Server privacy protection system based on big data

Abstract

The invention discloses a server privacy protection system based on big data, which comprises a data acquisition module, a data analysis module and an output module, and is characterized in that the data analysis module is used for analyzing and judging whether preference data has significant infringement on the privacy of a user, and the data acquisition module, the data analysis module and the output module are in mutual communication connection. The time coefficient analysis module is used for extracting and processing time node related information of user browsing records, and browsing parameters of users for preference information are subjected to three-dimensional visualization, so that the detection dimension of user privacy protection is improved, and the detection accuracy and the humanization degree are improved; through the information disclosure analysis module, the time coefficient model constructed by the user browsing record is approached in a mode of detecting the area of the model, it is guaranteed that the extracted time node is in the constructed time coefficient model, the sampling inspection accuracy is guaranteed, and the method has the advantages of being high in detection information coverage capacity and high in humanization degree.

Description

Technical Field

[0001] The present invention relates to the technical field of network data management, and in particular to a server privacy protection system based on big data. Background Art

[0002] With the continuous development of science and technology networks, network terminal systems automatically locate users' locations, voices and other information through videos to obtain users' preferred content and push it to various clients. Even the user's age, city of residence, gender, occupation range, and even more subtle driving type, pet type, fitness equipment, etc.... This information can be recorded and analyzed through daily behavior. At the same time, since most of the data between clients are interoperable, the user's behavior on various software is open source and then collected. All kinds of information registered with the mobile phone number of each software, as well as those not registered but recorded on the IP address and IMEI device serial number, will be stored in the client server.

[0003] However, in the process of analyzing, extracting and pushing user preferences, the system cannot understand the user's interest in information push. Due to the insufficient information pick-up of some pushed information, these pushes are not helpful to users. Especially when the information is pushed too frequently, it may infringe on the user's privacy and greatly affect the user's experience of using the device. Therefore, it is necessary to design a server privacy protection system based on big data with strong detection information coverage and high humanization. Summary of the invention

[0004] The purpose of the present invention is to provide a server privacy protection system based on big data to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a server privacy protection system based on big data, including a data acquisition module, a data analysis module and an output module, the data acquisition module is used to obtain the user preference data results analyzed in the client; the data analysis module is used to analyze and judge whether the preference data has a significant infringement on the user's privacy; the output module is used to test and feedback the data information pushed to the user, and improve the parameters of the push mode, the data acquisition module, the data analysis module and the output module are communicatively connected to each other.

[0006] According to the above technical solution, the data collection module includes a preference data extraction module and a user information collection module. The preference data extraction module is used to extract user preference information collected by the client; the user information collection module is used to obtain user personal information.

[0007] According to the above technical solution, the data analysis module includes a time coefficient analysis module and an information disclosure analysis module. The time coefficient analysis module is used to obtain the timeliness parameters of user information collected by the client; the information disclosure analysis module is used to analyze the rationality of pushing public target information to users.

[0008] According to the above technical solution, the information disclosure analysis module further includes an information relevance detection submodule and an acceptability coefficient detection submodule. The information relevance detection submodule is used to obtain the relevance parameters of information retrieved by users on various platforms; the acceptability coefficient detection submodule is used to determine whether the current type of preference information is suitable for continued push.

[0009] According to the above technical solution, the output module includes an information shielding module and a pushing module. The information shielding module is used to output instructions for shielding and pushing such information; and the pushing module is used to output instructions for pushing such information.

[0010] According to the above technical solution, the operation method of the server privacy protection system mainly includes the following steps:

[0011] Step S1: accessing the server privacy protection system in the client to extract the user preference information and user personal information data obtained by the client;

[0012] Step S2: The system analyzes the time coefficient of the user's preference data and the information relevance of the browsing data, and predicts the user's acceptance of the disclosure of personal information;

[0013] Step S3: the access test module pushes the target information to the user, and the feedback result submodule determines the user's feedback to determine whether the current type of preference information is suitable for continued push;

[0014] Step S4: The system implements shielding and pushing instructions for the preference information through the information shielding module and the pushing module respectively.

[0015] According to the above technical solution, step S2 further includes:

[0016] Step S21: The time coefficient analysis module records the time node data of the preference information obtained by the client, and obtains the elapsed time value of each detected time node from the current time, which is arranged from low to high as t 1 ,t 2 ……t n+1 , then the browsing time interval of user preference data Where n is the number of intervals between recorded time nodes, then the time difference between the last browsing of the user of this data The unit of A is hours;

[0017] Step S22: The time coefficient analysis module randomly selects two end and start time nodes from the time nodes obtained in step S21, sets the interval between the two selected time nodes as the monitoring time period, and records the number of time nodes that appear within the monitoring time period. The average time for preference information to appear once within the monitoring time period is Where T 0 is the time interval of the monitoring period, T 0 The unit is hour, K is the number of times the preference information appears in the monitoring time period; Step S23: The time coefficient analysis module records the K browsing information of the preference information appearing in the client in the monitoring time period obtained in the step S22, and the browsing time data of the K browsing information, and obtains the minimum value T of the browsing information time 1 and the maximum browsing time T 2 , calculate the browsing time difference T 差 =T 2 -T 1 ,Will The record is the average value of the sampled browsing time data, where the browsing information time T is in seconds, and φ is the parameter multiple matching value of the component time coefficient model;

[0018] Step S24: Establish a spatial rectangular coordinate system with the user browsing time difference record as the X-axis of the time coefficient model, the browsing record interval time record within the monitoring time period as the Y-axis of the time coefficient model, and the sampled browsing time data record as the Z-axis of the time coefficient model, wherein the coordinate origin is (0, 0), the length of the positive half axis of the X-axis is the user's most recent browsing time difference A, the length of the positive half axis of the Y-axis is the average appearance time ΔT, and the length of the positive half axis of the Z-axis is the difference H of the sampled browsing time data.

[0019] According to the above technical solution, step S3 further includes:

[0020] Step S31: the information relevance detection submodule randomly selects a non-linked preference reference information from the preference information, and obtains the relevance threshold between the preference information entered in different clients and the target reference information;

[0021] Step S32: The correlation threshold value X is in the time coefficient model. The total volume of the current time coefficient model V = AHΔT. When the volume of the time coefficient model is higher than the rated value V 0 When , the time coefficient analysis module records the preference information data into the information relevance detection submodule;

[0022] Step S33: randomly select a target preference information in step S23, and obtain the user's most recent browsing time difference corresponding to the target preference information Average appearance time of target preference information The sample browsing time data corresponding to the target preference information is H G , then the coordinates of the target preference information in the rectangular coordinate system are (A G , ΔT G , H G );

[0023] Step S34: the acceptability coefficient detection submodule analyzes whether the pre-pushed preference information may infringe the user's privacy.

[0024] According to the above technical solution, in step S31, the specific method for obtaining the correlation threshold between the information entered into the preference in different clients and the target reference information includes:

[0025] After entering the history interface from the user browsing information in the same client, enter the corresponding time node of the target preference information interface through the link of the history interface, and obtain the L time nodes crossed by the user entering the preference information interface through the historical order interface for the second time. If it is detected that the L time nodes completely include the K time nodes in step S23, the association threshold is set to η1%; if it is detected that the L time nodes do not include the K time nodes in step S23 at all, the association threshold is set to η2%; if it is detected that the L time nodes partially include the K time nodes in step S23, the association threshold is set to η3%, wherein η1, η2, and η3 are respectively the ratio values ​​of the nodes crossed when entering the history interface under different inclusion conditions within the K time nodes.

[0026] According to the above technical solution, in step S33, the possible analysis method of whether the pre-pushed preference information will cause privacy infringement on the user specifically includes:

[0027] Step S331: In the spatial rectangular coordinate system constructed in step S2, the target preference information detected by the system has a probability of infringing the user's privacy. Where λ is the correlation threshold ratio value;

[0028] Step S332: When the probability of the target preference information violating user privacy is higher than the safety limit value, it is determined that the current type of preference information is not suitable for continued push; otherwise, it is determined that the current type of preference information is suitable for continued push, and the corresponding judgment result is input into the output module.

[0029] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention extracts and processes the time node related information of the user's browsing records through the time coefficient analysis module, and successively uses the user's most recent browsing time difference A, the average appearance time ΔT and the difference H of the sampled browsing time data as the X, Y and Z axes in the spatial rectangular coordinate system of the time coefficient model, and stereoscopically visualizes the user's browsing parameters for preference information, thereby improving the detection dimension of user privacy protection, and increasing the accuracy and humanization of the detection; through the information disclosure analysis module, the time coefficient model constructed by the user's browsing records is approximated by the detection model area, and by extracting random time nodes within the monitoring time period, it not only has sufficient randomness, but also can ensure that the extracted time nodes are within the constructed time coefficient model, thereby ensuring the accuracy of the random inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0031] Figure 1 It is a schematic diagram of the system module composition of the present invention. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] See also Figure 1 The present invention provides a technical solution: a server privacy protection system based on big data, comprising:

[0034] Data collection module, data analysis module and output module. The data collection module is used to obtain the user preference data results analyzed in the client; the data analysis module is used to analyze and judge whether the preference data has a significant infringement on the user's privacy; the output module is used to test and feedback the data information pushed to the user, and improve the parameters of the push mode. The data collection module, data analysis module and output module are connected to each other in communication.

[0035] The present invention extracts and processes the time node related information of the user's browsing record through the time coefficient analysis module, and sequentially uses the user's most recent browsing time difference A, the average appearance time ΔT and the difference H of the sampled browsing time data as the X, Y and Z axes in the spatial rectangular coordinate system of the time coefficient model, and stereoscopically visualizes the user's browsing parameters for preference information, thereby improving the detection dimension of user privacy protection and increasing the accuracy and humanization of the detection; through the information disclosure analysis module, the time coefficient model constructed by the user's browsing record is approximated by the detection model area, and by extracting random time nodes within the monitoring time period, it not only has sufficient randomness, but also can ensure that the extracted time nodes are within the constructed time coefficient model, thereby ensuring the accuracy of the random inspection.

[0036] The data collection module includes a preference data extraction module and a user information collection module. The preference data extraction module is used to extract the user preference information collected by the client; the user information collection module is used to obtain the user's personal information.

[0037] The data analysis module includes a time coefficient analysis module and an information disclosure analysis module. The time coefficient analysis module is used to obtain the timeliness parameters of user information collected by the client; the information disclosure analysis module is used to analyze the rationality of pushing public target information to users.

[0038] The information disclosure analysis module further includes an information relevance detection submodule and an acceptability coefficient detection submodule. The information relevance detection submodule is used to obtain the relevance parameters of information retrieved by users on various platforms; the acceptability coefficient detection submodule is used to determine whether the current type of preference information is suitable for continued push.

[0039] The output module includes an information shielding module and a pushing module. The information shielding module is used to output instructions for shielding and pushing such information; the pushing module is used to output instructions for pushing such information.

[0040] In a preferred embodiment, the operation method of the server privacy protection system mainly includes the following steps:

[0041] Step S1: accessing the server privacy protection system in the client to extract the user preference information and user personal information data obtained by the client;

[0042] Step S2: The system analyzes the time coefficient of the user's preference data and the information relevance of the browsing data, and predicts the user's acceptance of the disclosure of personal information;

[0043] Step S3: the access test module pushes the target information to the user, and the feedback result submodule determines the user's feedback to determine whether the current type of preference information is suitable for continued push;

[0044] Step S4: The system implements shielding and pushing instructions for the preference information through the information shielding module and the pushing module respectively.

[0045] In this embodiment, step S2 further includes:

[0046] Step S21: The time coefficient analysis module records the time node data of the preference information obtained by the client, and obtains the elapsed time value of each detected time node from the current time, which is arranged from low to high as t 1 ,t 2 ……t n+1 , where t 1 Indicates the time from the last time the system pushed preference information to the user to now, which is the browsing interval of the user's preference data. Where n is the number of intervals between recorded time nodes, then the time difference between the last browsing of the user of this data The unit of A is hours;

[0047] Step S22: The time coefficient analysis module randomly selects two end and start time nodes from the time nodes obtained in step S21, sets the interval between the two selected time nodes as the monitoring time period, and records the number of time nodes that appear within the monitoring time period. The average appearance time of the preference information once within the monitoring time period is Where T 0 is the time interval of the monitoring period, T 0 The unit is hour, K is the number of times the preference information appears in the monitoring time period; Step S23: The time coefficient analysis module records the K browsing information of the preference information appearing in the client in the monitoring time period obtained in Step S22, and the browsing time data of the K browsing information, and obtains the minimum value T of the browsing information time 1 and the maximum browsing time T 2 , calculate the browsing time difference T 差 =T 2 -T 1 ,Will The record is the average value of the sampled browsing time data, where the browsing information time T is in seconds, and φ is the parameter multiple matching value of the component time coefficient model;

[0048] Step S24: Establish a spatial rectangular coordinate system with the user browsing time difference record as the X-axis of the time coefficient model, the browsing record interval time record within the monitoring time period as the Y-axis of the time coefficient model, and the sampled browsing time data record as the Z-axis of the time coefficient model, wherein the coordinate origin is (0, 0), the length of the positive half axis of the X-axis is the user's most recent browsing time difference A, the length of the positive half axis of the Y-axis is the average appearance time ΔT, and the length of the positive half axis of the Z-axis is the difference H of the sampled browsing time data.

[0049] This technical solution solves the problem that the three time coefficients obtained in the process of pushing information to users are too independent and have low correlation, and that simple linearity cannot achieve accurate user privacy monitoring. By extracting and processing the relevant information of the time nodes of the user's browsing records, the user's most recent browsing time difference A, the average appearance time ΔT, and the difference H of the sampled browsing time data are used as the X, Y, and Z axes in the spatial rectangular coordinate system of the time coefficient model, and the user's browsing parameters for preference information are visualized in three dimensions, which improves the detection dimension of user privacy protection and increases the accuracy and humanization of detection.

[0050] In this embodiment, step S3 further includes:

[0051] Step S31: the information relevance detection submodule randomly selects a non-linked preference reference information from the preference information, and obtains the relevance threshold between the preference information entered in different clients and the target reference information;

[0052] Step S32: The correlation threshold value X is in the time coefficient model. The total volume of the current time coefficient model V = AHΔT. When the volume of the time coefficient model is higher than the rated value V 0 When , the time coefficient analysis module records the preference information data into the information relevance detection submodule;

[0053] Step S33: randomly select a target preference information in step S23, and obtain the user's most recent browsing time difference corresponding to the target preference information Average appearance time of target preference information The sample browsing time data corresponding to the target preference information is H G , then the coordinates of the target preference information in the rectangular coordinate system are (A G , ΔT G , H G );

[0054] Step S34: the acceptability coefficient detection submodule analyzes whether the pre-pushed preference information may infringe the user's privacy.

[0055] Through this technical solution, the time coefficient model constructed by the user browsing records is approximated by detecting the model area, and by extracting random time nodes within the monitoring time period, it not only has sufficient randomness, but also can ensure that the extracted time nodes are within the constructed time coefficient model, thereby ensuring the accuracy of the random inspection.

[0056] In step S31 of this embodiment, the specific method of obtaining the correlation threshold between the information entered into the preference in different clients and the target reference information includes:

[0057] After the user browses the history interface in the same client, the corresponding time node of the user entering the target preference information interface through the link of the history interface is obtained, and the L time nodes spanned by the user entering the preference information interface through the history order interface for the second time are obtained. If it is detected that the L time nodes completely include the K time nodes in step S23, the association threshold is set to η 1 %; if it is detected that the L time nodes do not contain the K time nodes in step S23 at all, the association threshold is set to η 2 %; if it is detected that the L time nodes partially include the K time nodes in step S23, the association threshold is set to η 3 %, where η 1 , η 2 , η 3 They are the ratio values ​​of entering the historical interface and crossing nodes under different inclusion conditions within K time nodes.

[0058] Through this technical solution, the problem of different correlations between information extraction and user privacy monitoring caused by the uncertain length of the information push time span of the preference message in step S2 in the preference reference message randomly selected by the system is solved. By introducing a correlation threshold, the time overlap ratio of extracting different information and the time coefficient model is analyzed, thereby ensuring the controllability of the extracted push information samples.

[0059] In step S33 of this embodiment, the possible analysis method of whether the pre-pushed preference information will cause privacy infringement on the user specifically includes:

[0060] Step S331: In the spatial rectangular coordinate system constructed in step S2, the probability of the target preference information detected by the system infringing the user's privacy Where λ is the correlation threshold ratio value;

[0061] When it is detected that L time nodes completely include the K time nodes in step S23, the value of λ is 100; when it is detected that L time nodes completely do not include the K time nodes in step S23, the value of λ is 0; if it is detected that L time nodes partially include the K time nodes in step S23, the value of λ is η.

[0062] Step S332: When the probability of the target preference information violating user privacy is higher than the safety limit value, it is determined that the current type of preference information is not suitable for continued push; otherwise, it is determined that the current type of preference information is suitable for continued push, and the corresponding judgment result is input into the output module.

[0063] This technical solution solves the problem that when the system pushes preference information to users, the user's redundant push impressions generated by the short interval of push information, the user's short-term repeated impressions generated by the time difference of the latest push messages, and the user's reaction feedback to the push messages will affect the user's privacy. As a result, the preference information is often pushed to the user meaninglessly, which in turn causes the user's solution problem. By limiting the reference point in the constructed time coefficient model to obtain the range of push messages that the user can accept, the possibility of the push information bringing adverse effects to the target user is effectively reduced, the timely feedback of human-computer interaction is enhanced, and the reliability of the system privacy protection mechanism is improved.

[0064] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0065] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. Server privacy protection system based on big data, Features: It includes a data collection module, a data analysis module and an output module. The data collection module is used to obtain the user preference data results analyzed in the client; the data analysis module is used to analyze and judge whether the preference data has a significant infringement on the user's privacy; The output module is used to test and feedback the data information pushed to the user, and to improve the parameters of the push mode. The data acquisition module, the data analysis module and the output module are connected to each other for communication.

2. According to the server privacy protection system based on big data according to claim 1, Features: The data collection module includes a preference data extraction module and a user information collection module. The preference data extraction module is used to extract user preference information collected by the client; the user information collection module is used to obtain user personal information.

3. According to the server privacy protection system based on big data according to claim 2, Features: The data analysis module includes a time coefficient analysis module and an information disclosure analysis module. The time coefficient analysis module is used to obtain the timeliness parameters of user information collected by the client; the information disclosure analysis module is used to analyze the rationality of pushing public target information to users.

4. According to the server privacy protection system based on big data according to claim 3, Features: The information disclosure analysis module further includes an information relevance detection submodule and an acceptability coefficient detection submodule. The information relevance detection submodule is used to obtain the relevance parameters of information retrieved by users in various platforms; the acceptability coefficient detection submodule is used to determine whether the current type of preference information is suitable for continued push.

5. According to the server privacy protection system based on big data according to claim 4, Features: The output module includes an information shielding module and a pushing module. The information shielding module is used to output instructions for shielding and pushing such information; and the pushing module is used to output instructions for pushing such information.

6. According to the server privacy protection system based on big data according to claim 5, Features: The operation method of the server privacy protection system mainly includes the following steps: Step S1: accessing the server privacy protection system in the client to extract the user preference information and user personal information data obtained by the client; Step S2: The system analyzes the time coefficient of the user's preference data and the information relevance of the browsing data, and predicts the user's acceptance of the disclosure of personal information; Step S3: the access test module pushes the target information to the user, and the feedback result submodule determines the user's feedback to determine whether the current type of preference information is suitable for continued push; Step S4: The system implements shielding and pushing instructions for the preference information through the information shielding module and the pushing module respectively.

7. The server privacy protection system based on big data according to claim 6, Features: The step S2 further comprises: Step S21: The time coefficient analysis module records the time node data of the preference information obtained by the client, and obtains the elapsed time value of each detected time node from the current time, which is arranged from low to high as t 1 ,t 2 ……t n+1 , then the browsing time interval of user preference data Where n is the number of intervals between recorded time nodes, then the time difference between the last browsing of the user of this data The unit of A is hours; Step S22: The time coefficient analysis module randomly selects two end and start time nodes from the time nodes obtained in step S21, sets the interval between the two selected time nodes as the monitoring time period, and records the number of time nodes that appear within the monitoring time period. The average time for preference information to appear once within the monitoring time period is Where T 0 is the time interval of the monitoring period, T 0 The unit of is hour, K is the number of preference information appearing in the monitoring time period; Step S23: The time coefficient analysis module records the K browsing information of the preference information appearing in the client within the monitoring time period obtained in step S22, and the browsing time data of the K browsing information, and obtains the minimum value T of the browsing information time 1 and the maximum browsing time T 2 , calculate the browsing time difference T 差 =T 2 -T 1 ,Will The record is the average value of the sampled browsing time data, where the browsing information time T is in seconds, and φ is the parameter multiple matching value of the component time coefficient model; Step S24: Establish a spatial rectangular coordinate system with the user browsing time difference record as the X-axis of the time coefficient model, the browsing record interval time record within the monitoring time period as the Y-axis of the time coefficient model, and the sampled browsing time data record as the Z-axis of the time coefficient model, wherein the coordinate origin is (0, 0), the length of the positive half axis of the X-axis is the user's most recent browsing time difference A, the length of the positive half axis of the Y-axis is the average appearance time ΔT, and the length of the positive half axis of the Z-axis is the difference H of the sampled browsing time data.

8. According to the server privacy protection system based on big data according to claim 7, Features: The step S3 further comprises: Step S31: the information relevance detection submodule randomly selects a non-linked preference reference information from the preference information, and obtains the relevance threshold between the preference information entered in different clients and the target reference information; Step S32: The correlation threshold value X is in the time coefficient model. The total volume of the current time coefficient model V = AHΔT. When the volume of the time coefficient model is higher than the rated value V 0 When , the time coefficient analysis module records the preference information data into the information relevance detection submodule; Step S33: randomly select a target preference information in step S23, and obtain the user's most recent browsing time difference corresponding to the target preference information Average appearance time of target preference information The sample browsing time data corresponding to the target preference information is H G , then the coordinates of the target preference information in the spatial rectangular coordinate system are (A G , ΔT G , H G ); Step S34: the acceptability coefficient detection submodule analyzes whether the pre-pushed preference information may infringe the user's privacy.

9. The server privacy protection system based on big data according to claim 8, Features: In step S31, the specific method for obtaining the correlation threshold between the information entered into the preference list in different clients and the target reference information includes: After entering the history interface from the user browsing information in the same client, enter the corresponding time node of the target preference information interface through the link of the history interface, and obtain the L time nodes crossed by the user entering the preference information interface through the historical order interface for the second time. If it is detected that the L time nodes completely include the K time nodes in step S23, the association threshold is set to η1%; if it is detected that the L time nodes do not completely include the K time nodes in step S23, the association threshold is set to η2%; if it is detected that the L time nodes partially include the K time nodes in step S23, the association threshold is set to η3%, wherein η1, η2, and η3 are respectively the ratio values ​​of the nodes crossed when entering the history interface under different inclusion conditions within the K time nodes.

10. The server privacy protection system based on big data according to claim 9, Features: In step S33, the possible analysis method of whether the pre-pushed preference information will cause privacy infringement on the user specifically includes: Step S331: In the spatial rectangular coordinate system constructed in step S2, the probability of the target preference information detected by the system infringing the user's privacy Where λ is the correlation threshold ratio value; Step S332: When the probability of the target preference information violating user privacy is higher than the safety limit value, it is determined that the current type of preference information is not suitable for continued push; otherwise, it is determined that the current type of preference information is suitable for continued push, and the corresponding judgment result is input into the output module.

Images