CN113177166B - A method and system for personalized location semantic publishing based on differential privacy - Google Patents

Abstract

The invention relates to a method and system for publishing personalized location semantics based on differential privacy, belonging to the field of data mining and privacy protection. Firstly, according to the privacy protection requirements of users, a semantic privacy protection level with a parameter of 1 is set, and the distance to be protected semantics is calculated. The semantics of the most recent l-1 locations; then, the semantic sensitivity of all locations is calculated according to the number of semantic visits of the user, and the release probabilities of l semantics are obtained based on the semantic sensitivity; finally, the Gaussian variables and exponential distribution variables of specific parameters are used to generate The Laplacian variable with a specific probability, that is, the posted user location that meets a specific semantic sensitivity. It solves the problem of not protecting the location semantics in the existing differential privacy publishing location methods.

Description

A method and system for personalized location semantic publishing based on differential privacy

technical field

The invention belongs to the field of data mining and privacy protection, and relates to a method and system for publishing personalized location semantics based on differential privacy.

Background technique

With the widespread use of mobile terminal equipment such as mobile phones and the rapid development of wireless communication technology, location-based services (Location-Based Service, LBS) are used more and more frequently. Surrounding store search, information push and other services. In the process of location service, a large amount of spatial location data will be generated. In order to make relevant pushes according to the user's preferences, the LBS provider will upload, publish and share the collected user location data. However, some sensitive information of users may be involved in the shared location data, and the data owner may not want to share their location data directly.

At present, there are mainly three types of location privacy protection methods: based on spatial anonymity, based on encryption and based on perturbation. Spatial anonymity mainly hides the user's location, sets the corresponding anonymous parameter level, and confuses the user's original value with the anonymous value to protect the privacy of the user's location. However, the anonymous parameter level is difficult to set based on anonymity-based protection methods, and The availability of data after anonymity is not high; encryption-based location privacy protection methods usually use symmetric encryption and asymmetric encryption algorithms to encrypt location data, thereby hiding the true value of location data, but encryption-based methods are often more complicated, and the impact on communication resources is limited. The consumption is very huge; among the perturbation-based methods, the differential privacy protection method is the most important privacy protection method for location privacy protection due to its rigorous mathematical reasoning model and no restrictions on the background knowledge of the attacker.

The current location differential privacy protection usually uses the Laplacian noise mechanism to perturb the latitude and longitude of the original location in a small range, which can provide high data availability while protecting the precise latitude and longitude data of the location. However, location semantics, as an important part of location information, often contains sensitive information of users (for example, home address, check-in location, etc.). The existing differential privacy protection method only protects the latitude and longitude data of the location, but does not protect the location semantics of users. The attacker can obtain the location semantic information of the user through location semantic inference. How to protect both the user's spatial location data and the user's location semantics when publishing the user's location is an urgent problem to be solved.

SUMMARY OF THE INVENTION

In view of this, the purpose of the present invention is to provide a method and system for publishing personalized location semantics based on differential privacy. First, according to the user's privacy protection requirements, a semantic privacy protection level with a parameter of 1 is set, and the semantic privacy protection level closest to the semantics to be protected is calculated. l-1 location semantics; then calculate the semantic sensitivity of all locations according to the number of semantic visits of the user, and obtain the release probability of l semantics based on the semantic sensitivity; finally, the Gaussian variables and exponential distribution variables of specific parameters are used to generate certain probabilities. The Laplacian variable of , which is the published user location that meets a certain semantic sensitivity.

To achieve the above object, the present invention provides the following technical solutions:

In one aspect, the present invention provides a method for publishing personalized location semantics based on differential privacy, comprising the following steps:

其中

分别表示第i个位置的经度、纬度和语义；S1: Data preprocessing, performing data cleaning and reduction on the originally collected location data, and obtaining the location-sensitive data to be protected X={x ₁ ,..., _xi ,...,x _n }, a total of n position, where x _i represents the ith position,

in

respectively represent the longitude, latitude and semantics of the i-th position;

S2: Set the corresponding semantic privacy protection level l according to the semantic privacy protection requirements;

S3: Calculate semantic sensitivity;

S4: Noise generation, according to the generation principle of Laplacian noise to generate Laplacian noise that conforms to a specific semantic sensitivity;

S5: adding noise, adding the obtained Laplacian noise to the position data to obtain new position data X' _lng =X _lng +Y _lng , X' _lat =X _lat +Y _lat ;

S6: iterative processing, iterative processing of each position, repeating steps S2-S5, until all position data processing is completed;

分别表示加噪声后第i个位置的经度、纬度和语义。S7: Data release, for each processed location data, there is new perturbed location data corresponding to it, and these locations are in at least l different semantics, and one location semantics is selected as the user's location release , the new position released is X′={x′ ₁ ,x′ ₂ ,...,x′ _i ,...,x′ _n }, where x′ _i represents the ith position after adding noise

represent the longitude, latitude and semantics of the i-th position after adding noise, respectively.

Further, step S3 specifically includes the following steps:

表示第i个语义的经纬度范围；S31: Calculate the nearest l-1 semantics of the i-th position x _i according to the Euclidean distance, and use the semantics to which it belongs and the l-1 semantics as a semantic set Sem(sem ₁ ,sem ₂ ,...,sem _i ,...,sem _l ), where

Indicates the latitude and longitude range of the i-th semantic;

计算公式如下式，S32: position semantic sensitivity calculation, calculating the sensitivity of the l semantics obtained in step S3-1

The calculation formula is as follows,

Among them, H(sem _i ) represents the total number of times that semantic sem _i is accessed, and L represents the sum of the times that all semantics are accessed.

Further, the step S4 specifically includes the following steps:

S41: Calculate the standard deviation σ ₁ , σ ₂ of the Gaussian inverse cumulative distribution function according to the semantic sensitivity and the position semantic range:

Among them, μ is 0, σ ₁ , σ ₂ are the required Gaussian standard deviation parameters;

S42: Calculate the parameters λ ₁ , λ ₂ required by the inverse cumulative exponential distribution function to generate the exponential distribution according to the semantic sensitivity and the position semantic range:

S43: generate Gaussian distribution noise Z _lng , Z _lat according to the Gaussian distribution parameters obtained in step S41;

S44: generate exponentially distributed noises W _lng , W _lat according to the exponential distribution parameters obtained in step S42;

其中Y_lng，Y_lat即为产生的符合特定语义敏感度的经纬度噪声。S45: Compute generalized Laplace variables

Among them, Y _lng and Y _lat are the generated longitude and latitude noises that conform to specific semantic sensitivity.

In another aspect, the present invention provides a personalized location semantic publishing system based on differential privacy, comprising:

分别表示第i个位置的经度、纬度和语义。Data preprocessing module: used to clean and reduce the original collected position data to obtain the position data to be protected X={x ₁ ,x ₂ ,..., _xi ,...,x _n }, where x _i represents the ith position,

respectively represent the longitude, latitude and semantics of the ith position.

Parameter setting module: used to set the semantic location privacy level protection parameter l.

Semantic sensitivity calculation module: used to calculate the sensitivity of the l semantics P _sem = (p _sem1 , p _sem2 , . . . , p _seml );

Noise generation module: used to generate Laplacian noise that conforms to a specific semantic sensitivity;

Noise adding module: used to add the generalized Laplacian noise obtained by the fifth unit in the noise generating module to the position data to obtain new position data X′ _lng =X _lng +Y _lng , X′ _lat =X _lat +Y _lat ;

Iterative processing module: used to iteratively process each position until all position data updates are completed;

分别表示加噪声后第i个位置的经度、纬度和语义。Data publishing module: For each processed position data, there is new perturbed position data corresponding to it, and these positions are in at least l different semantics, and one position semantics is selected as the user's position to publish, The new position published is X′={x′ ₁ ,x′ ₂ ,...,x′ _i ,...,x′ _n }, where x′ _i represents the ith position after adding noise

represent the longitude, latitude and semantics of the i-th position after adding noise, respectively.

Further, the semantic sensitivity calculation module includes the following subunits:

表示第i个语义的经纬度范围；The first unit of semantic sensitivity calculation: calculate the l-1 semantics closest to the i-th position according to the Euclidean distance, and take the semantics to which it belongs and the l-1 semantics as a semantic set Sem=(sem ₁ , sem ₂ , .. .,sem _l ), where

Indicates the latitude and longitude range of the i-th semantic;

如下式：The second unit of semantic sensitivity calculation: positional semantic sensitivity calculation, calculating the sensitivity of these l semantics

The formula is as follows:

Among them, H(sem _i ) represents the total number of times that semantic sem _i is accessed, and L represents the sum of the times that all semantics are accessed.

Further, the noise generation module includes the following subunits:

The first unit of noise generation, according to the semantic sensitivity and the position semantic range, calculate the standard deviation of the Gaussian inverse cumulative distribution function σ ₁ , σ ₂ ,

Among them, μ is 0, σ ₁ , σ ₂ are the parameters we seek;

The second unit of noise generation, according to the semantic sensitivity and the position semantic range, obtains the parameters λ ₁ , λ ₂ required by the inverse cumulative exponential distribution function to generate the exponential distribution;

The third unit of noise generation, according to the Gaussian distribution parameters obtained in step S4-1, generates Gaussian distribution noise Z _lng , Z _lat ;

The fourth unit of noise generation, according to the exponential distribution parameters obtained in step S4-2, generates exponentially distributed noises W _lng , W _lat ;

Y_lng，Y_lat即为产生的符合特定语义敏感度的经纬度噪声。Noise Generation Unit 5, computes generalized Laplace variables

Y _lng , Y _lat are the generated latitude and longitude noise that conforms to specific semantic sensitivity.

The beneficial effects of the present invention are: the present invention can generate Laplacian noise with a specific probability, not only can protect the privacy of the latitude and longitude data of the location, but also protect the semantic privacy of the user's location; the privacy protection can be based on the user's different semantic sensitivities It generates noise that meets specific semantic sensitivity and realizes personalized protection of user location semantics; the implementation process and steps are simple and easy to implement, improve the availability of published data and reduce the consumption of communication resources, and have important market value.

Other advantages, objects, and features of the present invention will be set forth in the description that follows, and will be apparent to those skilled in the art based on a study of the following, to the extent that is taught in the practice of the present invention. The objectives and other advantages of the present invention may be realized and attained by the following description.

Description of drawings

In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be preferably described in detail below with reference to the accompanying drawings, wherein:

FIG. 1 is a flowchart showing the steps of a method for publishing personalized location semantics based on differential privacy according to an embodiment of the present invention;

2 is an overall flow chart provided by an embodiment of the present invention;

FIG. 3 is a schematic diagram of a system for publishing personalized location semantics based on differential privacy according to an embodiment of the present invention.

Detailed ways

Embodiments of the present invention are described below through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the drawings provided in the following embodiments are only used to illustrate the basic idea of the present invention in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

Among them, the accompanying drawings are only used for exemplary description, and represent only schematic diagrams, not physical drawings, and should not be construed as limitations of the present invention; in order to better illustrate the embodiments of the present invention, some parts of the accompanying drawings will be omitted, The enlargement or reduction does not represent the size of the actual product; it is understandable to those skilled in the art that some well-known structures and their descriptions in the accompanying drawings may be omitted.

The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar components; in the description of the present invention, it should be understood that if there are terms “upper”, “lower”, “left” and “right” , "front", "rear" and other indicated orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the indicated device or element must be It has a specific orientation, is constructed and operated in a specific orientation, so the terms describing the positional relationship in the accompanying drawings are only used for exemplary illustration, and should not be construed as a limitation of the present invention. situation to understand the specific meaning of the above terms.

Please refer to FIG. 1 to FIG. 3. FIG. 1 and FIG. 2 are flowcharts of the overall method for implementing the present invention, respectively. The specific steps of the embodiment of the method for generating Laplacian noise with specific semantic sensitivity provided by the present invention include:

分别表示第i个位置的经度、纬度和语义。Step S1, data preprocessing, performing data cleaning and reduction on the originally collected location data, to obtain the location-sensitive data to be protected X={x ₁ ,..., _xi ,...,x _n }, a total of n positions, where x _i represents the ith position,

respectively represent the longitude, latitude and semantics of the ith position.

In the embodiment, the check-in data is cleaned and reduced to obtain the check-in data X={x ₁ , x ₂ , . . . , x ₁₃₀₃ } to be protected.

Step S2, setting the semantic privacy protection level parameter, and the user sets the corresponding privacy protection level 1 according to his own semantic privacy protection requirements.

In the embodiment, the semantic privacy protection level parameter l=4 is set, and a person skilled in the art may set the semantic privacy protection level parameter by themselves during specific implementation.

Step S3, semantic sensitivity calculation, including the following steps,

表示第i个语义的经纬度范围。Step S3-1, calculate the nearest l-1 semantics of the i-th position x _i according to the Euclidean distance, and take the semantics to which it belongs and the l-1 semantics as a semantic set Sem=(sem ₁ , sem ₂ , .. .,sem _l ), where

Indicates the latitude and longitude range of the i-th semantic.

In the embodiment, take the first position point x ₁ ={106.61704, 29.541919, <Xinhua Bookstore>}, and calculate the nearest three semantics near the first position according to the Euclidean distance as {<New Century Supermarket>, <China Mobile >,<KFC>} The four semantic ranges are respectively, the semantic range of Xinhua Bookstore is {[106.61647,106.61760],[29.541427,29.542410]}, the semantic range of New Century Supermarket is {[106.61495,106.616128],[ 29.541529, 29.54255]}, the semantic range of China Mobile is {[106.613461, 106.61486], [29.54121, 29.5424362]}, the semantic range of KFC is {[106.617366, 106.61809], [29.54063, 29.54126]};

计算公式如下式，Step S3-2, calculating the positional semantic sensitivity, calculating the sensitivity of the l semantics obtained in step S3-1

The calculation formula is as follows,

Among them, H(sem _i ) represents the total number of times that semantic sem _i is accessed, and L represents the sum of the times that all semantics are accessed.

In the example, Xinhua Bookstore has been visited 30 times, New Century Supermarket has been visited 200 times, China Mobile has been visited 15 times, and KFC has been visited 10 times. , 0.007;

Step S4, noise generation, generating Laplacian noise conforming to a specific semantic sensitivity according to the Laplacian noise generation principle, including the following steps:

Step S4-1, according to the semantic sensitivity and the position semantic range, the standard deviations σ ₁ and σ ₂ of the Gaussian inverse cumulative distribution function are obtained:

Among them, μ is 0, σ ₁ , σ ₂ are the required Gaussian standard deviation parameters;

In the embodiment, four semantic parameters σ ₁ ={0.00346, 0.0000213, 0.000317, 0.0000278}, σ ₂ ={0.00023001, 0.000321, 0.0000378, 0.00001265} are respectively obtained according to the semantic range and the calculated semantic sensitivity in step S3

Step S4-2, according to the semantic sensitivity and the position semantic range, obtain the parameters λ ₁ , λ ₂ required by the inverse cumulative exponential distribution function to generate the exponential distribution:

In the embodiment, four semantic parameters λ ₁ ={0.000075, 0.00067, 0.0000379, 0.0000543}, λ ₂ ={0.0003954, 0.0001534, 0.000069023, 0.00001357} are respectively obtained according to the semantic range and the calculated semantic sensitivity in step S3;

Step S4-3, according to the Gaussian distribution parameters obtained in step S4-1, generate Gaussian distribution noise Z _lng , Z _lat ;

In an embodiment, one of the sets of σ ₁ and σ ₂ obtained in step S4-1 is randomly selected as a Gaussian distribution parameter. Here, we select 0.000317 from σ ₁ to generate Z _lng = 0.0000131, and select 0.000321 from σ ₂ to generate Z _lat = 0.000678;

Step S4-4, according to the exponential distribution parameters obtained in step S4-2, generate exponentially distributed noises W _lng , W _lat ;

In the embodiment, one is randomly selected from the collection of λ ₁ and λ ₂ obtained in step S4-2 as an exponential distribution parameter, here we select 0.00067 from λ ₁ to generate W _lng =0.000362, and select 0.000354 from σ ₂ to generate W _lat = 0.000714;

Y_lng，Y_lat即为产生的符合特定语义敏感度的经纬度噪声。Step S4-5, calculate the generalized Laplace variable

Y _lng , Y _lat are the generated latitude and longitude noise that conforms to specific semantic sensitivity.

In an embodiment, according to Z _lng , Z _lat , W _lng , and W _lat obtained in step S4-3 and step S4-4, Y _lng =0.0002492444, Y _lat =0.000181166;

Step S5, adding noise, adding the generalized Laplacian noise obtained in step S4-5 to the position data to obtain new position data X' _lng =X _lng +Y _lng , X' _lat =X _lat +Y _lat .

In an embodiment, the generalized Laplacian noise generated in step S4-5 is added to the first position data x ₁ ={106.61704, 29.541919, <Xinhua Bookstore>} to obtain new position data to obtain position data x′ ₁ ={106.6172892444, 29.54210016}.

Step S6, iterative processing, iterative processing of each position, and repeating the above steps S2-S5 until all position data processing is completed.

In the embodiment, traverse each position data, and perform the above-mentioned steps S2-S5 processing on all 1303 positions, until all position data processing is completed;

分别表示加噪声后第i个位置的经度、纬度和语义。Step S7, data release, for each processed location data, there is new perturbed location data corresponding to it, and these locations are in at least l different semantics, we select one location semantics as the user's location semantics. The position is published, and the new position published is X′={x′ ₁ ,x′ ₂ ,...,x′ _i ,...,x′ _n }, where x′ _i represents the ith position after adding noise ,

represent the longitude, latitude and semantics of the i-th position after adding noise, respectively.

_In the embodiment, the position data released by x1 is the longitude and latitude data after adding noise 106.6172892444, 29.54210016, and the semantics are Xinhua Bookstore, New Century Supermarket, China Mobile, and KFC. The size of the semantic sensitivity is used as the probability measurement unit to select a semantic release.

In specific implementation, the method provided by the present invention can realize the automatic running process based on software technology, and can also realize the corresponding system in a modular way.

分别表示第i个位置的经度、纬度和语义。The data preprocessing module is used to clean and reduce the originally collected position data to obtain the position data to be protected X={x ₁ ,x ₂ ,..., _xi ,...,x _n }, where x _i represents the ith position,

respectively represent the longitude, latitude and semantics of the ith position.

The parameter setting module is used to set the semantic location privacy level protection parameter l.

The semantic sensitivity calculation module is used to calculate the sensitivity of the l semantics P _sem = (p _sem1 , p _sem2 ,..., p _seml ), including the following subunits,

表示第i个语义的经纬度范围。The first unit calculates the nearest l-1 semantics of the i-th position according to the Euclidean distance, and takes the semantics to which it belongs and the l-1 semantics as a semantic set Sem=(sem ₁ ,sem ₂ ,...,sem _i ,...sem _l ), where

Indicates the latitude and longitude range of the i-th semantic.

如下式，The second unit, location semantic sensitivity calculation, calculates the sensitivity of these l semantics

as the following formula,

Among them, H(sem _i ) represents the total number of times that semantic sem _i is accessed, and L represents the sum of the times that all semantics are accessed.

A noise generation module for generating Laplacian noise conforming to a specific semantic sensitivity, including the following subunits,

The first unit, according to the semantic sensitivity and the position semantic range, calculate the standard deviation of the Gaussian inverse cumulative distribution function σ ₁ , σ ₂ ,

Among them, μ is 0, σ ₁ , σ ₂ are the parameters we seek;

The second unit is to obtain the parameters λ ₁ , λ ₂ required by the inverse cumulative exponential distribution function to generate the exponential distribution according to the semantic sensitivity and the location semantic range;

The third unit, according to the Gaussian distribution parameters obtained in step S4-1, generates Gaussian distribution noise Z _lng , Z _lat ;

The fourth unit, according to the exponential distribution parameters obtained in step S4-2, generates exponential distribution noises W _lng , W _lat ;

Y_lng，Y_lat即为产生的符合特定语义敏感度的经纬度噪声。Unit 5, Computing Generalized Laplace Variables

Y _lng , Y _lat are the generated latitude and longitude noise that conforms to specific semantic sensitivity.

The noise adding module is used to add the generalized Laplace noise obtained by the fifth unit in the noise generating module to the position data to obtain new position data X′ _lng =X _lng +Y _lng , X′ _lat =X _lat +Y _lat .

The iterative processing module is used to iteratively process each position, and repeat the above steps S2-S5 until all the position data are updated.

分别表示加噪声后第i个位置的经度、纬度和语义。In the data publishing module, for each processed location data, there is new perturbed location data corresponding to it, and these locations are in at least l different semantics, and we select one location semantics as the user's location release. , the new position released is X′={x′ ₁ ,x′ ₂ ,...,x′ _i ,...,x′ _n }, where x′ _i represents the ith position after adding noise,

represent the longitude, latitude and semantics of the i-th position after adding noise, respectively.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be Modifications or equivalent replacements, without departing from the spirit and scope of the technical solution, should all be included in the scope of the claims of the present invention.

Claims (2)

1. A personalized position semantic publishing method based on differential privacy is characterized in that: the method comprises the following steps:

s1: data preprocessing, namely cleaning and stipulating the originally acquired position data to obtain position sensitive data X = { X = to-be-protected ₁ ,...,x _i ,...,x _n H, a total of n positions, where x _i Which represents the (i) th position of the (c),

wherein

Respectively representing longitude, latitude and semantics of the ith position;

s2: setting a corresponding semantic privacy protection level l according to the semantic privacy protection requirement;

s3: calculating semantic sensitivity; the step S3 specifically includes the following steps:

s31: calculating the ith position x according to Euclidean distance _i Last l-1 semanticsThe semantic meaning of the self and the 1 semantic meaning are taken as a semantic set Sem (Sem) ₁ ,sem ₂ ,...,sem _i ,...,sem _l ) Wherein

Representing a latitude and longitude range of the ith semantic;

s32: position semantic sensitivity calculation, namely calculating the sensitivity of the semantic words obtained in the step S31

The calculation formula is as follows,

wherein, H (sem) _i ) Representing semantics sem _i The total number of times of access, L represents the sum of the number of times of access of all semantics;

s4: generating noise, namely generating the Laplacian noise which accords with the specific semantic sensitivity according to the generation principle of the Laplacian noise; the step S4 specifically includes the following steps:

s41: calculating standard deviation sigma of inverse Gaussian cumulative distribution function according to semantic sensitivity and position semantic range ₁ ，σ ₂ ：

Wherein μ is 0 and σ ₁ ，σ ₂ The calculated Gaussian standard deviation parameter is obtained;

s42: calculating the parameter lambda required by generating the exponential distribution by the inverse cumulative exponential distribution function according to the semantic sensitivity and the position semantic range ₁ ,λ ₂ ：

S43: generating Gaussian distribution noise Z based on the Gaussian distribution parameters obtained in step S41 _lng ，Z _lat ；

S44: generating an exponential distribution noise W based on the exponential distribution parameter obtained in step S42 _lng ，W _lat ；

S45: calculating a generalized Laplace variable

Wherein Y is _lng ，Y _lat The latitude and longitude noise which is generated and accords with the specific semantic sensitivity is obtained;

s5: adding the obtained Laplacian noise to the position data to obtain new position data X' _lng ＝X _lng +Y _lng ，X′ _lat ＝X _lat +Y _lat ；

S6: performing iterative processing, namely performing iterative processing on each position, and repeating the steps S2-S5 until the data processing of all the positions is finished;

s7: data issuing, wherein for each processed position data, new disturbed position data corresponds to the position data, the positions are at least in l different semantics, one position semantic is selected from the position data to be issued as the position of a user, and the issued new position is X '= { X' ₁ ,x' ₂ ,...,x′ _i ,...,x' _n X 'therein' _i Indicating the ith position after noise addition

Respectively representing longitude, latitude and semantics of the ith position after noise addition.

2. A personalized position semantic publishing system based on differential privacy is characterized in that: comprises that

A data preprocessing module: the method is used for carrying out data cleaning and specification on the originally acquired position data to obtain position data X = { X = to-be-protected ₁ ,x ₂ ,...,x _i ,...,x _n In which x _i Which represents the (i) th position of the (c),

respectively representing longitude, latitude and semantics of the ith position;

a parameter setting module: the method is used for setting a semantic location privacy level protection parameter l;

a semantic sensitivity calculation module: sensitivity for calculating this semantic

A noise generation module: generating laplacian noise that conforms to a particular semantic sensitivity;

a noise adding module: is used for adding the generalized Laplacian noise obtained by the fifth unit in the noise generation module into the position data to obtain new position data X' _lng ＝X _lng +Y _lng ，X′ _lat ＝X _lat +Y _lat ；

An iteration processing module: the system is used for iteratively processing each position until all position data are updated;

the data release module: for each processed position data, new perturbed position data is corresponding to the processed position data, and the positions are at least in one different semantic meaning, and are selected from the one or more different semantic meaningsOne location semantic is issued as the location of the user, and the new location issued is X '= { X' ₁ ,x' ₂ ,...,x′ _i ,...,x' _n X 'therein' _i Indicating the i-th position after the addition of noise,

respectively representing longitude, latitude and semantics of the ith position after noise addition;

the semantic sensitivity calculation module comprises the following subunits:

semantic sensitivity calculation first unit: calculating l-1 semanteme closest to the ith position according to the Euclidean distance, and taking the semanteme to which the semanteme belongs and the l-1 semantemes as a semantic set Sem = (Sem) ₁ ,sem ₂ ,...,sem _i ,...,sem _l ) Wherein

Representing a latitude and longitude range of the ith semantic;

semantic sensitivity calculation second unit: position semantic sensitivity calculation, calculating the sensitivity of the semantic

The following formula:

wherein, H (sem) _i ) Representing semantics sem _i The total number of times of access, L represents the sum of the number of times of access of all semantics;

the noise generation module comprises the following subunits:

a first noise generation unit for calculating standard deviation sigma of Gaussian inverse cumulative distribution function according to semantic sensitivity and position semantic range ₁ ，σ ₂ ，

Wherein μ is 0 and σ ₁ ,σ ₂ The parameters are obtained;

a second noise generation unit for calculating the parameter lambda needed by the inverse cumulative exponential distribution function to generate the exponential distribution according to the semantic sensitivity and the position semantic range ₁ ,λ ₂ ；

A third noise generation unit for generating Gaussian distribution noise Z based on the Gaussian distribution parameter obtained in step S41 _lng ，Z _lat ；

A fourth noise generation unit for generating an exponential distribution noise W based on the exponential distribution parameter obtained in step S42 _lng ，W _lat ；

A fifth unit for generating noise and calculating generalized Laplace variable

Y _lng ，Y _lat I.e. the generated longitude and latitude noise which conforms to the specific semantic sensitivity.

Images