Financial data security management system and method based on artificial intelligence

Abstract

The invention discloses a financial data security management system and method based on artificial intelligence, and relates to the technical field of financial data management. A data acquisition module is responsible for collecting historical financial data and imported financial files; the data preprocessing module preprocesses the collected data; the data analysis module adopts a machine learning algorithm, predicts the importance degree and the safety degree of imported data according to a historical data training model, and calculates a backup coefficient and a safety degree score; the sensitive data management module carries out fine-grained analysis on the data by combining a natural language processing technology based on a sensitive data identification model, automatically marks the sensitive data, and stores a sensitive data list and a mapping relation thereof by using a database technology; the data backup reminding module sets a threshold value and judges whether a user needs to be reminded to carry out data backup or not according to the backup coefficient; the sensitive data destroying module destroys the sensitive data to ensure that the data is unrecoverable, and the validity of the destroying process is ensured through multi-time iteration processing and destroying effect evaluation.

Description

Technical Field

[0001] The present invention belongs to the technical field of financial data management, and specifically relates to a financial data security management system and method based on artificial intelligence. Background Art

[0002] With the rapid development of information technology, enterprises increasingly rely on electronic systems to manage and process financial data. Financial data not only involves the economic activities of enterprises, but also contains a large amount of sensitive information, such as customer information, transaction records, financial statements, etc. Therefore, ensuring the security and integrity of financial data has become an important challenge facing enterprises.

[0003] At present, many companies have the following problems in financial data management:

[0004] Insufficient data security: Traditional data security management methods mainly rely on manual review and manual backup, which are prone to omissions and errors and cannot fully cover all data types.

[0005] Inaccurate data importance assessment: Existing methods usually assess the importance of data based on simple rules or experience, which lacks scientificity and accuracy.

[0006] Inaccurate identification of sensitive data: Traditional sensitive data identification methods rely on keyword matching or simple pattern recognition, which are difficult to deal with complex text content and multiple data types.

[0007] Unreasonable data backup strategy: Many companies lack an effective data backup strategy, resulting in the inability to restore data in a timely manner when data is lost or damaged.

[0008] Incomplete data destruction: Existing data destruction methods are often not thorough enough, and there is a risk of data residue, which may lead to the leakage of sensitive information. Summary of the invention

[0009] In order to overcome the shortcomings and deficiencies of the above-mentioned prior art, the first purpose of the present invention is to provide a financial data security management system based on artificial intelligence; the second purpose of the present invention is to provide a financial data security management method based on artificial intelligence.

[0010] The first object of the present invention adopts the following technical solution:

[0011] Financial data security management system based on artificial intelligence, including data collection module, data preprocessing module, data analysis module, sensitive data management module, data backup reminder module and sensitive data destruction module;

[0012] Data collection module: responsible for collecting historical financial data and imported financial files;

[0013] Data preprocessing module: used to preprocess the collected data, including cleaning, format conversion, etc., to ensure data quality;

[0014] Data analysis module: uses machine learning algorithms to train models based on historical data to predict the importance and security of imported data and calculate backup coefficients and security scores;

[0015] Sensitive data management module: Based on the sensitive data identification model of deep learning, it combines natural language processing technology to perform fine-grained analysis of data, automatically mark sensitive data, and use database technology to store sensitive data lists and their mapping relationships;

[0016] Data backup reminder module: Set the threshold and determine whether to remind the user to back up data based on the backup coefficient;

[0017] Sensitive data destruction module: Destroy sensitive data to ensure that the data is irrecoverable, and ensure the effectiveness of the destruction process through multiple iterative processing and evaluation of the destruction effect.

[0018] The second object of the present invention adopts the following technical solution:

[0019] The financial data security management method based on artificial intelligence is used to implement the financial data security management system based on artificial intelligence. The process is as follows:

[0020] Collect data, obtain historical financial data and historical retrieved data on the corresponding device, obtain the type and content data of historical financial data, and at the same time obtain the type and content data of historical retrieved data. The types of data include documents, pictures and videos; the content of data includes the text content of the document, the pixel value content of the picture and the frame image data content of each frame of the video; obtain the type and content data of the imported financial files; calculate data importance; calculate data security; tag sensitive data; count sensitive data; backup important data and security data; and destroy sensitive data.

[0021] Preferably, the data importance is calculated as follows:

[0022] S21: Import the acquired category data of historical financial data and the category data of imported financial files into the first backup coefficient calculation formula to calculate the first backup coefficient C 1 Calculation of the first backup coefficient B 1 The calculation formula is as follows: Where N is the number of historical backup data; C i is the number of types of data in the i-th historical backup data that are the same as the data types in the import file; T i is the number of data types in the i-th historical backup data;

[0023] S22: Calculate the second backup coefficient based on the historical financial data and the content of the imported financial file; import the acquired content data of the historical financial data and the content data of the imported financial file into the second backup coefficient calculation formula to calculate the second backup coefficient B 2 Calculation of the second backup coefficient B 2 The calculation formula is as follows: Where S() is the number of content data in brackets; K i is the content data in the i-th historical backup data; K is the content data in the imported file; ∪ is the union; ∩ is the intersection;

[0024] S23: Calculate the data importance by combining the first backup coefficient and the second backup coefficient; substitute the calculated first backup coefficient and the second backup coefficient into the data importance calculation formula to calculate the data importance I, the data importance I calculation formula is as follows: I = α·B 1 +(1-α)B 2 ; Among them, α is the first accounting coefficient; 1-α is the second accounting coefficient; 0≤α≤1.

[0025] Preferably, the data security degree is calculated as follows:

[0026] The types, content data, retrieval times data of the historical retrieved data, and the types and content data of the imported financial file data are imported into the data security calculation formula to calculate the data security S. The data security S calculation formula is: Among them, f j represents the number of times the i-th historical retrieved data is retrieved; K j represents the content data of the jth historical retrieved data; M represents the number of historical retrieved data.

[0027] Preferably, the sensitive data is marked as follows:

[0028] Using word segmentation technology, the data in the target financial data file is segmented into sentences according to punctuation marks;

[0029] Perform data division for each sentence, divide the target financial data file into multiple data units, and obtain a data unit list of the target financial data file;

[0030] Input all data units included in the data unit list of the target financial data file into each sensitive data recognition model corresponding to the pre-trained sensitive data type set;

[0031] For the sensitive data of each sensitive data type, whether each data unit included in the data unit list of the target financial data file output by the corresponding sensitive data identification model is a sensitive data label of the corresponding sensitive data type, the sensitive data label is one of 0 or 1, and the data unit with the sensitive data label of 1 is regarded as the sensitive data of the corresponding sensitive data type, and the data unit with the sensitive data label of 0 is regarded as a unit that does not belong to the corresponding sensitive data type;

[0032] The data units with the sensitive data label of 1 are combined into a sensitive data list corresponding to the sensitive data type;

[0033] Match the data unit list of the target financial data file with the sensitive data list one by one, and establish a mapping relationship between the data units contained in the target financial data file and the corresponding sensitive data types;

[0034] In the target financial data file, data units that need to be marked as corresponding sensitive data types are marked according to the sensitive data types.

[0035] Preferably, the sensitive data statistics are as follows:

[0036] According to the sensitive data type set and the sensitive data tags of the target financial data files, the target financial data files are continuously subjected to n rounds of sensitive data security destruction processing, each round processing one sensitive data type, and n is the total number of sensitive data types of the target financial data files;

[0037] Count all sensitive data types of the target financial data file;

[0038] Obtain a set of sensitive data types defined in advance, match all sensitive data types of the target financial data file with the set of sensitive data types, and obtain weight coefficient values ​​of all sensitive data types in the target financial data file;

[0039] Arrange all sensitive data types in the target financial data file in descending order according to the corresponding weight coefficient values, and number the sensitive data types in the target financial data file in this order, and mark the number of the sensitive data type as i, where i = 1, 2, ..., n.

[0040] Preferably, important data and security data backup are as follows:

[0041] The calculated data importance and data security of the imported financial file data are substituted into the backup coefficient calculation model to calculate the backup coefficient; the calculated backup coefficient is compared with the set backup coefficient threshold. If it is greater than or equal to the threshold, a data backup reminder is issued; if it is less than the threshold, no data backup reminder is issued.

[0042] Preferably, sensitive data destruction is as follows:

[0043] Carry out n rounds of security destruction of sensitive data in sequence according to the numbering order of sensitive data types;

[0044] Obtain the target financial data file after the j-th round of security destruction and mark it as the j-th data file;

[0045] According to the mark of the unit of the i-th sensitive data type in the j-th data file, count the number of units of the i-th sensitive data type and the number of units that do not belong to the i-th sensitive data type;

[0046] If the number of units of the i-th sensitive data type is equal to 0, the j+1-th data file is deleted, the current round of security destruction processing ends, and the next round of security destruction processing begins; otherwise, the destruction processing of the sensitive data of the i-th sensitive data type in the j-th data file is started;

[0047] The j-th data file after this round of destruction is divided into data units using word segmentation technology, all data units are obtained, and input into the sensitive data recognition model of the i-th sensitive data type to obtain the sensitive data label of each data unit, and the number of data units with sensitive data labels of 1 and 0 are counted.

[0048] Preferably, the destruction of sensitive data also includes evaluating the destruction effect, as follows:

[0049] Calculate the destruction evaluation value P: Where β is the scale factor of missed deletions; (1-β) is the scale factor of mistaken deletions; β∈[0,1]; L j The total number of data units contained in the data unit list of the target financial data file; L j 1 is the number of data units with sensitive data label 1; L j 0 is the number of data units with a sensitive data label of 0;

[0050] Compare the destruction evaluation value P with the preset destruction evaluation threshold PS. If P ≥ PS, delete the j-th data file, end the current round of secure destruction, and enter the next round of secure destruction processing; if P < PS, obtain the j-th data file, back it up, and obtain the (j + 1)-th data file; label the data units with sensitive data label 1 in the j-th data file after destruction in the (j + 1)-th data file; compare the data units corresponding to sensitive data label 0 in the j-th data file after destruction with the data units that do not belong to the i-th sensitive data type in the (j + 1)-th data file, obtain the inconsistent data units, label them as not belonging to the i-th sensitive data type, and map them to the (j + 1)-th data file, and restart the destruction processing of the (j + 1)-th data file until the destruction evaluation value is greater than or equal to the preset destruction evaluation threshold.

[0051] In summary, due to the adoption of the above technical solutions, the beneficial effects of the present invention are as follows:

[0052] 1. Through the calculation of the first backup coefficient and the second backup coefficient, the present invention comprehensively considers the types and contents of financial data, comprehensively evaluates the importance of data; uses historical backup data and historical retrieved data, combines with the data imported from financial files, and conducts reasonable security evaluation, improving the accuracy and reliability of the evaluation.

[0053] 2. Based on the sensitive data recognition model of deep learning, combined with natural language processing technology, the present invention conducts fine-grained analysis on financial data and automatically marks sensitive data; through a large amount of data training, the model can efficiently and accurately identify various types of sensitive data, reducing misjudgment and missed judgment.

[0054] 3. By calculating the backup coefficient and comparing it with the preset threshold, the present invention dynamically determines whether to remind the user to back up data, ensuring that important data is protected in a timely manner. The user can adjust the backup threshold according to actual needs, improving the adaptability and flexibility of the system.

[0055] 4. In accordance with the serial number order of sensitive data types, the present invention continuously implements multiple rounds of secure destruction processing to ensure that each type of sensitive data is effectively processed. After each round of destruction, an effect evaluation is carried out to ensure the effectiveness and thoroughness of the destruction process, and to avoid potential security hazards caused by data residue. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0057] Figure 1 The module diagram of the financial data security management system based on artificial intelligence of the present invention is shown;

[0058] Figure 2 A flow chart showing the financial data security management method based on artificial intelligence of the present invention is shown;

[0059] Figure 3 A flow chart of sensitive data marking according to the present invention is shown. DETAILED DESCRIPTION

[0060] 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.

[0061] In addition, the described features, structures or characteristics may be combined in one or more example embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the example embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, steps, etc. may be adopted. In other cases, well-known structures, methods, implementations or operations are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0062] Embodiment 1:

[0063] See also Figure 1 As shown, the artificial intelligence-based financial data security management system of this embodiment includes a data acquisition module, a data preprocessing module, a data analysis module, a sensitive data management module, a data backup reminder module and a sensitive data destruction module.

[0064] Data collection module: responsible for collecting historical financial data and imported financial files.

[0065] Data preprocessing module: used to preprocess the collected data, including cleaning, format conversion, etc., to ensure data quality.

[0066] Data analysis module: uses machine learning algorithms to train models based on historical data to predict the importance and security of imported data and calculate backup coefficients and security scores.

[0067] Sensitive data management module: Based on the sensitive data identification model of deep learning, it combines natural language processing technology to perform fine-grained analysis of data, automatically mark sensitive data, and use database technology to store sensitive data lists and their mapping relationships.

[0068] Data backup reminder module: Set a threshold and determine whether it is necessary to remind the user to back up data based on the backup coefficient.

[0069] Sensitive data destruction module: Destroy sensitive data to ensure that the data is irrecoverable, and ensure the effectiveness of the destruction process through multiple iterative processing and evaluation of the destruction effect.

[0070] The beneficial effects of this embodiment include: through multi-dimensional data collection and processing, intelligent data evaluation and management, and automated sensitive data identification and destruction, it not only improves the security and management efficiency of financial data, but also reduces costs and risks, and enhances data compliance and user experience.

[0071] Embodiment 2:

[0072] See also Figure 2 As shown, the financial data security management method based on artificial intelligence in this embodiment has the following steps:

[0073] Step 1: Collect data.

[0074] S11: Acquire historical financial data and historical retrieved data on the corresponding device, acquire the type and content data of the historical financial data, and acquire the type and content data of the historical retrieved data at the same time. The types of data include documents, pictures, and videos; the content of the data includes the text content of the document, the pixel value content of the picture, and the frame image data content of each frame of the video.

[0075] S12: Obtain the type and content data of the imported financial document.

[0076] Step 2: Calculate data importance.

[0077] S21: Import the acquired category data of historical financial data and the category data of imported financial files into the first backup coefficient calculation formula to calculate the first backup coefficient C 1 Calculation of the first backup coefficient B 1 The calculation formula is as follows:

[0078]

[0079] Where N is the number of historical backup data; C i is the number of types of data in the i-th historical backup data that are the same as the data types in the import file; T i is the number of data types in the i-th historical backup data.

[0080] S22: Calculate a second backup coefficient based on the historical financial data and the content of the imported financial file.

[0081] Import the content data of the historical financial data obtained and the content data of the imported financial file into the second backup coefficient calculation formula to calculate the second backup coefficient B 2 The second backup coefficient B 2 The calculation formula is as follows:

[0082]

[0083] Among them, S () is the number of content data in brackets; K i is the content data in the i-th historical backup data; K is the content data in the imported file; ∪ is the union; ∩ is the intersection.

[0084] S23: Calculate the data importance by combining the first backup coefficient and the second backup coefficient.

[0085] Substitute the calculated first backup coefficient and the second backup coefficient into the data importance calculation formula to calculate the data importance I. The data importance I calculation formula is as follows:

[0086] I=α·B 1 +(1-α)B 2 ;

[0087] Among them, α is the first accounting coefficient; 1-α is the second accounting coefficient; 0≤α≤1.

[0088] Step 3: Calculate data security.

[0089] S31: Acquire the type, content data and retrieval frequency data of historical retrieved data, and simultaneously acquire the type and content data of imported financial document data.

[0090] S32: Import the acquired historical retrieved data types, content data, retrieved times data, and imported financial document data types and content data into a data security calculation formula to calculate the data security S. The data security S calculation formula is:

[0091]

[0092] Among them, f j represents the number of times the i-th historical retrieved data is retrieved; K j represents the content data of the jth historical retrieved data; M represents the number of historical retrieved data.

[0093] Step 4: Mark sensitive data.

[0094] Based on the pre-trained sensitive data identification model, sensitive data of all sensitive data types contained in the target financial data file are marked.

[0095] See also Figure 3 As shown, the sensitive data tags are as follows:

[0096] Data segmentation: Use word segmentation technology to segment the data in the target financial data file into sentences according to punctuation marks.

[0097] Data division: Data division is performed for each sentence, thereby dividing the target financial data file into multiple data units, and obtaining a data unit list of the target financial data file.

[0098] Input sensitive data identification model: input all data units included in the data unit list of the target financial data file into each sensitive data identification model corresponding to the sensitive data type set defined in the early stage.

[0099] Output sensitive data labels: For sensitive data of each sensitive data type, the corresponding sensitive data identification model outputs whether each data unit contained in the data unit list of the target financial data file has a sensitive data label of the corresponding sensitive data type. The sensitive data label is either 0 or 1. The data unit with a sensitive data label of 1 is regarded as sensitive data of the corresponding sensitive data type, and the data unit with a sensitive data label of 0 is regarded as a unit that does not belong to the corresponding sensitive data type.

[0100] Generate a sensitive data list: organize the data units with the sensitive data label of 1 into a sensitive data list corresponding to the sensitive data type.

[0101] Establish a mapping relationship: Match the data unit list of the target financial data file with the sensitive data list one by one, and establish a mapping relationship between the data units contained in the target financial data file and the corresponding sensitive data types.

[0102] Mark sensitive data: In the target financial data file, mark the data units that need to be marked as the corresponding sensitive data types according to the sensitive data types.

[0103] Step 5: Sensitive data statistics.

[0104] According to the sensitive data type set and the sensitive data tag of the target financial data file, the target financial data file is continuously subjected to n rounds of sensitive data security destruction processing, with each round processing one sensitive data type. n is the total number of sensitive data types in the target financial data file.

[0105] Count sensitive data types: Count all sensitive data types in the target financial data file.

[0106] Match sensitive data types: obtain the sensitive data type set defined in advance, match all sensitive data types of the target financial data file with the sensitive data type set, and obtain the weight coefficient values ​​of all sensitive data types in the target financial data file.

[0107] Sorting sensitive data types: Arrange all sensitive data types in the target financial data file in descending order according to the corresponding weight coefficient values, and number the sensitive data types in the target financial data file in this order, marking the number of the sensitive data type as i, where i = 1, 2, ..., n.

[0108] Step 6. Back up important and secure data.

[0109] Calculate the backup coefficient: substitute the calculated data importance and data security of the imported financial file data into the backup coefficient calculation model to calculate the backup coefficient.

[0110] Security reminder: The calculated backup factor is compared with the set backup factor threshold. If it is greater than or equal to the threshold, a data backup reminder is issued; if it is less than the threshold, no data backup reminder is issued.

[0111] Step 7: Destroy sensitive data.

[0112] Continuous destruction processing: Carry out n rounds of secure destruction processing of sensitive data in sequence according to the serial number of sensitive data types.

[0113] Get data file: Get the target financial data file after the j-th round of security destruction and mark it as the j-th data file.

[0114] Count the number of statistical data units: According to the label of the unit of the i-th sensitive data type in the j-th data file, count the number of units of the i-th sensitive data type and the number of units that do not belong to the i-th sensitive data type.

[0115] Determine the destruction conditions: If the number of units of the i-th sensitive data type is equal to 0, delete the j+1-th data file, end this round of security destruction processing, and enter the next round of security destruction processing; otherwise, start the destruction processing of the sensitive data of the i-th sensitive data type in the j-th data file.

[0116] Destruction processing: Use word segmentation technology to divide the j-th data file after this round of destruction into data units, obtain all data units, input them into the sensitive data recognition model of the i-th sensitive data type, and obtain the sensitive data label of each data unit. Count the number of data units with sensitive data labels of 1 and the number of data units with sensitive data labels of 0.

[0117] Calculate the destruction assessment value:

[0118]

[0119] Among them, β is the proportionality factor of the missed deletion value; (1 - β) is the proportionality factor of the mis-deletion value; β ∈ [0, 1]; L j The total number of data units included in the data unit list of the target financial data file; L j 1 is the number of data units with a sensitive data label of 1; L j 0 is the number of data units with a sensitive data label of 0.

[0120] Evaluate the destruction effect: Compare the destruction evaluation value P with the preset destruction evaluation threshold PS.

[0121] If P ≥ PS, delete the j-th data file, end this round of secure deletion, and enter the next round of secure deletion processing;

[0122] If P < PS, obtain the j-th data file, back it up, and obtain the (j + 1)-th data file. Mark the data units with a sensitive data label of 1 in the j-th data file after destruction in the (j + 1)-th data file; Compare the data units corresponding to the sensitive data label of 0 in the j-th data file after destruction with the data units that do not belong to the i-th sensitive data type in the (j + 1)-th data file, obtain the inconsistent data units, mark them as not belonging to the i-th sensitive data type, and map them to the (j + 1)-th data file. Restart the destruction processing of the (j + 1)-th data file until the destruction evaluation value is greater than or equal to the preset destruction evaluation threshold.

[0123] The beneficial effects of this embodiment are as follows: Through data collection and preprocessing, the system can efficiently collect and organize financial data, providing a reliable basis for subsequent analysis; Secondly, by using machine learning algorithms, it accurately predicts the importance and security of data, providing a scientific basis for data backup and destruction, and effectively improving the intelligent level of data security management; Using deep learning technology, it automatically marks and manages sensitive data, reducing the risk of human intervention and enhancing the accuracy of data protection. In addition, it intelligently reminds users to back up according to the backup coefficient, avoiding the risk of data loss. Finally, through multiple iterative processes and effect evaluations, it ensures that sensitive data is completely destroyed and cannot be recovered, guaranteeing information security.

[0124] The proportion coefficients and proportional factors of the present invention are used to measure the degree of influence of different factors or variables on a certain result or decision. The definition of proportion coefficients and proportional factors refers to the numerical value assigned to each factor when comparing and evaluating multiple factors to reflect its importance or priority. These proportion coefficients and proportional factors can be determined according to specific circumstances and needs, and are usually jointly formulated and confirmed by professionals or relevant stakeholders. By reasonably setting the proportion coefficients and proportional factors, it can help the program or system make decisions or predictions more accurately.

[0125] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

[0126] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. Financial data security management system based on artificial intelligence, characterized by: The system includes a data acquisition module, a data preprocessing module, a data analysis module, a sensitive data management module, a data backup reminder module and a sensitive data destruction module; Data collection module: responsible for collecting historical financial data and imported financial files; Data preprocessing module: used to preprocess the collected data, including cleaning, format conversion, etc., to ensure data quality; Data analysis module: uses machine learning algorithms to train models based on historical data to predict the importance and security of imported data and calculate backup coefficients and security scores; Sensitive data management module: Based on the sensitive data identification model of deep learning, it combines natural language processing technology to perform fine-grained analysis of data, automatically mark sensitive data, and use database technology to store sensitive data lists and their mapping relationships; Data backup reminder module: Set the threshold and determine whether to remind the user to back up data based on the backup coefficient; Sensitive data destruction module: Destroy sensitive data to ensure that the data is irrecoverable, and ensure the effectiveness of the destruction process through multiple iterative processing and evaluation of the destruction effect.

2. A financial data security management method based on artificial intelligence, used to implement the financial data security management system based on artificial intelligence as claimed in claim 1, characterized in that: The method flow is as follows: Collect data, obtain historical financial data and historical retrieved data on the corresponding device, obtain the type and content data of historical financial data, and at the same time obtain the type and content data of historical retrieved data. The types of data include documents, pictures and videos; the content of data includes the text content of documents, the pixel value content of pictures and the frame image data content of each frame of videos; obtain the type and content data of imported financial files; calculate data importance; calculate data security; mark sensitive data; Sensitive data statistics; Backup of important data and security data; destruction of sensitive data.

3. The financial data security management method based on artificial intelligence according to claim 2 is characterized in that: The data importance calculation is specifically as follows: S21: Import the acquired type data of historical financial data and the type data of imported financial files into the first backup coefficient calculation formula to calculate the first backup coefficient C1. The calculation formula of the first backup coefficient B1 is as follows: Where N is the number of historical backup data; C i is the number of types of data in the i-th historical backup data that are the same as the data types in the import file; T i is the number of data types in the i-th historical backup data; S22: Calculate the second backup coefficient based on the historical financial data and the content of the imported financial file; import the acquired content data of the historical financial data and the content data of the imported financial file into the second backup coefficient calculation formula to calculate the second backup coefficient B2. The calculation formula of the second backup coefficient B2 is as follows: Among them, K i is the content data in the i-th historical backup data; K is the content data in the imported file; ∪ is the union; ∩ is the intersection; S(K i ∩K) is K i The number of intersections with K; S(K i ∪K) is K i The number of sets that are combined with K; S23: Calculate the data importance by combining the first backup coefficient and the second backup coefficient; substitute the calculated first backup coefficient and the second backup coefficient into the data importance calculation formula to calculate the data importance I. The data importance I calculation formula is as follows: I=α·B1+(1-α)B2; wherein α is the first proportion coefficient; 1-α is the second proportion coefficient; 0≤α≤1.

4. The financial data security management method based on artificial intelligence according to claim 2 is characterized in that: The data security calculation is specifically as follows: The types, content data, retrieval times data of the historical retrieved data, and the types and content data of the imported financial file data are imported into the data security calculation formula to calculate the data security S. The data security S calculation formula is: Among them, f j represents the number of times the i-th historical retrieved data is retrieved; K j represents the content data of the jth historical retrieved data; M represents the number of historical retrieved data; K represents the content data in the imported file; S(K j ∩K) is K j The number of intersections with K; S(K j ∪K) is K j The number of sets that are union with K.

5. The financial data security management method based on artificial intelligence according to claim 2 is characterized in that: The sensitive data markings are as follows: Using word segmentation technology, the data in the target financial data file is segmented into sentences according to punctuation marks; Perform data division for each sentence, divide the target financial data file into multiple data units, and obtain a data unit list of the target financial data file; Input all data units included in the data unit list of the target financial data file into each sensitive data recognition model corresponding to the pre-trained sensitive data type set; For the sensitive data of each sensitive data type, whether each data unit included in the data unit list of the target financial data file output by the corresponding sensitive data identification model is a sensitive data label of the corresponding sensitive data type, the sensitive data label is one of 0 or 1, and the data unit with the sensitive data label of 1 is regarded as the sensitive data of the corresponding sensitive data type, and the data unit with the sensitive data label of 0 is regarded as a unit that does not belong to the corresponding sensitive data type; The data units with the sensitive data label of 1 are combined into a sensitive data list corresponding to the sensitive data type; Match the data unit list of the target financial data file with the sensitive data list one by one, and establish a mapping relationship between the data units contained in the target financial data file and the corresponding sensitive data types; In the target financial data file, data units that need to be marked as corresponding sensitive data types are marked according to the sensitive data types.

6. The financial data security management method based on artificial intelligence according to claim 2 is characterized in that: The sensitive data statistics are as follows: According to the sensitive data type set and the sensitive data tags of the target financial data files, the target financial data files are continuously subjected to n rounds of sensitive data security destruction processing, each round processing one sensitive data type, and n is the total number of sensitive data types of the target financial data files; Count all sensitive data types of the target financial data file; Obtain a set of sensitive data types defined in advance, match all sensitive data types of the target financial data file with the set of sensitive data types, and obtain weight coefficient values ​​of all sensitive data types in the target financial data file; Arrange all sensitive data types in the target financial data file in descending order according to the corresponding weight coefficient values, and number the sensitive data types in the target financial data file in this order, and mark the number of the sensitive data type as i, where i = 1, 2, ..., n.

7. The financial data security management method based on artificial intelligence according to claim 2 is characterized in that: The important data and security data backup are as follows: Substituting the calculated data importance and data security of the imported financial file data into the backup coefficient calculation model to calculate the backup coefficient; comparing the calculated backup coefficient with the set backup coefficient threshold, and if it is greater than or equal to the threshold, a data backup reminder is issued; If it is less than the threshold, no data backup reminder will be issued.

8. The financial data security management method based on artificial intelligence according to claim 2 is characterized in that: The sensitive data destruction is as follows: Carry out n rounds of security destruction of sensitive data in sequence according to the numbering order of sensitive data types; Obtain the target financial data file after the j-th round of security destruction and mark it as the j-th data file; According to the mark of the unit of the i-th sensitive data type in the j-th data file, count the number of units of the i-th sensitive data type and the number of units that do not belong to the i-th sensitive data type; If the number of units of the i-th sensitive data type is equal to 0, the j+1-th data file is deleted, the current round of security destruction processing ends, and the next round of security destruction processing begins; Otherwise, start destroying the sensitive data of the i-th sensitive data type in the j-th data file; The j-th data file after this round of destruction is divided into data units using word segmentation technology, all data units are obtained, and input into the sensitive data recognition model of the i-th sensitive data type to obtain the sensitive data label of each data unit, and the number of data units with sensitive data labels of 1 and 0 are counted.

9. The financial data security management method based on artificial intelligence according to claim 2 is characterized in that: The sensitive data destruction also includes evaluating the destruction effect, as follows: Calculate the destruction evaluation value P: Where β is the scale factor of missed deletions; (1-β) is the scale factor of mistaken deletions; β∈[0,1]; L j The total number of data units contained in the data unit list of the target financial data file; L j 1 is the number of data units with sensitive data label 1; L j 0 is the number of data units with a sensitive data label of 0; Compare the destruction evaluation value P with the preset destruction evaluation threshold PS. If P≥PS, delete the j-th data file, end the current round of secure destruction, and enter the next round of secure destruction processing; if P<PS, obtain the j-th data file, back it up, and obtain the (j + 1)-th data file; label the data units with sensitive data label 1 in the j-th data file after destruction in the (j + 1)-th data file; compare the data units corresponding to sensitive data label 0 in the j-th data file after destruction with the data units that do not belong to the i-th sensitive data type in the (j + 1)-th data file, obtain the inconsistent data units, label them as not belonging to the i-th sensitive data type, and map them to the (j + 1)-th data file, and restart the destruction processing of the (j + 1)-th data file until the destruction evaluation value is greater than or equal to the preset destruction evaluation threshold.

Images