SMS signature intelligent filing verification system

Abstract

This invention discloses an intelligent SMS signature registration and verification system, belonging to the field of internet security service technology. This invention addresses the security vulnerabilities of existing technologies that rely on static keys and fixed policies, lacking behavioral analysis and security mechanisms, leading to key leaks, permission abuse, and content fraud. By introducing multi-factor authentication combined with behavioral profiling technology, the authenticity and legitimacy of the registration entity are ensured. Based on dynamic trust assessment and permission policy adjustment, intelligent access control is achieved, adapting not only to normal business needs but also responding promptly to abnormal behavior to ensure system security. Furthermore, hash operations and dynamic instruction binding ensure the uniqueness and immutability of each SMS request. A response mechanism driven by a multi-layered verification and decision tree model further guarantees the efficient, secure, and reliable operation of SMS communication.

Description

Technical Field

[0001] This invention relates to the field of internet security service technology, specifically to an intelligent SMS signature registration and verification system. Background Technology

[0002] SMS service is an indispensable basic communication tool in areas such as enterprise notifications, identity verification, and marketing promotion. SMS signatures are key information that identifies the sender.

[0003] Traditional filing systems primarily rely on manual submission and qualification review. Once approved, a fixed signature and static API key are assigned. However, this model has the following inherent drawbacks:

[0004] 1. Once a static API key is leaked, it can be misused indefinitely, causing a serious information protection vulnerability.

[0005] 2. Access control policies are usually fixed and cannot be dynamically adjusted based on historical patterns of sending behavior, content risks, recipient groups, or other dimensions.

[0006] 3. It cannot distinguish between normal business transmissions and malicious fraudulent transmissions, and lacks the ability to identify and intervene in illegal content or illegal activities under legitimate signatures.

[0007] 4. Traditional SMS protocols do not have a strong security mechanism that binds them to registration information, and therefore cannot provide end-to-end guarantees for data protection.

[0008] Therefore, in order to meet the existing needs, a smart SMS signature filing and verification system is proposed. Summary of the Invention

[0009] The purpose of this invention is to provide an intelligent SMS signature registration and verification system. By introducing multi-factor authentication and combining it with behavioral profiling technology, the system ensures the authenticity and legality of the registration entity. Based on dynamic trust assessment and permission policy adjustment, it achieves intelligent access control, which can not only adapt to normal business needs but also respond to abnormal behavior in a timely manner, ensuring system security. Through hash operations and dynamic instruction binding, it ensures the uniqueness and immutability of each SMS request. Based on a multi-layer verification and decision tree model-driven response mechanism, it further ensures the efficient, secure, and reliable operation of SMS communication, thus solving the problems mentioned in the background technology.

[0010] To achieve the above objectives, the present invention provides the following technical solution:

[0011] The SMS signature intelligent filing and verification system includes:

[0012] The intelligent filing unit is configured to register user accounts and file entity records, introduce multi-factor authentication, and build behavioral profiles;

[0013] The access control unit is configured to perform dynamic access control based on behavioral profiles and account registration duration.

[0014] The dynamic security unit is configured to initialize a seed for each registered entity. Before each request is sent, the client uses IoT hash operation to iteratively calculate the initialization seed and generate a one-time dynamic instruction.

[0015] The verification and execution unit is configured to intercept and verify all outbound SMS requests at the millisecond level. The verification content includes: signature and token verification, policy compliance verification, and real-time risk scoring.

[0016] The secure communication protocol unit is configured to embed dynamic instructions, signature IDs, and content hash values ​​as standard fields into the communication protocol request header, enabling the receiving end to verify the integrity and authenticity of the request by querying the public key or shared seed of the filing system.

[0017] Furthermore, the intelligent registration unit includes:

[0018] The multi-factor authentication module is configured to verify the legal identity credentials and official communication channels of the filing initiator, and to verify the biometrics of the authorized person responsible for the filing initiator.

[0019] Once all three verifications are successful, the entity registration is confirmed as successful, and a unique registration root identifier is generated.

[0020] Build behavioral profiles and assign them compliant API communication key pairs;

[0021] The audit log module is configured to record all operations, verification attempts, timestamps, IP addresses, and device fingerprint information throughout the multi-factor authentication process, and is set to read-only mode.

[0022] Furthermore, the multi-factor authentication module includes:

[0023] The first authentication module is configured to receive the legal identity core credential document submitted by the applicant for filing, which corresponds to the type of the sending entity. It extracts the key fields in the document through image recognition technology and performs real-time interface verification with the database to confirm the authenticity and validity of the credential.

[0024] The second authentication module is configured to receive one or more official communication channels associated with the applicant's identity and send randomly generated dynamic instructions to the official communication channels to verify whether the content filled in by the applicant is correct.

[0025] The third authentication module is configured to guide the authorized person to complete a liveness detection sequence by calling the face recognition service, and compare the captured biometric information with the benchmark information in the identity information database to verify whether the current applicant is consistent with the identity of the authorized person.

[0026] Further, the access control unit includes:

[0027] The dynamic trust assessment module is configured to comprehensively and quantitatively assess the credibility of the registered entity. Its assessment dimensions include: assessing the static attributes of the registered entity, assessing the strength of the identity authentication passed by the registered entity in the initial registration stage, assessing the historical and real-time behavioral data of the registered entity, assessing the normal sending success rate of the registered entity within a specified period, assessing the degree of deviation of the registered entity's sending behavior in time distribution, frequency fluctuation and content template from the historical record, and receiving real-time risk signals.

[0028] The dynamic permission policy module is configured with multiple permission policy levels that are pre-defined and associated with the trust assessment result range, forming a policy library.

[0029] Furthermore, the access control unit also includes:

[0030] The permission dynamic execution module is configured to automatically adjust the corresponding permission policy based on real-time evaluation results;

[0031] The closed-loop feedback module is configured to provide real-time feedback on execution results and new entity behavior data, dynamically update behavior profiles, and optimize the assessment accuracy and predictive capabilities of the dynamic trust assessment module.

[0032] Furthermore, the dynamic security unit includes:

[0033] The instruction binding module is configured to use a hash function to perform a hash operation on the SMS content, and bind the hash value with a dynamic instruction to ensure that the request is non-replayable and tamper-proof.

[0034] The binding verification module is configured to verify the validity of the binding relationship on the receiving end, ensuring the consistency of dynamic commands, SMS content, and timestamps; if the binding relationship is tampered with or the timestamp is invalid, the receiving end will reject the request.

[0035] Furthermore, the dynamic security unit also includes:

[0036] The instruction verification module is configured to parse the received dynamic instructions, extract their hash values ​​and iteration counts, and verify whether the format and content of the dynamic instructions meet expectations.

[0037] The key update module sets a lifecycle for each seed. When the lifecycle ends, it automatically triggers the key update process to generate a new seed and distribute it securely.

[0038] Furthermore, the verification and execution unit includes:

[0039] The response action module is configured to pre-set corresponding actions based on historical verification results. A decision tree model is used to cluster and associate the actions with the verification results to form an action library.

[0040] The verification response module is configured to obtain multi-level verification results, associate them with preset handling actions in the action library, and execute the corresponding handling strategies.

[0041] Furthermore, the method for constructing behavioral profiles includes:

[0042] Obtain multi-dimensional associated data of the SMS signature filing entity. The multi-dimensional associated data includes core verification data, operational behavior data, and environmental associated data. The core verification data includes legal identity certificate data, official communication channel data, and biometric data. The operational behavior data includes filing operation logs and interaction process data. The environmental associated data includes device feature data, network feature data, and scene feature data.

[0043] The collected multi-dimensional associated data is subjected to dual integrity verification, which includes original data hash verification and feature consistency verification. Valid associated data is obtained after the verification is passed.

[0044] The effective associated data is subjected to hierarchical desensitization and security hardening processing. The effective associated data is divided into first-level sensitive data, second-level sensitive data and third-level sensitive data. According to the preset hierarchical desensitization rules, the sensitive data of different levels is desensitized in a non-deletion manner. Then, the desensitized data is stored by double hashing, local desensitization processing and AES-256 algorithm encryption to obtain secure data.

[0045] The security processing data is split into fine-grained behavioral dimensions, divided into single operation behaviors, associated verification behaviors, and environment-related behaviors. A three-dimensional behavior matrix is ​​constructed based on the filing entity ID, behavior dimensions, and filing process time window. The matrix value of the three-dimensional behavior matrix is ​​calculated by multiplying the base value and the preset base weight. For missing values ​​corresponding to behaviors that have not occurred in the matrix, the average value of similar entities in the same industry or the minimum value of all entities is used to fill in the missing values, resulting in a complete behavior matrix.

[0046] The final weights are calculated based on a preset two-level weight system and dynamic adjustment factors and then normalized. The preset two-level weight system includes safety verification weights and environmental correlation weights. The dynamic adjustment factors include verification effectiveness factors, behavioral consistency factors, environmental safety factors, and risk level factors. After normalization, the target weights for each behavioral dimension are obtained.

[0047] Based on the complete behavior matrix and target weights, a four-level tagging system is constructed. The four-level tagging system includes subject attribute layer tags, verification status layer tags, operation behavior layer tags, and environmental security layer tags. Each level of tags is generated through quantification threshold rules and logical association rules to obtain a behavior profile.

[0048] Furthermore, the comprehensive quantitative assessment of the credibility of the registered entity includes:

[0049] The comprehensive compliance index of the filing entity is determined based on the static attributes of the filing entity, the strength of the initial filing identity authentication, historical and real-time behavioral data, the normal sending success rate within a specified period, and the degree of behavioral deviation.

[0050] ;

[0051] in, The overall compliance index for registered entities; For static attribute weights; Quantify scores for static attributes; As a weight for identity authentication strength; Assign a score to the strength of identity authentication; Weighting based on historical behavior; Scoring based on historical behavior compliance; Weights for real-time behavior; Assigning scores for real-time behavioral compliance. Weighted by the success rate of transmission; This represents the normal success rate of sending. Weights for behavioral deviation; This refers to the overall deviation. The historical-real-time behavior collaborative incentive coefficient;

[0052] Based on the comprehensive compliance index and dynamic behavioral characteristics of the registered entity, the credibility score of the registered entity is determined;

[0053] ;

[0054] in, Assign a credibility score to the entity being registered; This is the time decay coefficient; The incentive coefficient for high-quality behavior; This represents the penalty coefficient for serious violations. The normalized value of the number of serious violations within a specified period; This is the deviation penalty coefficient; This is the deviation penalty threshold.

[0055] Compared with the prior art, the beneficial effects of the present invention are:

[0056] This invention introduces multi-factor authentication combined with behavioral profiling technology to ensure the authenticity and legitimacy of the registered entity; based on dynamic trust assessment and permission policy adjustment, it achieves intelligent access control, which not only adapts to normal business needs but also responds promptly to abnormal behavior, ensuring system security; and through hash operations and dynamic instruction binding, it ensures the uniqueness and immutability of each SMS request; based on a multi-layer verification and decision tree model-driven response mechanism, it further strengthens the ability to review and process SMS requests. This enables the system to not only effectively prevent security risks in SMS communication but also improve its flexibility and adaptability through dynamic adjustment and optimization mechanisms, ensuring efficient, secure, and reliable operation of SMS communication and providing users with robust information security protection. Attached Figure Description

[0057] Figure 1 This is a flowchart of the SMS signature intelligent filing and verification system of the present invention. Detailed Implementation

[0058] 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.

[0059] To address the inherent flaws of traditional registration systems that rely heavily on manual submission and qualification review, assigning a fixed signature and static API key upon approval, the following points should be considered: Static API keys, once leaked, can be abused indefinitely, creating serious information protection vulnerabilities; access control policies are typically fixed and cannot be dynamically adjusted based on historical sending patterns, content risk, recipient groups, etc.; they cannot distinguish between legitimate business sending and malicious fraudulent sending, lacking identification and intervention capabilities; and traditional SMS protocols themselves lack a strong security mechanism bound to registration information, failing to provide end-to-end data protection guarantees. Please refer to [link / reference needed]. Figure 1 This embodiment provides the following technical solution:

[0060] The SMS signature intelligent filing and verification system includes:

[0061] The intelligent filing unit is configured to register user accounts and file entity records, introduce multi-factor authentication, and build behavioral profiles. It continuously updates the normal behavior patterns of each filing entity, including but not limited to sending time distribution, frequency, content templates, target number sets, and success rates, forming a dynamic security baseline. The intelligent filing unit includes:

[0062] The multi-factor authentication module is configured to verify the legal identity credentials and official communication channels of the filing initiator, and to verify the biometrics of the authorized person responsible for the filing initiator; once all three are verified, the filing is confirmed as successful and a unique filing root identifier is generated; a behavioral profile is constructed and a compliant API communication key pair is assigned to it;

[0063] The multi-factor authentication module includes:

[0064] The first authentication module is configured to receive the legal identity core credential document submitted by the applicant for filing, which corresponds to the type of the sending entity. It extracts key fields from the document using image recognition technology and performs real-time interface verification with the database to confirm the authenticity and validity of the credential. For example, when an individual user submits a filing application and uploads proof of identity through the system interface, the first authentication module receives the proof of identity document, uses image recognition technology to extract the name, ID number, etc., and performs real-time interface verification with the personal information in the database to confirm the authenticity and validity of the credential.

[0065] The second authentication module is configured to receive one or more official communication channels associated with the applicant's identity and send randomly generated dynamic instructions to these official communication channels to verify the accuracy of the information provided by the applicant. For example, if an individual user submits their mobile phone number as an official communication channel, the second authentication module will send a randomly generated dynamic instruction, such as an SMS verification code, to the user's mobile phone number and require the user to fill in the verification code for verification, thus ensuring the authenticity of the mobile phone number.

[0066] The third authentication module is configured to guide the authorized person to complete a liveness detection sequence by calling the facial recognition service, and compare the captured biometric information with the baseline information in the identity information database to verify whether the current applicant is consistent with the authorized person's identity. For example, if an individual user performs fingerprint recognition verification through the system interface, the third authentication module calls the fingerprint recognition service to capture the user's fingerprint feature information and compares it with the baseline information in the identity information database to confirm that the current applicant is consistent with the individual user's identity. After the legal identity certificate verification, official communication channel verification, and biometric verification are all passed, the system confirms that the individual user's registration is successful. The system generates a unique registration root identifier, initializes the individual user's corresponding behavioral profile and security baseline, and assigns a compliant API communication key pair to them.

[0067] The audit log module is configured to record all operations, verification attempts, timestamps, IP addresses, and device fingerprint information throughout the multi-factor authentication process. Administrators can view the specific results of each verification step, such as the verification results of legal identity credentials, the verification content of official communication channels, and the comparison information of biometric verification. Based on the analysis results of the audit logs, administrators can take corresponding security measures, such as suspending the SMS sending permissions of the registered entity, requiring re-registration, or further investigating abnormal behavior. The module is set to read-only mode to ensure that the audit logs cannot be tampered with, for security auditing and accountability.

[0068] The access control unit is configured to perform dynamic access control based on behavioral profiles and account registration duration; the access control unit includes:

[0069] The dynamic trust assessment module is configured to comprehensively and quantitatively evaluate the trustworthiness of registered entities. Its assessment dimensions include: evaluating the static attributes of the registered entity, such as the entity's duration of registration (a longer duration results in a better basic trust assessment); evaluating the strength of identity authentication passed during the initial registration phase (higher authentication strength results in a better basic trust assessment); evaluating the historical and real-time behavioral data of the registered entity; evaluating the entity's successful sending rate within a specified period; evaluating the deviation of the entity's sending behavior from historical records in terms of time distribution, frequency fluctuations, and content templates; and receiving real-time risk signals. For example, for newly registered applications, the initial sending frequency limit is set to a lower level; when the behavioral profile shows trustworthy behavior, access restrictions are automatically relaxed; conversely, when abnormal behavior is detected, permissions are immediately tightened or temporary bans are implemented, achieving intelligent access control.

[0070] The dynamic permission policy module is configured with multiple permission policy levels associated with the trust assessment result range, forming a policy library. The permission policy has multiple control dimensions, including: sending frequency thresholds, such as the maximum number of messages that can be sent per second, per minute, per hour, and per day; allowed time period windows for sending; allowed target number ranges or number segments; and allowed range of SMS content templates.

[0071] The dynamic permission execution module is configured to automatically adjust the corresponding permission policy based on real-time evaluation results. Its execution logic includes: for newly registered entities, automatically applying the initial permission policy; the initial permission policy includes: a maximum of 50 messages per hour, a allowed sending time of 8:00-22:00 daily, a range of allowed target numbers from the individual's address book, and allowed SMS content templates that are already registered personal notification templates; when the evaluation result is consistently better than the threshold and enters the high trust zone, automatically triggering permission enhancement operations, such as: relaxing the sending frequency limit, extending the sending time window, or opening more content templates; when the evaluation result drops to the low trust or risk zone due to detected abnormal behavior, immediately triggering permission tightening operations or temporary blocking operations, such as: restoring to the initial permission or imposing stricter restrictions, and temporary blocking will suspend sending permissions and trigger management alarms.

[0072] The closed-loop feedback module is configured to provide real-time feedback on execution results and new entity behavior data, dynamically update behavior profiles, and optimize the assessment accuracy and predictive capabilities of the dynamic trust assessment module.

[0073] A dynamic security unit is configured to initialize a seed for each registered entity. The hash value of the seed is mixed with the entity's registration root identifier and a fragment of the system master key to ensure that the seed cannot be ported or forged between different entities. Before each request is sent, the client uses IoT hash operations to iteratively calculate the initialization seed, generating a one-time dynamic instruction. The dynamic security unit includes:

[0074] The instruction binding module is configured to use a hash function to perform hash calculations on the SMS content and bind the hash value to dynamic instructions to ensure that the request is non-replayable and tamper-proof. Content binding includes: binding dynamic instructions to SMS content to ensure the integrity and authenticity of the SMS content; and binding dynamic instructions to the timestamp of the sent request to ensure the timeliness and non-replayability of the request. The timestamp is used to ensure that each request is unique and to prevent replay attacks.

[0075] The binding verification module is configured to verify the validity of the binding relationship on the receiving end, ensuring the consistency of dynamic commands, SMS content, and timestamps; if the binding relationship is tampered with or the timestamp is invalid, the receiving end will reject the request.

[0076] The instruction verification module is configured to parse received dynamic instructions, extract their hash values ​​and iteration counts, and verify whether the format and content of the dynamic instructions meet expectations. Specifically, it uses a hash operation algorithm and a seed to verify the received dynamic instructions, comparing the hash values ​​with the received hash values ​​to ensure the validity of the dynamic instructions; it also verifies the validity of the bound SMS content and timestamp to ensure the integrity and timeliness of the request; if the verification fails, the receiving end will reject the request and record detailed information.

[0077] The key update module sets a lifecycle for each seed, such as based on time or usage count. When the lifecycle ends, it automatically triggers the key update process, generates a new seed, and distributes it securely to achieve forward security. When a security threat or change in entity registration status is detected, the previous seed is automatically revoked and a new seed is updated to minimize the risk window.

[0078] The verification and execution unit is configured to intercept and verify all outbound SMS requests at the millisecond level. Verification content includes: signature and token verification: verifying the legality of the signature and the validity of the dynamic token; policy compliance verification: checking whether the request complies with the current access control policy defined by the behavior profiling engine, such as frequency limits; real-time risk scoring: performing real-time risk analysis based on request content and recipient characteristics, such as whether it is a new number, complaint rate, etc. The verification and execution unit includes:

[0079] The response action module is configured to pre-set corresponding actions based on historical verification results. A decision tree model is used to cluster and associate the actions with the verification results to form an action library.

[0080] The verification response module is configured to obtain multi-layer verification results, associate them with pre-defined actions in the action library, and execute corresponding handling strategies. For example:

Pass

Reject

Delayed Sending

Secondary Verification

Transfer to Manual Review

Simultaneous Blocking

Handling Pipeline

[0081] The secure communication protocol unit is configured to embed dynamic instructions, signature ID, and content hash value as standard fields into the communication protocol request header, enabling the receiving end to verify the integrity and authenticity of the request by querying the public key or shared seed of the filing system; thereby realizing an inherently secure communication protocol and achieving end-to-end data and information protection.

[0082] The beneficial effects achieved by the above are as follows: By introducing multi-factor authentication combined with behavioral profiling technology, the authenticity and legality of the registered entity are ensured; based on dynamic trust assessment and permission policy adjustment, intelligent access control is achieved, which can not only adapt to normal business needs but also respond to abnormal behavior in a timely manner, ensuring system security; through hash operations and dynamic instruction binding, the uniqueness and immutability of each SMS request are ensured; based on a multi-layer verification and decision tree model-driven response mechanism, the ability to review and handle SMS requests is further strengthened; thus, the system not only effectively prevents security risks in SMS communication but also improves the system's flexibility and adaptability through dynamic adjustment and optimization mechanisms, ensuring the efficient, secure, and reliable operation of SMS communication and providing users with strong information security protection.

[0083] Working principle: By utilizing multi-factor authentication and audit logs, the system ensures the authenticity and legitimacy of the registered entity and records the authentication process; based on dynamic trust assessment, it dynamically adjusts access permissions and updates behavioral profiles; through instruction binding and key updates, it ensures the uniqueness and long-term security of SMS requests; using verification and execution units for millisecond-level interception and multi-layer verification, it executes corresponding handling strategies; and based on the secure communication protocol unit, it embeds dynamic instructions and other information to achieve end-to-end secure communication; through the collaborative work of multiple modules, it effectively ensures the legitimacy, integrity, and security of SMS communication, and improves the system's flexibility and adaptability.

[0084] The method for constructing behavioral profiles includes:

[0085] Obtain multi-dimensional associated data of the SMS signature filing entity. The multi-dimensional associated data includes core verification data, operational behavior data, and environmental associated data. The core verification data includes legal identity certificate data, official communication channel data, and biometric data. The operational behavior data includes filing operation logs and interaction process data. The environmental associated data includes device feature data, network feature data, and scene feature data.

[0086] The collected multi-dimensional associated data is subjected to dual integrity verification, which includes original data hash verification and feature consistency verification. Valid associated data is obtained after the verification is passed.

[0087] The effective associated data is subjected to hierarchical desensitization and security hardening processing. The effective associated data is divided into first-level sensitive data, second-level sensitive data and third-level sensitive data. According to the preset hierarchical desensitization rules, the sensitive data of different levels is desensitized in a non-deletion manner. Then, the desensitized data is stored by double hashing, local desensitization processing and AES-256 algorithm encryption to obtain secure data.

[0088] The security processing data is split into fine-grained behavioral dimensions, divided into single operation behaviors, associated verification behaviors, and environment-related behaviors. A three-dimensional behavior matrix is ​​constructed based on the filing entity ID, behavior dimensions, and filing process time window. The matrix value of the three-dimensional behavior matrix is ​​calculated by multiplying the base value and the preset base weight. For missing values ​​corresponding to behaviors that have not occurred in the matrix, the average value of similar entities in the same industry or the minimum value of all entities is used to fill in the missing values, resulting in a complete behavior matrix.

[0089] The final weights are calculated based on a preset two-level weight system and dynamic adjustment factors and then normalized. The preset two-level weight system includes safety verification weights and environmental correlation weights. The dynamic adjustment factors include verification effectiveness factors, behavioral consistency factors, environmental safety factors, and risk level factors. After normalization, the target weights for each behavioral dimension are obtained.

[0090] Based on the complete behavior matrix and target weights, a four-level tagging system is constructed. The four-level tagging system includes subject attribute layer tags, verification status layer tags, operation behavior layer tags, and environmental security layer tags. Each level of tags is generated through quantification threshold rules and logical association rules to obtain a behavior profile.

[0091] In this embodiment, the dual integrity verification includes original data hash verification and feature consistency verification, including:

[0092] Raw data hash verification: Calculate the SHA-256 hash value for each type of collected data and store it in a local secure cache. After the data is transmitted to the processing node, the hash value is recalculated. If the two are consistent, the verification is successful.

[0093] Feature consistency verification: For key fields of legal identity credentials (such as the first 8 digits of the business license number, the first 6 digits of the ID card number plus the last 4 digits), additional feature hash values ​​are calculated and compared with the verification results returned by the government platform to ensure that the data has not been tampered with.

[0094] Verification failure handling: If the hash value is inconsistent or the feature verification fails, the data re-collection process is triggered, and the failure log is recorded at the same time.

[0095] In this embodiment, Level 1 sensitive data includes: ID card number, complete business license number, original biometric data, and complete mobile phone number; Level 2 sensitive data includes: IP address, MAC address, and detailed hardware information of the operating device; Level 3 sensitive data includes: the name of the legal representative of the filing entity and the complete corporate email address.

[0096] In this embodiment, the hierarchical desensitization rules are as follows: Level 1 sensitive data: ID card number → first 6 digits + "" + last 4 digits, business license number → first 8 digits + "" + last 4 digits, mobile phone number → first 3 digits + "" + last 4 digits, original biometric data → converted to irreversible feature code (such as facial feature point vector hash value); Level 2 sensitive data: IP address → IPv4 hides the last segment (such as 192.168.1.), IPv6 hides the last 64 digits, MAC address → hides the middle 4 digits (such as AA:BB:**:CC:DD); Level 3 sensitive data: legal representative's name → surname + "" (such as Zhang), corporate email → first 3 characters of username + "" + domain name.

[0097] In this embodiment, dual hash storage is implemented: SHA-256 hash values ​​are calculated for both the data before and after de-identification, and a mapping relationship is established. The hash calculation and mapping relationship generation are both completed on the local hardware analysis platform. Local de-identification processing is also implemented: all de-identification operations are completed on the local hardware analysis platform, and the original sensitive data and the hash value before de-identification are not transmitted to the cloud or third-party nodes. Encrypted storage is also implemented: the de-identified data is encrypted using the AES-256 algorithm and stored in a distributed secure database. Only authorized filing verification nodes can decrypt and access the data through the filing root identifier.

[0098] In this embodiment, single operational behavior (S-type): independent filing operation actions, including S1 (submitting legal identity certificate), S2 (submitting communication channel information), S3 (submitting biometric data), S4 (submitting signature filing content), S5 (modifying filing information), S6 (withdrawing filing application), S7 (supplementing filing materials), and S8 (operation integrity score); associated verification behavior (U-type): a combination of multi-factor verification behavior sequences, including U1 (S1→S2→S3, normal verification order), U2 (S1→S3→S2, abnormal order), U3 (S1→S2→verification failure→resubmit verification→S3, retest verification), and U4 (S3→S1→S2→S4, complete verification + submission); environmental associated behavior (E-type): the association characteristics between the operating environment and the subject, including E1 (device and subject binding degree), E2 (IP location and registration address consistency), E3 (operation time period and subject working time period matching degree), and E4 (network type stability).

[0099] In this embodiment, the matrix is ​​defined as follows: A matrix M (Subject ID, Behavior Dimension, Time Window) is constructed, where "Subject ID" is the associated code of the filing root identifier, "Time Window" is based on a single filing process (from the first operation to the completion / rejection of the filing review, with a maximum of 72 hours), and "Behavior Dimension" covers all sub-items of categories S, U, and E; Matrix value calculation: Base value: For category S behaviors, S1-S7 take the frequency of occurrence (e.g., if S5 occurs 2 times, the base value = 2), S8 (operation integrity score) = 1 - (number of occurrences of S6 + number of occurrences of S7) / (total number of occurrences of S1-S7), and the result is normalized to [0,1]; For category U behaviors, the sequence occurrence count × preset coefficient (U1 / U4 coefficient = 1.0, U2 / U3 coefficient = 0.5); For category E behaviors... Calculate the relevance score (out of 10, e.g., 10 points for complete consistency and 2 points for complete inconsistency in E2); Weight adjustment: Preset basic weights (S class: S3=5, S1=3, S2=3, S4=2, S5=1, S6=1, S7=2, S8=4; U class: U4=8, U1=6, U3=4, U2=2; E class: E1=4, E2=3, E3=2, E4=2), matrix value = basic value × basic weight; Matrix completion: For dimensions of behaviors that have not occurred, use a pre-built industry benchmark library for completion. "Similar entities" are defined as a set of entities with the same business license industry code (GB / T4754) and the same entity type. The completion value = the average value of similar entities in this dimension × 0.8; if there is no similar entity data, use the minimum value of all entities in this dimension for completion.

[0100] In this embodiment, a two-level weighting system is established with a total weight of 10, where: security verification weight (S-class + U-class) accounts for 60% (6 points), and environment-related weight (E-class) accounts for 40% (4 points); Dynamic adjustment of factors is applied hierarchically: Verification validity factor: only applies to security verification weights; a 20% increase is applied if all three factors pass on the first attempt; a 10% increase is applied if the first attempt passes; a 15% decrease is applied if the second or more attempts pass; Environment security factor: only applies to environment-related weights; a 15% increase is applied if the IP address and registration address are consistent and the device binding history is ≥3 months; a 25% decrease is applied if the distance between the IP address and registration address is >500 kilometers and there is no historical binding record; Behavioral consistency factor: applies to the S8 dimension base value; a 30% increase is applied if the consistency between the current operation and the subject's historical filing behavior is ≥80% (consistent operation time, device type, IP region); a 20% decrease is applied if the consistency is <50%; Risk level factor: applies to... Category weights: In high-risk scenarios (IP location and registered address distance > 500 km + first-time filing + operation time between 22:00-6:00), the weights for security verification and environment-related categories are doubled. Weight calculation process: Calculate the adjusted weights: Security verification weight = 6 × (1 + verification validity factor + risk level factor adjustment value); Environment-related weight = 4 × (1 + environmental security factor + risk level factor adjustment value); Allocate category weights to sub-dimensions: S-category dimension weight = Security verification weight × (S-category basic weight / S-category total basic weight); U-category dimension weight = Security verification weight × (U-category basic weight / U-category total basic weight); E-category dimension weight = Environment-related weight × (E-category basic weight / E-category total basic weight); Apply the behavior consistency factor to the S8 dimension: S8 final weight = S8 allocated weight × (1 + behavior consistency factor); Normalization: Sum of all dimension weights is scaled to 10 to obtain the target weight.

[0101] In this embodiment, a multi-level behavioral profile tagging system is constructed: Level 1 Tag (Subject Attribute Layer): T1 (Subject Type): Enterprise Subject / Individual Subject (based on legal identity certificate type); T2 (Industry Attribute): Finance / Education / E-commerce / Government Affairs / Other (based on business license industry code GB / T4754 or keywords in the filing signature content); T3 (Filing Level): First-time Filing / Second-time Filing / Multiple-time Filing (based on historical filing records; no history indicates first-time filing); Level 2 Tag (Verification Status Layer): V1 (Verification Completeness): All three factors passed / Partial pass (number of pass items < 3); V2 (Validation efficiency): One-time pass (no retries) / Pass with 1 retry / Pass with ≥ 2 retries; V3 (Validation order): Normal order (U1 / U4 sequence percentage ≥ 80%) / Abnormal order (U2 / U3 sequence percentage > 20%); Level 3 labels (Operation behavior layer): O1 (Operation frequency): High-frequency operation (total frequency of S1-S7 ≥ 8) / Medium-frequency operation (3 ≤ total frequency < 8) / Low-frequency operation (total frequency < 3); O2 (Operation stability): Stable operation (variance of operation time interval < 300 seconds) / Fluctuation operation (300 seconds ≤ variance < 1000 seconds) / Abnormal fluctuation (variance ≥ 1000 seconds); O3 (Operational Integrity): Complete operation (S8 score ≥ 0.8 and no S6) / Interrupted operation (S8 score < 0.5 or S6 executed but S4 not completed); Level 4 Labels (Environmental Security Layer): S1 (IP Consistency): Fully consistent (E2 score ≥ 9) / Partially consistent (6 ≤ E2 score < 9) / Completely inconsistent (E2 score < 6); S2 (Device Binding): High binding (E1 score ≥ 8) / Medium binding (5 ≤ E1 score < 9) <8) / Low binding (E1 score <5); S3 (Time period matching degree): Match (E3 score ≥ 8) / Partial match (5 ≤ E3 score < 8) / No match (E3 score < 5); Tag generation rules only use quantitative threshold rules and logical association rules: Quantitative threshold rules: If the total frequency of S1-S7 is ≥ 8, then O1 = high frequency operation; E2 score ≥ 9, then S1 = completely consistent; Logical association rules: If the number of verification retries = 0 and the number of passed items = 3, then V2 = passed in one go; If S6 is executed and S4 is not resubmitted, then O3 = interrupted operation.

[0102] In this embodiment, the behavioral profile is further associated and stored based on the unique root identifier of the filing entity. The associated storage adopts a distributed document storage architecture, and the stored content includes basic information, core data, and auxiliary information. The basic information includes the profile ID, the root identifier of the filing entity, and time information. The core data includes a complete behavioral matrix, a target weight table, and a four-level tag system. The auxiliary information includes de-identified logs, verification records, and iteration history, thus completing the construction of the behavioral profile. The industry benchmark library is constructed during system initialization and is generated based on the industry code of the business license and the entity type cluster of historical filing entities for matrix completion.

[0103] The working principle and beneficial effects of the above technical solution are as follows: Through dual integrity verification, it can effectively identify possible errors, omissions, or tampering issues in the collected multi-dimensional related data; only data that passes the verification will be recognized as valid related data, which ensures that the data used for subsequent analysis and profile construction is accurate and complete, providing a solid data foundation for accurate behavioral profiles; accurate data can avoid behavioral analysis bias caused by inaccurate data, enabling the behavioral profiles built based on this data to more realistically reflect the actual behavioral characteristics and status of the SMS signature filing entity; valid related data is divided into first-level sensitive data, second-level sensitive data, and third-level sensitive data, and non-deletion desensitization processing is performed according to preset hierarchical desensitization rules. This tiered processing approach employs different protection measures based on the sensitivity of the data, maximizing the protection of the privacy and sensitive information of the filing entity without losing critical information, thus reducing the risk of data leakage. The de-identified data undergoes dual hash storage, local de-identification processing, and AES-256 encryption, further enhancing data security. Even if the data is attacked during storage or transmission, attackers will find it difficult to obtain valuable sensitive information, ensuring data confidentiality and integrity. Based on a pre-set two-level weighting system and dynamically adjusting factors, the final weights are calculated and normalized. This approach comprehensively considers the impact of multiple factors on behavioral profiles, dynamically adjusting the weights of each behavioral dimension according to different situations, making behavioral profiles more flexible and adaptable to different business scenarios and risk conditions.

[0104] The comprehensive quantitative assessment of the credibility of the registered entity includes:

[0105] The comprehensive compliance index of the filing entity is determined based on the static attributes of the filing entity, the strength of the initial filing identity authentication, historical and real-time behavioral data, the normal sending success rate within a specified period, and the degree of behavioral deviation.

[0106] ;

[0107] in, The overall compliance index for registered entities; For static attribute weights; Quantify scores for static attributes; As a weight for identity authentication strength; Assign a score to the strength of identity authentication; Weighting based on historical behavior; Scoring based on historical behavior compliance; Weights for real-time behavior; Assigning scores for real-time behavioral compliance. Weighted by the success rate of transmission; This represents the normal success rate of sending. Weights for behavioral deviation; This refers to the overall deviation. The historical-real-time behavior collaborative incentive coefficient;

[0108] Based on the comprehensive compliance index and dynamic behavioral characteristics of the registered entity, the credibility score of the registered entity is determined;

[0109] ;

[0110] in, Assign a credibility score to the entity being registered; This is the time decay coefficient; The incentive coefficient for high-quality behavior; This represents the penalty coefficient for serious violations. The normalized value of the number of serious violations within a specified period; This is the deviation penalty coefficient; This is the deviation penalty threshold.

[0111] In this embodiment, This refers to the overall deviation. ; This indicates a deviation in the time distribution. This is due to frequency fluctuation deviation; Content template deviation; .

[0112] In this embodiment, the static attribute quantification score is used. Identity authentication strength score Historical compliance score Real-time behavioral compliance score Normal transmission success rate Static attribute quantification score Time distribution deviation Frequency fluctuation deviation Content template deviation .

[0113] In this embodiment, , , , , , The sum of all values ​​is 1, and all are greater than 0.

[0114] In this embodiment, the collaborative excitation coefficient Incentive coefficient for high-quality behavior Penalty coefficient for serious violations Deviation penalty coefficient Deviation penalty threshold .

[0115] In this embodiment, the time decay coefficient , The number of days since the last active period of the registered entity; decay rate. The value is 0.1.

[0116] The working principle and beneficial effects of the above technical solution are as follows: It calculates the comprehensive compliance index and credibility score of the registered entity, quantifying the credibility of the registered entity. This quantification makes the credibility of different registered entities comparable, facilitating managers to sort and screen numerous registered entities and quickly identify entities with high and low credibility. By comprehensively evaluating the credibility of registered entities, it is possible to promptly identify registered entities with low credibility and provide early warnings of potential risks. For example, when the credibility score of a registered entity falls below a certain threshold, relevant personnel can take timely measures, such as strengthening supervision and conducting investigations, to prevent the risk from escalating. The quantified credibility score can provide strong support for relevant decision-making. In decision-making processes such as resource allocation and cooperation selection, managers can prioritize registered entities with higher credibility, reducing business risks and improving the scientific nature and accuracy of decision-making.

[0117] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0118] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A smart SMS signature filing and verification system, characterized in that, include: The intelligent filing unit is configured to register user accounts and file entity records, introduce multi-factor authentication, and build behavioral profiles; The access control unit is configured to perform dynamic access control based on behavioral profiles and account registration duration. The dynamic security unit is configured to initialize a seed for each registered entity. Before each request is sent, the client uses IoT hash operation to iteratively calculate the initialization seed and generate a one-time dynamic instruction. The verification and execution unit is configured to intercept and verify all outbound SMS requests at the millisecond level. The verification content includes: signature and token verification, policy compliance verification, and real-time risk scoring. The secure communication protocol unit is configured to embed dynamic instructions, signature IDs, and content hash values ​​as standard fields into the communication protocol request header, enabling the receiving end to verify the integrity and authenticity of the request by querying the public key or shared seed of the filing system.

2. The SMS signature intelligent filing and verification system according to claim 1, characterized in that, The intelligent filing unit includes: The multi-factor authentication module is configured to verify the legal identity credentials and official communication channels of the filing initiator, and to verify the biometrics of the authorized person responsible for the filing initiator. Once all three verifications are successful, the entity registration is confirmed as successful, and a unique registration root identifier is generated. Build behavioral profiles and assign them compliant API communication key pairs; The audit log module is configured to record all operations, verification attempts, timestamps, IP addresses, and device fingerprint information throughout the multi-factor authentication process, and is set to read-only mode.

3. The SMS signature intelligent filing and verification system according to claim 2, characterized in that, The multi-factor authentication module includes: The first authentication module is configured to receive the legal identity core credential document submitted by the applicant for filing, which corresponds to the type of the sending entity. It extracts the key fields in the document through image recognition technology and performs real-time interface verification with the database to confirm the authenticity and validity of the credential. The second authentication module is configured to receive one or more official communication channels associated with the applicant's identity and send randomly generated dynamic instructions to the official communication channels to verify whether the content filled in by the applicant is correct. The third authentication module is configured to guide the authorized person to complete a liveness detection sequence by calling the face recognition service, and compare the captured biometric information with the benchmark information in the identity information database to verify whether the current applicant is consistent with the identity of the authorized person.

4. The SMS signature intelligent filing and verification system according to claim 1, characterized in that, The access control unit includes: The dynamic trust assessment module is configured to comprehensively and quantitatively assess the credibility of the registered entity. Its assessment dimensions include: assessing the static attributes of the registered entity, assessing the strength of the identity authentication passed by the registered entity in the initial registration stage, assessing the historical and real-time behavioral data of the registered entity, assessing the normal sending success rate of the registered entity within a specified period, and assessing the degree of deviation of the registered entity's sending behavior in time distribution, frequency fluctuation, and content template from historical records. The dynamic permission policy module is configured with multiple permission policy levels that are pre-defined and associated with the trust assessment result range, forming a policy library.

5. The SMS signature intelligent filing and verification system according to claim 4, characterized in that, The access control unit further includes: The permission dynamic execution module is configured to automatically adjust the corresponding permission policy based on real-time evaluation results; The closed-loop feedback module is configured to provide real-time feedback on execution results and new entity behavior data, dynamically update behavior profiles, and optimize the assessment accuracy and predictive capabilities of the dynamic trust assessment module.

6. The SMS signature intelligent filing and verification system according to claim 1, characterized in that, The dynamic security unit includes: The instruction binding module is configured to use a hash function to perform a hash operation on the SMS content, and bind the hash value with a dynamic instruction to ensure that the request is non-replayable and tamper-proof. The binding verification module is configured to verify the validity of the binding relationship on the receiving end, ensuring the consistency of dynamic commands, SMS content, and timestamps; if the binding relationship is tampered with or the timestamp is invalid, the receiving end will reject the request.

7. The SMS signature intelligent filing and verification system according to claim 6, characterized in that, The dynamic security unit further includes: The instruction verification module is configured to parse the received dynamic instructions, extract their hash values ​​and iteration counts, and verify whether the format and content of the dynamic instructions meet expectations. The key update module sets a lifecycle for each seed. When the lifecycle ends, it automatically triggers the key update process to generate a new seed and distribute it securely.

8. The SMS signature intelligent filing and verification system according to claim 1, characterized in that, The verification and execution unit includes: The response action module is configured to pre-set corresponding actions based on historical verification results. A decision tree model is used to cluster and associate the actions with the verification results to form an action library. The verification response module is configured to obtain multi-level verification results, associate them with preset handling actions in the action library, and execute the corresponding handling strategies.

9. The SMS signature intelligent filing and verification system according to claim 2, characterized in that, The method for constructing behavioral profiles includes: Obtain multi-dimensional associated data of the SMS signature filing entity. The multi-dimensional associated data includes core verification data, operational behavior data, and environmental associated data. The core verification data includes legal identity certificate data, official communication channel data, and biometric data. The operational behavior data includes filing operation logs and interaction process data. The environmental associated data includes device feature data, network feature data, and scene feature data. The collected multi-dimensional associated data is subjected to dual integrity verification, which includes original data hash verification and feature consistency verification. Valid associated data is obtained after the verification is passed. The effective associated data is subjected to hierarchical desensitization and security hardening processing. The effective associated data is divided into first-level sensitive data, second-level sensitive data and third-level sensitive data. According to the preset hierarchical desensitization rules, the sensitive data of different levels is desensitized in a non-deletion manner. Then, the desensitized data is stored by double hashing, local desensitization processing and AES-256 algorithm encryption to obtain secure data. The security processing data is split into fine-grained behavioral dimensions, divided into single operation behaviors, associated verification behaviors, and environment-related behaviors. A three-dimensional behavior matrix is ​​constructed based on the filing entity ID, behavior dimensions, and filing process time window. The matrix value of the three-dimensional behavior matrix is ​​calculated by multiplying the base value and the preset base weight. For missing values ​​corresponding to behaviors that have not occurred in the matrix, the average value of similar entities in the same industry or the minimum value of all entities is used to fill in the missing values, resulting in a complete behavior matrix. The final weights are calculated based on a preset two-level weight system and dynamic adjustment factors and then normalized. The preset two-level weight system includes safety verification weights and environmental correlation weights. The dynamic adjustment factors include verification effectiveness factors, behavioral consistency factors, environmental safety factors, and risk level factors. After normalization, the target weights for each behavioral dimension are obtained. Based on the complete behavior matrix and target weights, a four-level tagging system is constructed. The four-level tagging system includes subject attribute layer tags, verification status layer tags, operation behavior layer tags, and environmental security layer tags. Each level of tags is generated through quantification threshold rules and logical association rules to obtain a behavior profile.

10. The SMS signature intelligent filing and verification system according to claim 4, characterized in that, The comprehensive quantitative assessment of the credibility of the registered entity includes: The comprehensive compliance index of the filing entity is determined based on the static attributes of the filing entity, the strength of the initial filing identity authentication, historical and real-time behavioral data, the normal sending success rate within a specified period, and the degree of behavioral deviation. ； in, The overall compliance index for registered entities; For static attribute weights; Quantify scores for static attributes; As a weight for identity authentication strength; Assign a score to the strength of identity authentication; Weighting based on historical behavior; Scoring based on historical behavior compliance; Weights for real-time behavior; Assigning scores for real-time behavioral compliance. Weighted by the success rate of transmission; This represents the normal success rate of sending. Weights for behavioral deviation; This refers to the overall deviation. The historical-real-time behavior collaborative incentive coefficient; Based on the comprehensive compliance index and dynamic behavioral characteristics of the registered entity, the credibility score of the registered entity is determined; ； in, Assign a credibility score to the entity being registered; This is the time decay coefficient; The incentive coefficient for high-quality behavior; This represents the penalty coefficient for serious violations. The normalized value of the number of serious violations within a specified period; This is the deviation penalty coefficient; This is the deviation penalty threshold.

Images