noise-adding behavior mode distribution function to obtain the noise-adding behavior pattern data comprises:

Obtaining a data query request carrying a behavior mode data query parameter, where the behavior pattern data query parameter includes a user set and a feature;

Generating a random parameter corresponding to the feature;

And inputting the random parameter and the behavior pattern data query parameter into the noise-added behavior pattern distribution function to obtain the noise-added behavior pattern data.
The terminal device according to claim 11, wherein the preset random noise function obeys a scale parameter
Laplace distribution;

Wherein Δf is a global sensitivity of the behavior pattern distribution function, and the ε is a preset privacy protection budget parameter.
A computer readable storage medium storing computer readable instructions, wherein the computer readable instructions, when executed by at least one processor, implement the following steps:

Get social media data published by multiple users;

Establishing a behavior pattern distribution function corresponding to the social media data;

Integrating the preset random noise function with the behavior pattern distribution function to obtain a noise-added behavior pattern distribution function;

Entering the behavior mode data query parameter into the noise-added behavior pattern distribution function to obtain the noise-added behavior pattern data;

The noise-added behavior pattern data is released.
The computer readable storage medium according to claim 16, wherein the establishing a behavior pattern distribution function corresponding to the social media data comprises:

Dividing the plurality of users into a plurality of user sets;

Performing parsing processing on the social media data corresponding to each of the user sets to obtain behavior pattern data of the user set;

Inputting behavior pattern data of each of the user sets into a preset behavior pattern distribution model;

Maximizing an entropy of the behavior pattern distribution model to determine an attribute parameter in the behavior pattern distribution model;

The behavior pattern distribution model determined by the attribute parameter is output as a behavior pattern distribution function corresponding to the social media data.
The computer readable storage medium of claim 17, wherein the behavior pattern distribution model is:

Wherein the x is a set of users, the y is a set containing respective words in the social media data, and the f j (x, y) is a proportion of the j-th feature of the user in the y, λ j,y is the attribute parameter corresponding to the j-th feature of the user, and the Z(x) is a normalization factor.
The computer readable storage medium according to claim 17 or 18, wherein the inputting the behavior mode data query parameter into the noise-added behavior pattern distribution function to obtain the noise-added behavior pattern data comprises:

Obtaining a data query request carrying a behavior mode data query parameter, where the behavior pattern data query parameter includes a user set and a feature;

Generating a random parameter corresponding to the feature;

And inputting the random parameter and the behavior pattern data query parameter into the noise-added behavior pattern distribution function to obtain the noise-added behavior pattern data.
The computer readable storage medium according to claim 16, wherein said predetermined said random noise function obeys a scale parameter
Laplace distribution;

Wherein Δf is a global sensitivity of the behavior pattern distribution function, and the ε is a preset privacy protection budget parameter.