A data sharing method and system based on data desensitization

Abstract

This invention relates to the field of computer technology, and in particular to a data sharing method and system based on data anonymization. This method uses a sequence labeling model to predict the field categories of an initial word sequence, and analyzes the initial field sequence in conjunction with preset fields to achieve a quantitative assessment of the importance of the initial fields. This avoids the blindness of traditional rule-based anonymization methods, balancing data security and data availability. Priority is set on an initial word unit, and the anonymization granularity is dynamically adjusted based on the terminal's preset permission level. High-privilege terminals can obtain complete data, while low-privilege terminals can only obtain non-sensitive initial words, thus protecting data privacy while ensuring data availability across different terminals.

Description

Technical Field

[0001] This invention relates to the field of data transmission technology, and in particular to a data sharing method and system based on data anonymization. Background Technology

[0002] Currently, in data sharing and transmission scenarios, text data often contains critical information such as personal privacy information and commercially sensitive data. Personal privacy information includes ID card numbers and bank accounts, while commercially sensitive data includes financial statements and customer information.

[0003] Traditional data anonymization methods often employ rule-based processing, such as direct replacement or simple masking. However, rule-based processing methods often fail to accurately distinguish between ordinary fields and key sensitive fields in the text, resulting in insufficient field recognition accuracy. This leads to either over- or under-anonymization of data, reducing data usability and security.

[0004] In addition, rule-based processing methods are difficult to flexibly adjust de-identification strategies according to different data scenarios and terminal permission levels, and have poor generalization ability, that is, poor ability to dynamically adapt to scenarios, making it difficult to achieve a balance between data privacy protection and data sharing availability.

[0005] Therefore, how to improve the security of data sharing without affecting data availability has become an urgent problem to be solved. Summary of the Invention

[0006] To address the aforementioned technical problems, the present invention provides a data sharing method based on data anonymization, which includes the following steps:

[0007] S101, the acquired text data to be transmitted is split into tokens to obtain an initial token sequence containing M initial tokens, where M is a positive integer and the priority of each initial token is set to the first preset value.

[0008] S102, the initial word sequence is input into the trained sequence labeling model to obtain the first predicted field category and the first predicted probability corresponding to the M initial words.

[0009] S103, based on the first prediction field category corresponding to the M initial words, form an initial field sequence containing N initial fields from the M initial words, where N is a positive integer less than M.

[0010] S104. Based on the preset fields and the initial field sequence, obtain the importance evaluation value corresponding to each initial field.

[0011] S105. Based on the importance evaluation value corresponding to each initial field, select several temporary key fields from all initial fields.

[0012] S106. Determine several target key fields based on each temporary key field and several reference key fields in the preset set of reference key fields.

[0013] S107, set the priority of each initial word in each target key field to the second preset value.

[0014] S108, when any terminal requests to obtain the text data to be transmitted, according to the preset permission level corresponding to the terminal, each initial word corresponding to the priority level less than or equal to the preset permission level is sent to the terminal.

[0015] The present invention also provides a data sharing system based on data anonymization, the data sharing system based on data anonymization comprising:

[0016] The word segmentation module is used to segment the acquired text data to be transmitted into words to obtain an initial word sequence containing M initial words, where M is a positive integer, and the priority of each initial word is set to the first preset value.

[0017] The category prediction module is used to input the initial word sequence into the trained sequence labeling model to obtain the first prediction field category and the first prediction probability corresponding to the M initial words.

[0018] The field formation module is used to form an initial field sequence containing N initial fields from the M initial words according to the first predicted field categories corresponding to the M initial words, where N is a positive integer less than M.

[0019] The field evaluation module is used to obtain the importance evaluation value corresponding to each initial field based on the preset fields and the initial field sequence.

[0020] The field filtering module is used to filter out several temporary key fields from all the initial fields based on the importance evaluation value corresponding to each initial field.

[0021] The field determination module is used to determine several target key fields based on various temporary key fields and several reference key fields in a preset set of reference key fields.

[0022] The priority update module is used to set the priority of each initial word in each target key field to the second preset value.

[0023] The data sharing module is used to send each initial word corresponding to a priority level less than or equal to the preset permission level to any terminal when the terminal requests to obtain the text data to be transmitted.

[0024] This invention has at least the following beneficial effects: by using a sequence labeling model to predict the field category of the initial word sequence, and combining it with preset fields to analyze the initial field sequence, it achieves a quantitative assessment of the importance of the initial fields, avoiding the blindness of traditional rule-based desensitization methods, and taking into account both data security and data availability. Priority is set on the basis of initial words, and the desensitization granularity is dynamically adjusted based on the preset permission level of the terminal. High-privilege terminals can obtain complete data, while low-privilege terminals can only obtain non-sensitive initial words, thus protecting data privacy while ensuring data availability for different terminals. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a flowchart illustrating a data sharing method based on data anonymization provided in Embodiment 1 of the present invention;

[0027] Figure 2 This is a schematic diagram of the structure of a data sharing system based on data anonymization provided in Embodiment 2 of the present invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It is understood that, where appropriate, the terms used to distinguish similar objects can be interchanged so that the invention can also be implemented in other embodiments besides the illustrated or described embodiments. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices.

[0030] Example 1

[0031] This first embodiment provides a data sharing method based on data anonymization, such as... Figure 1 The diagram shown is a flowchart of a data sharing method based on data anonymization provided in Embodiment 1 of the present invention. This data sharing method based on data anonymization includes the following steps:

[0032] S101, the acquired text data to be transmitted is split into words to obtain an initial word sequence containing M initial words, where M is a positive integer and the priority of each initial word is set to the first preset value.

[0033] S102, the initial word sequence is input into the trained sequence labeling model to obtain the first predicted field category and the first predicted probability corresponding to the M initial words respectively;

[0034] S103, based on the first prediction field category corresponding to the M initial words, form an initial field sequence containing N initial fields from the M initial words, where N is a positive integer less than M;

[0035] S104, Based on the preset fields and the initial field sequence, obtain the importance evaluation value corresponding to each initial field;

[0036] S105, Based on the importance evaluation value corresponding to each initial field, select several temporary key fields from all initial fields;

[0037] S106, Based on each temporary key field and several reference key fields in the preset set of reference key fields, determine several target key fields;

[0038] S107, set the priority of each initial word element in each target key field to the second preset value;

[0039] S108, when any terminal requests to obtain the text data to be transmitted, according to the preset permission level corresponding to the terminal, each initial word corresponding to the priority level less than or equal to the preset permission level is sent to the terminal.

[0040] The text data to be transmitted can refer to data to be shared with multiple terminals, and the word unit can refer to a token. Word unit splitting can be done using N-gram splitting method, dictionary-based splitting method, statistical model-based splitting method, etc., without any restrictions.

[0041] The M initial tokens are sorted according to their positions in the text data to be transmitted to obtain the initial token sequence.

[0042] The priority of an initial lexical unit can be used to indicate the degree of desensitization required for that initial lexical unit, or it can be understood as the semantic importance of that initial lexical unit.

[0043] Sequence labeling models can employ recurrent neural network models, long short-term memory network models, Transformer models, etc., and there are no restrictions here. The training process of sequence labeling models will not be elaborated here.

[0044] The first predicted field category can belong to a set of preset field categories. The set of preset field categories can include several preset field categories, such as location, amount, name, etc. The first predicted probability can represent the probability that the corresponding initial word belongs to its corresponding first predicted field category.

[0045] A single initial field can include multiple initial lexical units corresponding to the same first prediction field category. Similarly, N initial fields are sorted according to their positions in the text data to be transmitted to obtain an initial field sequence.

[0046] Preset fields can be used to mask the initial fields in the initial field sequence, and the importance evaluation value corresponding to each initial field can be determined by masking reconstruction analysis.

[0047] Temporary key fields can refer to key fields obtained from occlusion reconstruction analysis, while reference key fields can refer to key fields set by the implementer based on prior information.

[0048] The target key fields can refer to the key fields determined by comprehensive analysis and prior knowledge.

[0049] A terminal can refer to a terminal that has access to the shared text data to be transmitted. Based on the preset permission level and the priority of each initial word, the data anonymization result of the corresponding terminal can be determined, and then the data anonymization result can be sent to the corresponding terminal.

[0050] In one specific implementation, the step of forming an initial field sequence containing N initial fields from the M initial words according to the first predicted field categories corresponding to the M initial words includes:

[0051] Initialize the word identifier i=1, and initialize the field identifier j=1;

[0052] When the category of the first predicted field corresponding to the i-th initial word is the same as the category of the first predicted field corresponding to the (i-1)-th initial word, the i-th initial word is added to the initial field corresponding to the (i-1)-th initial word.

[0053] When the category of the first predicted field corresponding to the i-th initial word is different from the category of the first predicted field corresponding to the (i-1)-th initial word, or when i=1, the j-th initial field is constructed from the i-th initial word, and j=j+1 is updated;

[0054] Update i=i+1, return to the step of adding the i-th initial word to the initial field corresponding to the (i-1)-th initial word when the category of the first predicted field corresponding to the i-th initial word is the same as the category of the first predicted field corresponding to the (i-1)-th initial word, until i=M, determine the value of j-1 to N, and obtain N initial fields;

[0055] The initial field sequence is formed from the N initial fields.

[0056] In one specific implementation, obtaining the importance evaluation value corresponding to each initial field based on the preset fields and the initial field sequence includes:

[0057] For any initial field, replace the initial field with the preset field in the initial field sequence to obtain the intermediate field sequence;

[0058] The intermediate field sequence is input into the trained reconstruction model to obtain the reconstruction field sequence;

[0059] The difference is calculated based on the reconstructed field sequence and the initial field sequence to obtain the difference calculation result.

[0060] The difference calculation result is mapped to the importance evaluation value corresponding to the initial field.

[0061] Compared to the field occlusion method of directly deleting the initial field, setting a preset field aims to make the trained reconstruction model aware that the preset field needs to be reconstructed.

[0062] The reconstruction model can include convolutional modules and deconvolutional modules. The convolutional module is used to extract feature information from the intermediate field sequence of the input, and the deconvolutional module is used to reconstruct the reconstructed field sequence based on the feature information. The reconstruction model can adopt Transformer model, BERT model, BART model, etc. The training process of the reconstruction model will not be described in detail here.

[0063] Specifically, the difference between the reconstructed field sequence and the initial field sequence can be calculated using cosine distance. The mapping of the difference calculation results can be performed using the softmax function based on the difference calculation results corresponding to each initial field, thereby obtaining the importance evaluation value corresponding to each initial field.

[0064] In one specific implementation, determining several target key fields based on various temporary key fields and several reference key fields in a preset set of reference key fields includes:

[0065] Obtain several sets of basic key fields, each set of basic key fields corresponding to a sequence of basic field categories;

[0066] Based on the first prediction field category corresponding to the M initial lexical units, a prediction field category sequence is formed;

[0067] The set of basic key fields corresponding to the basic field category sequence that is the same as the predicted field category sequence is determined as the reference key field set;

[0068] Based on each temporary key field and several reference key fields in the set of reference key fields, several target key fields are determined.

[0069] Specifically, the first prediction field category corresponding to each of the M initial words is sorted according to their positions in the text data to be transmitted, thus obtaining the prediction field category sequence.

[0070] Specifically, this embodiment determines a set of reference key fields by comparing the predicted field category sequence with the basic field category sequence, thereby determining reference key fields that conform to the field category order. This can adapt to the expression of the text, reduce the number of reference key fields, and make the subsequently determined target key fields more reliable.

[0071] In one specific implementation, determining several target key fields based on each temporary key field and several reference key fields in the set of reference key fields includes:

[0072] A set of temporary key fields is formed from the various temporary key fields;

[0073] Calculate the union of the temporary key field set and the reference key field set, and use each temporary key field contained in the union as the target key field.

[0074] In this set, any temporary key field belongs to both the temporary key field set and the reference key field set.

[0075] In one specific implementation, the second preset value is greater than the first preset value, and the preset permission level is either the first preset value or the second preset value.

[0076] The first preset value can be 1, and the second preset value can be 2.

[0077] In one specific implementation, after step S107 and before step S108, the following steps are also included:

[0078] For any target key field, a reference word is randomly determined from the initial word elements contained in the target key field. The initial word elements corresponding to the reference word elements are updated with preset word elements in the initial word element sequence. Then, the initial word elements are input into the trained sequence labeling model to obtain the second predicted field category and the second predicted probability corresponding to M initial word elements respectively.

[0079] Determine the first field category and the first classification probability corresponding to the initial field to which the reference word belongs;

[0080] Determine the second field category and second classification probability of the preset word element corresponding to the reference word element;

[0081] If the first field category and the second field category are different, the priority corresponding to the reference term is set to a third preset value;

[0082] If the first field category and the second field category are the same, and the absolute value of the difference between the first classification probability and the second classification probability is greater than a preset probability threshold, then the priority corresponding to the reference word is set to a third preset value.

[0083] Otherwise, the priority corresponding to the reference term will not be updated.

[0084] Among them, the first field category corresponding to the initial field to which the reference word belongs is the first predicted field category corresponding to the reference word, and the first classification probability can be calculated based on the average of the first predicted probabilities corresponding to all initial words in the initial field to which the reference word belongs.

[0085] The second field category of the initial field to which the reference word belongs is the second prediction field category of the preset word. The first classification probability can be calculated based on the average of the second prediction probabilities corresponding to all initial words in the initial field to which the preset word belongs.

[0086] Specifically, when the first field category and the second field category are different, it indicates that the replacement of the reference word has caused a significant semantic change. Therefore, the reference word is considered to be of high importance, and the priority corresponding to the reference word is set to the third preset value.

[0087] When the first field category and the second field category are the same, and the absolute value of the difference between the first category probability and the second category probability is greater than the preset probability threshold, it also indicates that the replacement of the reference word causes a significant semantic change. Therefore, the reference word is considered to be of high importance, and the priority corresponding to the reference word is set to the third preset value.

[0088] In one specific implementation, the third preset value is greater than the second preset value, and the second preset value is greater than the first preset value.

[0089] The first preset value can be 1, the second preset value can be 2, and the third preset value can be 3.

[0090] In one specific implementation, the preset permission level is the first preset value, the second preset value, or the third preset value;

[0091] The step of sending initial lexical units corresponding to priorities less than or equal to the preset permission level to the terminal includes:

[0092] Based on the preset permission level corresponding to the terminal, each initial word element corresponding to a priority level greater than the preset permission level is replaced with a preset word element in the initial word element sequence to obtain the target word element sequence;

[0093] The target word sequence is sent to the terminal.

[0094] The target word sequence is the data anonymization result of the corresponding terminal.

[0095] This embodiment predicts the field category of the initial word sequence using a sequence labeling model, and analyzes the initial field sequence in conjunction with preset fields to achieve a quantitative assessment of the importance of the initial fields. This avoids the blindness of traditional rule-based desensitization methods and takes into account both data security and data availability. Priority is set on the basis of initial words, and the desensitization granularity is dynamically adjusted based on the preset permission level of the terminal. High-privilege terminals can obtain complete data, while low-privilege terminals can only obtain non-sensitive initial words, thus protecting data privacy while ensuring data availability for different terminals.

[0096] Example 2

[0097] This second embodiment provides a data sharing system based on data anonymization, such as... Figure 2 The diagram shown is a structural schematic of a data sharing system based on data anonymization provided in Embodiment 2 of the present invention. This data sharing system based on data anonymization includes:

[0098] The word segmentation module 201 is used to segment the acquired text data to be transmitted into words to obtain an initial word sequence containing M initial words, where M is a positive integer, and the priority of each initial word is set to the first preset value.

[0099] Category prediction module 202 is used to input the initial word sequence into the trained sequence labeling model to obtain the first prediction field category and the first prediction probability corresponding to the M initial words respectively;

[0100] The field forming module 203 is used to form an initial field sequence containing N initial fields from the M initial words according to the first predicted field categories corresponding to the M initial words, where N is a positive integer less than M;

[0101] The field evaluation module 204 is used to obtain the importance evaluation value corresponding to each initial field based on the preset fields and the initial field sequence.

[0102] The field filtering module 205 is used to filter out several temporary key fields from all the initial fields based on the importance evaluation value corresponding to each initial field.

[0103] The field determination module 206 is used to determine several target key fields based on each temporary key field and several reference key fields in a preset set of reference key fields;

[0104] Priority update module 207 is used to set the priority of each initial word element in each target key field to the second preset value.

[0105] The data sharing module 208 is used to send each initial word corresponding to a priority level less than or equal to the preset permission level to any terminal when the terminal requests to obtain the text data to be transmitted.

[0106] It should be noted that the specific limitations of the data sharing system based on data anonymization can be found in the limitations of the data sharing method based on data anonymization mentioned above, and will not be repeated here. The information interaction and execution process between the above modules are based on the same concept as the method embodiments of this invention, and their specific functions and technical effects can be found in the method embodiments section, and will not be repeated here.

[0107] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A data sharing method based on data anonymization, characterized in that, The data sharing method based on data anonymization includes the following steps: S101, the acquired text data to be transmitted is split into words to obtain an initial word sequence containing M initial words, where M is a positive integer and the priority of each initial word is set to the first preset value. S102, the initial word sequence is input into the trained sequence labeling model to obtain the first predicted field category and the first predicted probability corresponding to the M initial words respectively; S103, based on the first prediction field category corresponding to the M initial words respectively, form an initial field sequence containing N initial fields from the M initial words, where N is a positive integer less than M; S104, based on the preset fields and the initial field sequence, obtain the importance evaluation value corresponding to each initial field, wherein the preset fields are used to occlude the initial fields in the initial field sequence, so as to determine the importance evaluation value corresponding to each initial field through occlusion reconstruction analysis. Obtaining the importance evaluation value corresponding to each initial field based on the preset fields and the initial field sequence includes: For any initial field, replace the initial field with the preset field in the initial field sequence to obtain the intermediate field sequence; The intermediate field sequence is input into the trained reconstruction model to obtain the reconstruction field sequence; The difference is calculated based on the reconstructed field sequence and the initial field sequence to obtain the difference calculation result. The difference calculation result is mapped to the importance evaluation value corresponding to the initial field; S105, Based on the importance evaluation value corresponding to each initial field, select several temporary key fields from all initial fields; S106, Based on each temporary key field and several reference key fields in the preset set of reference key fields, determine several target key fields; S107, set the priority of each initial word element in each target key field to the second preset value; S108, when any terminal requests to obtain the text data to be transmitted, according to the preset permission level corresponding to the terminal, each initial word corresponding to the priority level less than or equal to the preset permission level is sent to the terminal.

2. The data sharing method based on data anonymization according to claim 1, characterized in that, The step of forming an initial field sequence containing N initial fields from the M initial words according to the first predicted field categories corresponding to the M initial words includes: Initialize the word identifier i=1, and initialize the field identifier j=1; When the category of the first predicted field corresponding to the i-th initial word is the same as the category of the first predicted field corresponding to the (i-1)-th initial word, the i-th initial word is added to the initial field corresponding to the (i-1)-th initial word. When the category of the first predicted field corresponding to the i-th initial word is different from the category of the first predicted field corresponding to the (i-1)-th initial word, or when i=1, the j-th initial field is constructed from the i-th initial word, and j=j+1 is updated; Update i=i+1, return to the step of adding the i-th initial word to the initial field corresponding to the (i-1)-th initial word when the category of the first predicted field corresponding to the i-th initial word is the same as the category of the first predicted field corresponding to the (i-1)-th initial word, until i=M, determine the value of j-1 to N, and obtain N initial fields; The initial field sequence is formed from the N initial fields.

3. The data sharing method based on data anonymization according to claim 1, characterized in that, The step of determining several target key fields based on various temporary key fields and several reference key fields in a preset set of reference key fields includes: Obtain several sets of basic key fields, each set of basic key fields corresponding to a sequence of basic field categories; Based on the first prediction field category corresponding to the M initial lexical units, a prediction field category sequence is formed; The set of basic key fields corresponding to the basic field category sequence that is the same as the predicted field category sequence is determined as the reference key field set; Based on each temporary key field and several reference key fields in the set of reference key fields, several target key fields are determined.

4. The data sharing method based on data anonymization according to claim 3, characterized in that, The step of determining several target key fields based on each temporary key field and several reference key fields in the set of reference key fields includes: A set of temporary key fields is formed from the various temporary key fields; Calculate the union of the temporary key field set and the reference key field set, and use each temporary key field contained in the union as the target key field.

5. The data sharing method based on data anonymization according to claim 1, characterized in that, The second preset value is greater than the first preset value, and the preset permission level is either the first preset value or the second preset value.

6. The data sharing method based on data anonymization according to claim 1, characterized in that, After step S107 and before step S108, the following steps are also included: For any target key field, a reference word is randomly determined from the initial word elements contained in the target key field. The initial word elements corresponding to the reference word elements are updated with preset word elements in the initial word element sequence. Then, the initial word elements are input into the trained sequence labeling model to obtain the second predicted field category and the second predicted probability corresponding to M initial word elements respectively. If the first prediction field category corresponding to the reference word is different from the second prediction field category corresponding to the preset word, then the priority corresponding to the reference word is set to the third preset value. If the first predicted field category corresponding to the reference word is the same as the second predicted field category corresponding to the preset word, and the absolute value of the difference between the first predicted probability corresponding to the reference word and the second predicted probability corresponding to the preset word is greater than the preset probability threshold, then the priority corresponding to the reference word is set to the third preset value. Otherwise, the priority corresponding to the reference term will not be updated.

7. The data sharing method based on data anonymization according to claim 6, characterized in that, The third preset value is greater than the second preset value, and the second preset value is greater than the first preset value.

8. The data sharing method based on data anonymization according to claim 7, characterized in that, The preset permission level is the first preset value, the second preset value, or the third preset value; The step of sending initial lexical units corresponding to priorities less than or equal to the preset permission level to the terminal includes: Based on the preset permission level corresponding to the terminal, each initial word element corresponding to a priority level greater than the preset permission level is replaced with a preset word element in the initial word element sequence to obtain the target word element sequence; The target word sequence is sent to the terminal.

9. A data sharing system based on data anonymization, characterized in that, The data sharing system based on data anonymization includes: The word segmentation module is used to segment the acquired text data to be transmitted into words to obtain an initial word sequence containing M initial words, where M is a positive integer, and the priority of each initial word is set to the first preset value. The category prediction module is used to input the initial word sequence into the trained sequence labeling model to obtain the first prediction field category and the first prediction probability corresponding to the M initial words respectively; The field formation module is used to form an initial field sequence containing N initial fields from the M initial words according to the first predicted field categories corresponding to the M initial words, where N is a positive integer less than M; The field evaluation module is used to obtain the importance evaluation value corresponding to each initial field based on preset fields and the initial field sequence. The preset fields are used to occlude the initial fields in the initial field sequence to determine the importance evaluation value corresponding to each initial field through occlusion reconstruction analysis. Obtaining the importance evaluation value corresponding to each initial field based on the preset fields and the initial field sequence includes: For any initial field, replace the initial field with the preset field in the initial field sequence to obtain the intermediate field sequence; The intermediate field sequence is input into the trained reconstruction model to obtain the reconstruction field sequence; The difference is calculated based on the reconstructed field sequence and the initial field sequence to obtain the difference calculation result. The difference calculation result is mapped to the importance evaluation value corresponding to the initial field; The field filtering module is used to filter out several temporary key fields from all the initial fields based on the importance evaluation value corresponding to each initial field. The field determination module is used to determine several target key fields based on various temporary key fields and several reference key fields in a preset set of reference key fields; The priority update module is used to set the priority of each initial word element in each target key field to the second preset value. The data sharing module is used to send each initial word corresponding to a priority level less than or equal to the preset permission level to any terminal when the terminal requests to obtain the text data to be transmitted.

Images