Data asset protection method suitable for financial information platform

Abstract

The invention discloses a data asset protection method suitable for a financial information platform, and the method comprises the steps: capturing an operation track, carrying out the dynamic control of frequency, carrying out the dynamic protection linkage, carrying out the hierarchical control of character reading, carrying out the data display, and carrying out the afterward tracing, and enabling the overall data safety and traceability of the financial data platform to be improved. A service related party is informed through a notification or mail method, response measures are prepared at the first time, and an access log is associated with a user identity and an encrypted fingerprint to generate an evidence obtaining report; the cross-modal cooperation of dynamic confusion and voiceprint encryption is combined with front-end page dynamic and biological feature encryption, the data acquisition difficulty is doubly improved, and the content and the identity of a leaked person can be synchronously traced when data is leaked by distinguishing manual and machine operations, dynamically adjusting a protection strategy and utilizing virtual identifier-voiceprint dual binding, so that the real-time security of the data is ensured. A dynamic anti-crawler strategy and a multi-mode content encryption technology are adopted, and active protection and afterward traceability tracking of financial text data and audio data are achieved.

Description

Technical Field

[0001] The present invention relates to a data asset protection method applicable to a financial information platform, and in particular to the field of software information security technology. Background Art

[0002] Traditional anti-crawler technologies rely on passive defenses such as IP rate limiting, User-Agent detection, and verification code verification. Their static page structures or API interface designs are susceptible to reverse engineering. Common financial information platforms use SSL / TLS data encryption for text content, but plaintext data can still be parsed by crawlers. Audio content uses general-purpose AES encryption, but decrypted content can be re-distributed.

[0003] Static defense is easy to crack: Traditional methods such as IP frequency limiting and verification codes rely on fixed rules and are easily bypassed by proxy pools and AI-based code-breaking platforms. For example, a platform uses verification codes to block crawlers, but the cracking rate of machine learning models is still over 50%.

[0004] Insufficient dynamic encryption: Some platforms use SSL / TLS transmission encryption or AES content encryption, but data can still be disseminated after decryption, lacking a dynamic encryption mechanism bound to user identity.

[0005] Low accuracy in human-machine identification: Protection strategies based on User-Agent or simple behavioral analysis have difficulty distinguishing between advanced anthropomorphic crawlers (such as simulated mouse tracks and page dwell time) and real users, with a misjudgment rate of up to 30%. Summary of the Invention

[0006] The present invention provides a data asset protection method applicable to a financial information platform to overcome the defects of the data encryption technology of a common financial information platform in the prior art, such as the easy cracking of static defense and the insufficient dynamic encryption.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] The present invention discloses a data asset protection method applicable to a financial information platform, comprising the following steps:

[0009] 1) Request initiated

[0010] Users access financial data pages / APIs through the client.

[0011] The behavioral feature analysis module captures operation trajectories in real time, including mouse movement speed and click frequency;

[0012] 2) Frequency dynamic control

[0013] The dynamic threshold currently available to users is divided into two parts. First, a personalized frequency baseline value A is generated based on the user's historical behavior data. Then, the dynamic threshold is calculated in real time based on the current system load. The value is calculated as: baseline A × (1 + current system load factor).

[0014] A gradient response mechanism is formed based on dynamic thresholds, and operational processes or restrictions are implemented according to the scale of exceeding the limit:

[0015] A threshold of 120% is a level 1 overlimit, triggering lightweight verification;

[0016] A threshold of 200% is a Level 2 over-limit: voiceprint verification is enabled and the data return granularity is reduced;

[0017] Continuous over-limit is classified as Level 3: the account is temporarily locked and manual review is initiated.

[0018] 3) Dynamic protection linkage

[0019] Normal user path:

[0020] a. Behavioral characteristics meet the baseline → Dynamic frequency control module releases bandwidth resources

[0021] b. The page obfuscation module generates low-noise virtual identifiers

[0022] c. The voiceprint encryption module enables standard encryption according to user level, where identity recognition is only embedded in the basic ID

[0023] Abnormal access path:

[0024] a. Behavioral deviation exceeds the threshold → triggering the gradient response mechanism

[0025] b. Improved page obfuscation strength

[0026] c. Forced voiceprint deep encryption

[0027] 4) Text reading level control

[0028] VIP / normal user levels and public / internal content sensitivity labels jointly determine reading permissions;

[0029] Unpublished financial reports or meeting minutes are highly sensitive text and must be in self-destructing mode.

[0030] If the page remains on for more than 30 seconds, it will be automatically blurred and require voiceprint verification to unlock; screenshots are prohibited; operations trigger the implantation of audio watermarks.

[0031] OCR interference layer: superimposes an invisible noise matrix on the bottom layer of the text to interfere with the crawler's OCR recognition accuracy; key numbers are dynamically replaced with image rendering;

[0032] 5) Data presentation

[0033] Display the content on the interface according to the control rules in 4);

[0034] 6) Retrospective

[0035] Voiceprint features in leaked data automatically trigger a traceability alert, notifying relevant parties via notification or email, allowing them to prepare countermeasures immediately.

[0036] Access logs are used to associate user identities with encrypted fingerprints to generate forensic reports.

[0037] Furthermore, the historical behavior data includes average daily visits and time period distribution.

[0038] Furthermore, the low-interference virtual identifier is a 30% page replacement rate.

[0039] Furthermore, the page obfuscation strength improvement includes a 70% page replacement rate and minute-level refresh of the DOM structure.

[0040] Furthermore, the mandatory voiceprint deep encryption includes embedding geographic location and device fingerprint composite features.

[0041] Furthermore, the prohibition of screenshots includes Canvas rendering and DRM encryption technology.

[0042] Furthermore, the operation triggers the audio watermark implantation to automatically insert the user ID audio fingerprint when copying and pasting the content.

[0043] The beneficial effects achieved by the present invention are:

[0044] 1. The overall data security and traceability of the financial data platform are enhanced through the combined effect of multiple technical means. The cross-modal collaboration of dynamic obfuscation and voiceprint encryption, combined with the dynamic front-end page and biometric encryption, doubles the difficulty of data acquisition. By distinguishing between manual and machine operations, protection strategies can be dynamically adjusted. By utilizing the dual binding of virtual identifiers and voiceprints, the content and identity of the leaker can be simultaneously traced in the event of a data leak.

[0045] 2. Systematically associate user and data asset portraits, dynamically associate data levels, visitor operations, and data return precision, and achieve a balance between resource consumption and security. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] 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:

[0047] Figure 1 It is a system architecture diagram of the present invention;

[0048] Figure 2 It is the core interaction flow chart of the present invention. DETAILED DESCRIPTION

[0049] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0050] Example 1

[0051] like Figures 1-2 As shown, a data asset protection method applicable to a financial information platform includes the following steps:

[0052] Step 1 Request initiation

[0053] Users access financial data pages / APIs through the client.

[0054] The behavioral feature analysis module captures operation trajectories in real time, including mouse movement speed and click frequency;

[0055] Step 2 Frequency Dynamic Control

[0056] The dynamic threshold currently available to users is divided into two parts. First, a personalized frequency baseline value A is generated based on historical user behavior data (such as average daily visits and time period distribution). Then, the dynamic threshold is calculated in real time based on the current system load. This value is calculated as: baseline A × (1 + current system load factor).

[0057] A gradient response mechanism is formed based on dynamic thresholds, and operational processes or restrictions are implemented according to the scale of exceeding the limit:

[0058] Level 1 overlimit (threshold 120%): triggers lightweight verification (such as slider verification);

[0059] Level 2 over-limit (threshold 200%): Enable voiceprint verification and reduce the granularity of data returned (such as only showing a summary);

[0060] Level 3 Overlimit (Continuous Overlimit): The account is temporarily locked and manual review is initiated.

[0061] Step 3 Dynamic protection linkage

[0062] Normal user path:

[0063] a. Behavioral characteristics meet the baseline → Dynamic frequency control module releases bandwidth resources

[0064] b. Page obfuscation module generates low-noise virtual identifiers (30% replacement rate)

[0065] c. The voiceprint encryption module enables standard encryption according to user level, where identity recognition is only embedded in the basic ID

[0066] Abnormal access path:

[0067] a. Behavioral deviation exceeds the threshold → triggering the gradient response mechanism

[0068] b. Improved page obfuscation strength (70% replacement rate + DOM structure refresh every minute)

[0069] c. Forced voiceprint deep encryption (embedded with geolocation + device fingerprint composite features)

[0070] Step 4 Text reading level control

[0071] User level (VIP / normal) and content sensitivity label (public / internal) jointly determine reading permissions;

[0072] Highly sensitive texts (such as unpublished financial reports and meeting minutes) are forced to use the self-destructing mode:

[0073] If the page stays on for more than 30 seconds, it will be automatically blurred and require re-voiceprint verification to unlock; screenshots are prohibited (based on Canvas rendering and DRM encryption technology); operations trigger audio watermark implantation (user ID audio fingerprint is automatically inserted when copying and pasting content).

[0074] OCR interference layer: An invisible noise matrix is ​​superimposed on the bottom layer of the text to interfere with the crawler's OCR recognition accuracy; key figures are dynamically replaced with image rendering (for example, "P / E ratio 25.3" is converted to SVG vector graphics);

[0075] Step 5 Data display

[0076] According to the control rules in step 4, display the content on the interface;

[0077] Step 6: Retrospective

[0078] Voiceprint features in leaked data automatically trigger a traceability alert, notifying relevant parties via notification or email, allowing them to prepare countermeasures immediately.

[0079] Access logs are used to associate user identities with encrypted fingerprints to generate forensic reports.

[0080] Specific steps: Taking the dynamic frequency control in step 2 as an example, the financial information platform has the characteristics of high data asset value and a large proportion of page reading operations. In this way, dynamically controlling the access frequency based on user behavior is an important part of protecting financial data assets. The first step is to collect user behavior data based on historical user behavior, including the average daily visits (the statistical period is 30 days, excluding data on abnormal days), the access period distribution model (divided into 24-hour segments, calculating the mean and standard deviation of the request volume in each period), and the operation type weight (such as query, repeated query, and download are assigned coefficients of 0.5, 1.2, and 1.5 respectively). Through this series, a personalized frequency baseline value A is generated.

[0081] The second step is to generate the load factor B of the system at that time through monitoring (sampling frequency 500ms):

[0082] index Weight Normalization algorithm CPU utilization (cpu) 40％ (Current value - Baseline) / Maximum load value Memory usage (mem) 30％ Current occupied / total memory capacity Request queue depth (queue) 30％ Waiting queue length / maximum queue threshold

[0083] The calculation formula is: B = 1 + (0.4*{cpu}+0.3*{mem}+0.3*{queue});

[0084] The third step is to calculate the current method coefficient C based on AB and compare it with the benchmarks D1, D2, and D3. If it exceeds D1, user access is restricted as follows: a dynamic slider component is injected into the front end, and a 9x9 pixel matrix trajectory verification code is generated in the back end; a time-limited token (valid for 2 minutes) is issued after successful verification, and the threshold for this period is temporarily relaxed to 110%; behavioral fingerprints (IP + device fingerprint + operation sequence) are asynchronously recorded to the risk control database. If it exceeds the benchmark D2, a second-level alarm is triggered, and user access is restricted as follows: the voice collection component is triggered, requiring the user to read a random 6-digit code, and the user's voiceprint is compared with the voiceprints of existing users. If the difference is too large, the user is blocked from access; a data degradation strategy is implemented, and text content is displayed in summary to prevent full text from being read. At the same time, image content is reduced in clarity and batch watermarking is used to reduce the risk of data loss. If it continues to exceed the benchmark D3, this method will reduce the risk of data loss through forced operations, as follows:

[0085] 30 minutes Account lock + SMS OTP verification unlock 2 hours Forced offline and frozen for 24 hours 24 hours Activate the manual review process (submit identity proof + usage scenario description)

[0086] The text reading classification control in step 4 realizes the classification of financial information according to the number / frequency of readings, and provides differentiated security protection strategies based on documents of different confidentiality levels to ensure the controllability of sensitive information in the entire link of access, reading, and dissemination. A mapping matrix of four levels of user levels (ordinary, VIP, audit, and executive) and four levels of content sensitivity labels (public, internal, confidential, and top secret) is established, and a semantic analysis engine is used to annotate content risks in real time. By automatically identifying high-risk entities such as financial data and technical parameters in the text, and combining contextual semantics to evaluate comprehensive sensitivity, content labels are dynamically generated. When determining permissions, the system implements dynamic interception based on the strength of user identity authentication (such as mobile phone number + real-name authentication for ordinary users, and mandatory biometric binding for executive accounts) and document sensitivity. Access is authorized when the user level ≥ the content label, and an emergency override switch is reserved for use in special scenarios. It achieves a balance between security protection and business efficiency, and is suitable for digital document management scenarios in highly sensitive industries.

[0087] Taking roadshow minutes as an example, this content is a core financial data asset and needs to be protected in both text and audio recording.

[0088] Anti-crawler protection performance is significantly improved

[0089] Improved crawler parsing failure rate: For traditional static pages, the crawler's success rate in locating data using XPath / CSS selectors can reach over 95%. By changing the hierarchical relationship of page elements every 5 minutes, the present invention reduces the crawler's locating failure rate to 98%. 70% of key data fields (such as stock prices and trading volumes) are replaced with meaningless hash values, reducing the completeness of valid data extraction to less than 5%. The experimental data is as follows:

[0090]

[0091] Data encryption and traceability upgrades

[0092] Audio content security is enhanced. Voiceprint feature embedding uses a frequency domain modulation algorithm with a 256-bit key. Cracking requires more than 10^15 calculations (referring to NIST standards). Leakage scenario testing:

[0093] Attack methods Traditional AES encrypted content availability Availability of the voiceprint encrypted content of the present invention Direct playback from illegal devices 100％ 0% (noise output only) Professional software to extract voiceprints not applicable Need to match the target user's real-time voiceprint Replay attack (recording playback) efficient Dynamic changes in voiceprints block replay

[0094] Improved leak tracing efficiency

[0095] Tracking mechanism upgrade: Any leaked data will contain dual tags of voiceprint fingerprint and virtual identifier; through reverse query of the voiceprint database, the identity of the leaker can be located within 5 minutes;

[0096] Actual case: During a trial period at a cooperative mutual fund company, two data leakage incidents were discovered: Case 1: An employee illegally took screenshots to disseminate an internal research report, which was then traced to a specific account by rendering a watermark; Case 2: A crawler using the company's account broke through the front-end obfuscation to obtain some data → the voiceprint features were missing, making the data unusable for commercial use.

[0097] Balance between system performance and user experience

[0098] Resource consumption compared to traditional solutions: Page loading delay: an average increase of 12ms (traditional anti-crawl solutions increase by 80ms); Voiceprint verification time: completed synchronously with audio playback, without additional user operation;

[0099] Module CPU usage (traditional solution) CPU occupancy rate (present invention) Anti-crawler engine 35％ 28％ Encryption 22％ 18％

[0100] It should be noted that the above descriptions are merely preferred embodiments of the present invention and are 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 will appreciate that they may modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, improvements, and the like that fall within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention. The terms used in the description of this application are intended only to describe specific embodiments and are not intended to limit the exemplary embodiments of the present invention. For ease of description, the dimensions of the various parts shown in the drawings are not drawn to scale. Technologies, methods, and devices known to persons skilled in the relevant art may not be discussed in detail, but where appropriate, such technologies, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely illustrative and not limiting. Therefore, other examples of the exemplary embodiments may have different values. It should be noted that similar reference numerals and letters denote similar items in the following figures. Therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0101] It should be noted that the terms "first," "second," etc. in the specification and claims of this application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way are interchangeable where appropriate, so that the embodiments of this application can be implemented in an order other than those illustrated or described herein, and that the objects distinguished by "first," "second," etc. are generally of the same type, and do not limit the number of objects. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0102] It should be noted that, in the description of this application, the directions or positional relationships indicated by directional terms such as "front, back, up, down, left, right", "horizontal, vertical, vertical, horizontal" and "top, bottom" are usually based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description. Unless otherwise specified, these directional terms do not indicate or imply that the device or element referred to must have a specific direction or be constructed and operated in a specific direction. Therefore, they cannot be understood as limiting the scope of protection of this application; the directional terms "inside and outside" refer to the inside and outside relative to the outline of each component itself.

Claims

1. A data asset protection method applicable to a financial information platform, characterized in that: The steps include: 1) Request initiated Users access financial data pages / APIs through the client. The behavioral feature analysis module captures operation trajectories in real time, including mouse movement speed and click frequency; 2) Frequency dynamic control The dynamic threshold currently available to users is divided into two parts. First, a personalized frequency baseline value A is generated based on the user's historical behavior data. Then, the dynamic threshold is calculated in real time based on the current system load. The value is calculated as: baseline A × (1 + current system load factor). A gradient response mechanism is formed based on dynamic thresholds, and operational processes or restrictions are implemented according to the scale of exceeding the limit: A threshold of 120% is a level 1 overlimit, triggering lightweight verification; A threshold of 200% is a Level 2 over-limit: voiceprint verification is enabled and the data return granularity is reduced; Continuous over-limit is classified as Level 3: the account is temporarily locked and manual review is initiated. 3) Dynamic protection linkage Normal user path: a. Behavioral characteristics meet the baseline → Dynamic frequency control module releases bandwidth resources b. The page obfuscation module generates low-noise virtual identifiers c. The voiceprint encryption module enables standard encryption according to user level, where identity recognition is only embedded in the basic ID Abnormal access path: a. Behavioral deviation exceeds the threshold → triggering the gradient response mechanism b. Improved page obfuscation strength c. Forced voiceprint deep encryption 4) Text reading level control VIP / normal user levels and public / internal content sensitivity labels jointly determine reading permissions; Unpublished financial reports or meeting minutes are highly sensitive text and must be in self-destructing mode. If the page remains on for more than 30 seconds, it will be automatically blurred and require voiceprint verification to unlock; screenshots are prohibited; operations trigger the implantation of audio watermarks. OCR interference layer: superimposes an invisible noise matrix on the bottom layer of the text to interfere with the crawler's OCR recognition accuracy; Key figures are dynamically replaced with image renderings; 5) Data presentation Display the content on the interface according to the control rules in 4); 6) Retrospective Voiceprint features in leaked data automatically trigger a traceability alert, notifying relevant parties via notification or email, allowing them to prepare countermeasures immediately. Access logs are used to associate user identities with encrypted fingerprints to generate forensic reports.

2. The data asset protection method applicable to a financial information platform according to claim 1, characterized in that: The historical behavior data includes average daily visits and time period distribution.

3. The data asset protection method applicable to a financial information platform according to claim 1, characterized in that: The low-interference virtual identifier is a 30% page replacement rate.

4. The data asset protection method applicable to a financial information platform according to claim 1, characterized in that: The improvement in page obfuscation strength includes a 70% page replacement rate and minute-level refresh of the DOM structure.

5. The data asset protection method applicable to a financial information platform according to claim 1, characterized in that: The mandatory voiceprint deep encryption includes embedding geographic location and device fingerprint composite features.

6. The data asset protection method applicable to a financial information platform according to claim 1, characterized in that: The prohibition on screenshots includes Canvas rendering and DRM encryption technology.

7. The data asset protection method applicable to a financial information platform according to claim 1, characterized in that: The operation triggers the audio watermark implantation to automatically insert the user ID audio fingerprint when copying and pasting the content.

Images