Digital collection verification method and system

Abstract

The invention relates to the technical field of digital collection verification, and particularly discloses a digital collection verification method and system, and the method comprises the steps: obtaining the metadata of a target digital collection from a block chain, and synchronously obtaining the historical transaction record of the target digital collection and the operation behavior data of a current dealer; generating a dynamic encryption key based on the creation hash and the issuing timestamp in the metadata of the target digital collection and the historical transaction record, and generating a time-sensitive verification token based on the dynamic encryption key; generating a multi-chain verification result based on the hash operation result of the dynamic encryption key, the time sequence characteristics of the historical transaction data, the operation behavior data of the current trader and the time-sensitive verification token; performing weighted voting on the multi-chain verification result based on a consensus mechanism to obtain a final verification result; multi-chain verification results are effectively integrated, the one-sidedness of single verification is avoided, and the reliability and credibility of the whole digital collection verification process are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of digital collection verification, and in particular to a digital collection verification method and system. Background Art

[0002] Traditional digital collection verification technology only relies on blockchain hash values ​​to verify digital collections, without comprehensively considering the multi-dimensional information of collections and traders, and cannot comprehensively evaluate the authenticity and legitimacy of digital collections in complex trading environments. Moreover, based on fixed hash values ​​and other information, it does not fully consider the dynamic changes of digital collections at different time points and trading stages. For example, the impact of factors such as issuance time and multiple transactions on the value and authenticity of collections makes it difficult to cope with new situations and potential risks that arise during the transaction process. Moreover, the current trader's operational behavior data is not analyzed, and it is impossible to determine whether the transaction behavior itself is abnormal, which may make malicious transactions or fraudulent behaviors difficult to detect in time, affecting the security of digital collection transactions. In addition, verification decisions are made based only on a single condition or simple rule, and there is a lack of scientific decision-making methods such as weighted voting based on a consensus mechanism, which may result in inaccurate and inauthoritative verification results, and fail to fully reflect the true status of digital collections.

[0003] Therefore, the present invention proposes a verification method and system for digital collections. Summary of the invention

[0004] The present invention provides a verification method and system for digital collections, including: by obtaining metadata, historical transaction records and current trader operation behavior data of target digital collections from blockchain, rich and comprehensive information is collected, providing sufficient data support for verification. The creation hash in the metadata can ensure the authenticity of the source of the collection creation, the issuance timestamp clarifies the issuance time sequence, the historical transaction records reflect the circulation of the collection, and the operation behavior data is related to the real-time dynamics of the current transaction. These data complement each other and comprehensively outline the "identity information" and transaction trajectory of the digital collection, laying a solid foundation for subsequent verification. A dynamic encryption key is generated based on the creation hash, issuance timestamp and historical transaction records in the metadata, and a time-sensitive verification token is further generated. The dynamic encryption key combines the key attributes and transaction history of the collection, has dynamic change characteristics, and enhances security; the time-sensitive verification token introduces the limitation of the time dimension, so that the verification is closely related to time, further improves the timeliness and security of the verification, effectively prevents delay attacks or tampering during the verification process, and ensures the effectiveness of the verification within a specific time range. The hash operation results of the dynamic encryption key, the time series characteristics of historical transaction data, the current trader's operation behavior data, and the time-sensitive verification token are used to generate multi-chain verification results, and digital collections are verified from multiple dimensions. The hash operation results provide verification at the encryption level, the time series characteristics reflect the timing rules of transactions, and the operation behavior data considers the actual situation of current transactions. The multi-chain verification results combine these factors to provide a more comprehensive and three-dimensional verification perspective, greatly improving the accuracy and reliability of verification. Based on the consensus mechanism, the multi-chain verification results are weighted voted to obtain the final verification results, and the consensus mechanism is used to ensure the fairness and transparency of the verification process. The weighted voting method takes into account the importance or credibility differences of different verification results, making the final verification results more authoritative and credible. This method effectively integrates the multi-chain verification results, avoids the one-sidedness of a single verification, enhances the reliability and credibility of the entire digital collection verification process, provides a strong guarantee for the transaction security and authenticity of digital collections, and promotes the healthy and orderly development of the digital collection market.

[0005] The present invention provides a method for verifying a digital collection, comprising: S1: Obtain the metadata of the target digital collection from the blockchain, and simultaneously obtain the historical transaction records of the target digital collection and the operation behavior data of the current trader; S2: Generate a dynamic encryption key based on the creation hash and release timestamp in the metadata of the target digital collection and the historical transaction record, and generate a time-sensitive verification token based on the dynamic encryption key; S3: Generates multi-chain verification results based on the hash operation results of the dynamic encryption key, the time series characteristics of historical transaction data, the operation behavior data of the current trader, and the time-sensitive verification token; S4: Perform weighted voting on the multi-chain verification results based on the consensus mechanism to obtain the final verification result.

[0006] Optionally, the verification method of the digital collection generates a dynamic encryption key based on the creation hash and the release timestamp in the metadata of the target digital collection and the historical transaction record, including: ; In the formula, is a dynamic encryption key, is a shared key generated based on the elliptic curve Diffie-Hellman algorithm, is the creator's public key, Is the release timestamp, is the exclusive OR operator, The hash message authentication code value calculated by the Merkle tree hash root based on historical transaction records and the pre-assigned basic key of the blockchain system and the HMAC-SHA3-256 algorithm. The Merkle tree hash root of historical transaction records. The base key pre-assigned for the blockchain system.

[0007] Optionally, the verification method of the digital collection generates a time-sensitive verification token based on a dynamic encryption key, including: ; In the formula, Token is a time-sensitive verification token. It is a hash message authentication code value calculated based on the dynamic encryption key, the Merkle tree hash root of historical transaction records, the Unix timestamp, the current transaction amount, and the HMAC-SHA3-256 algorithm. is a dynamic encryption key, The Merkle tree hash root of historical transaction records. is the exclusive OR operator, The first 16-bit integer of the Unix timestamp. is the transaction amount of the current transaction, It is the time window identifier, which consists of the first 8 binary bits of the timestamp.

[0008] Optionally, the verification method of the digital collection, S3: based on the hash operation result of the dynamic encryption key, the time series characteristics of the historical transaction data, the operation behavior data of the current trader and the time-sensitive verification token, a multi-chain verification result is generated, including: Perform a consistency check between the hash operation result of the dynamic encryption key and the genesis hash value stored in the smart contract to obtain the metadata integrity check result; The LSTM network based on the attention mechanism is used to analyze the time series characteristics of historical transaction data, generate transaction behavior feature vectors, and verify the transaction behavior based on the transaction behavior feature vectors to obtain the transaction behavior verification results; Based on the current trader's operation behavior data and time-sensitive verification token, verify the identity association between the current trader and the genesis trader, and obtain the identity association verification result; Among them, the multi-chain verification results include metadata integrity verification results, transaction behavior verification results, and identity association verification results.

[0009] Optionally, the verification method of the digital collection performs consistency verification on the hash operation result of the dynamic encryption key and the genesis hash value stored in the smart contract to obtain the metadata integrity verification result, including: Performing a hash operation on the dynamic encryption key based on a preset encryption function to obtain a hash operation result; Determine whether the hash operation result of the dynamic encryption key is consistent with the genesis hash value stored in the smart contract, and obtain the metadata integrity verification result.

[0010] Optionally, the verification method of the digital collection uses an LSTM network based on an attention mechanism to analyze the time series characteristics of historical transaction data, generate a transaction behavior feature vector, and perform transaction behavior verification based on the transaction behavior feature vector to obtain a transaction behavior verification result, including: Convert historical transaction data into time series format to obtain transaction record information for multiple time steps; Based on the LSTM network, all transaction record information is processed time-by-time to obtain the hidden state of the LSTM network at each time step; Calculate the attention weight of the LSTM network at time step t to the i-th time step in the historical transaction data; The historical transaction feature vector is weighted based on the attention weight of the LSTM network on the i-th time step in the historical transaction data at time step t to obtain the context vector; The output of the LSTM network based on the current transaction data is concatenated with the context vector to obtain the transaction behavior feature vector; The transaction behavior feature vector is input into the transaction behavior legitimacy classification model to obtain the transaction behavior verification result.

[0011] Optionally, the verification method of the digital collection calculates the attention weight of the LSTM network at time step t to the i-th time step in the historical transaction data, including: ; In the formula, is the attention weight of the LSTM network to the i-th time step in the historical transaction data at time step t, exp is an exponential function with the natural constant e as the base, v is a trainable parameter vector, T is the number of time steps contained in the historical transaction data, tanh is the hyperbolic tangent activation function, is used for the hidden state of the LSTM network at time step t-1 A trainable weight matrix that performs a linear transformation, is the hidden state of the LSTM network at time step t-1, is the feature vector used for the i-th time step in the historical trading data A trainable weight matrix that performs a linear transformation, is the feature vector of the i-th time step in the historical transaction data, is the feature vector used for the jth time step in the historical trading data A trainable weight matrix that performs a linear transformation, is the feature vector of the jth time step in the historical transaction data.

[0012] Optionally, the verification method of the digital collectible verifies the identity association between the current trader and the creation trader based on the operation behavior data of the current trader and the time-sensitive verification token, and obtains the identity association verification result, including: Extract features from the current trader's operation behavior data to obtain the current trader's operation behavior feature vector; Based on the similarity between the operation behavior feature vector of the current trader and the operation behavior feature vector of the genesis trader and the verification result based on the time-sensitive verification token, the identity association value between the current trader and the genesis trader is calculated as the identity association verification result.

[0013] Optionally, the verification method of the digital collection, S4: weighted voting on the multi-chain verification results based on the consensus mechanism to obtain the final verification result, including: Establish the credit rating of each chain in the multi-chain verification results based on historical verification results; Determine the relative contribution of each chain in the multi-chain verification results based on the credit rating of each chain in the multi-chain verification results; Based on the consensus mechanism and the relative contribution of each chain in the multi-chain verification results, a weighted vote is performed on the multi-chain verification results to obtain the final verification result.

[0014] Optionally, a verification system for a verification method of a digital collection includes: A two-way data acquisition module is used to obtain the metadata of the target digital collection from the blockchain, and simultaneously obtain the historical transaction records of the target digital collection and the operation behavior data of the current trader; A key and token generation module, used to generate a dynamic encryption key based on the creation hash and release timestamp in the metadata of the target digital collection and the historical transaction record, and to generate a time-sensitive verification token based on the dynamic encryption key; A multi-chain verification module, which is used to generate multi-chain verification results based on the hash operation results of dynamic encryption keys, the time series characteristics of historical transaction data, the operation behavior data of the current trader, and the time-sensitive verification token; The final verification module is used to perform weighted voting on the multi-chain verification results based on the consensus mechanism to obtain the final verification results.

[0015] Compared with the prior art, the present invention has the following beneficial effects: by obtaining the metadata, historical transaction records and current trader operation behavior data of the target digital collection from the blockchain, rich and comprehensive information is collected, providing sufficient data support for verification. The creation hash in the metadata can ensure the authenticity of the source of the collection, the issuance timestamp clarifies the issuance time sequence, the historical transaction records reflect the circulation of the collection, and the operation behavior data is related to the real-time dynamics of the current transaction. These data complement each other and comprehensively outline the "identity information" and transaction trajectory of the digital collection, laying a solid foundation for subsequent verification. A dynamic encryption key is generated based on the creation hash, issuance timestamp and historical transaction records in the metadata, and a time-sensitive verification token is further generated. The dynamic encryption key combines the key attributes and transaction history of the collection, has dynamic change characteristics, and enhances security; the time-sensitive verification token introduces the limitation of the time dimension, so that the verification is closely related to time, further improving the timeliness and security of the verification, effectively preventing delay attacks or tampering during the verification process, and ensuring the effectiveness of the verification within a specific time range. The hash operation results of the dynamic encryption key, the time series characteristics of historical transaction data, the current trader's operation behavior data, and the time-sensitive verification token are used to generate multi-chain verification results, and digital collections are verified from multiple dimensions. The hash operation results provide verification at the encryption level, the time series characteristics reflect the timing rules of transactions, and the operation behavior data considers the actual situation of current transactions. The multi-chain verification results combine these factors to provide a more comprehensive and three-dimensional verification perspective, greatly improving the accuracy and reliability of verification. Based on the consensus mechanism, the multi-chain verification results are weighted voted to obtain the final verification results, and the consensus mechanism is used to ensure the fairness and transparency of the verification process. The weighted voting method takes into account the importance or credibility differences of different verification results, making the final verification results more authoritative and credible. This method effectively integrates the multi-chain verification results, avoids the one-sidedness of a single verification, enhances the reliability and credibility of the entire digital collection verification process, provides a strong guarantee for the transaction security and authenticity of digital collections, and promotes the healthy and orderly development of the digital collection market.

[0016] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in this application document.

[0017] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] 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: Figure 1 is a flow chart of a verification method for a digital collection in an embodiment of the present invention; Figure 2 Schematic diagram of the functional modules of the verification system for digital collections in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The preferred embodiments of the present invention are described below in conjunction with 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.

[0020] Embodiment 1: The present invention provides a method for verifying digital collections, referring to Figure 1 ,include: S1: Obtain the metadata of the target digital collection from the blockchain, and simultaneously obtain the historical transaction records of the target digital collection and the operation behavior data of the current trader; collect multi-dimensional information for comprehensive verification of digital collections. Obtain the metadata of the target digital collection from the blockchain. The creation hash in the metadata can trace the source of the collection creation to ensure its authenticity; the issuance timestamp clarifies the time sequence of the collection issuance, which is of great significance for judging the development history of the collection. At the same time, obtain the historical transaction records of the digital collection, which can reflect the circulation of the collection at different times, such as the number of transactions, transaction time, transaction object and other information, which is helpful for analyzing the market circulation status of the collection. Also synchronously obtain the operation behavior data of the current trader, such as the transaction initiation time, transaction amount, operation steps, etc., which are related to the real-time dynamics of the current transaction. These data complement each other and comprehensively outline the "identity information" and transaction trajectory of the digital collection, laying a solid foundation for subsequent verification.

[0021] S2: Generate a dynamic encryption key based on the creation hash and release timestamp in the metadata of the target digital collection and the historical transaction record, and generate a time-sensitive verification token based on the dynamic encryption key; generate a dynamic encryption key by combining the creation hash, release timestamp and historical transaction record in the metadata of the target digital collection. The creation hash ensures the unique identification of the source of the collection, the release timestamp reflects the time sequence, and the historical transaction record reflects the circulation process. The dynamic encryption key generated by the combination of the three has the characteristics of dynamic change, which is more secure than the fixed key. For example, different changes in historical transaction records will make the generated dynamic encryption key different. Generate a time-sensitive verification token based on the dynamic encryption key. This token introduces a time dimension restriction, combines the dynamic encryption key with the Merkle tree hash root of the historical transaction record, the Unix timestamp, the current transaction amount, etc., and generates it through a specific algorithm. The time window identifier consists of the first 8 bits of the timestamp, so that the verification is closely related to time, which can effectively prevent delay attacks or tampering during the verification process and ensure the validity of the verification within a specific time range. For example, if the token is used outside the specified time window, the verification will fail.

[0022] S3: Generate multi-chain verification results based on the hash operation results of dynamic encryption keys, the time series characteristics of historical transaction data, the current trader's operation behavior data, and time-sensitive verification tokens; Generate multi-chain verification results based on the hash operation results of dynamic encryption keys, the time series characteristics of historical transaction data, the current trader's operation behavior data, and time-sensitive verification tokens. The hash operation results of dynamic encryption keys can verify the consistency of information related to digital collections at the encryption level; the time series characteristics of historical transaction data can reflect the timing rules of transactions, and by analyzing these characteristics, it can be determined whether the transaction behavior conforms to the normal mode; the current trader's operation behavior data considers the current transaction situation from the perspective of actual transaction operations; and the time-sensitive verification token is restricted from the perspective of time validity. By combining these factors, a more comprehensive and three-dimensional verification perspective is provided, greatly improving the accuracy and reliability of verification.

[0023] S4: Perform weighted voting on the multi-chain verification results based on the consensus mechanism to obtain the final verification results. First, establish the credit rating of each chain in the multi-chain verification results based on the historical verification results. The chain with a high credit rating indicates that its past verification results are more reliable. Then determine the relative contribution of each chain based on the credit rating. When performing weighted voting, the verification results of the chain with a high relative contribution have a greater weight in the final decision. This method uses the consensus mechanism to ensure the fairness and transparency of the verification process. The weighted voting takes into account the importance or credibility differences of different verification results, making the final verification results more authoritative and credible, effectively integrating the multi-chain verification results, avoiding the one-sidedness of a single verification, and enhancing the reliability and credibility of the entire digital collection verification process.

[0024] Embodiment 2: In an embodiment of the present invention, a dynamic encryption key is generated based on the creation hash and release timestamp in the metadata of the target digital collection and the historical transaction record, including: ; In the formula, is a dynamic encryption key, is a shared key generated based on the elliptic curve Diffie-Hellman algorithm, is the creator's public key, Is the release timestamp, is the exclusive OR operator, The hash message authentication code value calculated by the Merkle tree hash root based on historical transaction records and the pre-assigned basic key of the blockchain system and the HMAC-SHA3-256 algorithm. The Merkle tree hash root of historical transaction records. Pre-assigned basic keys for blockchain systems This formula generates an encryption key with dynamically changing characteristics by integrating multiple information closely related to digital collections, thereby enhancing the security and uniqueness of the digital collection verification process.

[0025] The further explanations of the parameters in this embodiment are as follows: Shared key generated based on elliptic curve Diffie-Hellman algorithm: The elliptic curve Diffie-Hellman algorithm is a method used for key exchange in cryptography. The generated shared key provides an initial security foundation for the encryption process. It is based on the mathematical properties of the elliptic curve, ensuring the security and difficulty of the key. For example, in the scenario of digital collections, it can be used as a basic encryption factor to ensure that the generated dynamic encryption key has a certain degree of confidentiality.

[0026] Creator's public key: The creator's public key is closely related to the source of the creation of the digital collection and is an important piece of information that identifies the creator. By incorporating the creator's public key into the formula, the dynamic encryption key is given a unique identifier related to the source of creation, which helps ensure that the authenticity of the creation of the digital collection is reflected at the encryption level. For example, different creators have different public keys, and the generated dynamic encryption keys will also be different, thereby distinguishing digital collections from different creators at the encryption level.

[0027] Release timestamp: The release timestamp records the specific time point when the digital collection is released. It plays the role of identifying the time dimension in the formula. The release time of the digital collection is of great significance to its value, legitimacy, and transaction traceability. Adding the release timestamp when generating the dynamic encryption key allows the key to reflect the time characteristics of the collection's release, increasing the dynamic nature of the key. For example, even if the digital collections are released at different times by the same creator, the generated dynamic encryption keys will be different due to different release timestamps.

[0028] XOR operator: XOR operator is used to perform logical operations on different parameters in the formula. This operation combines different information, further increasing the complexity and unpredictability of key generation. The characteristic of XOR operation is that the same is 0 and different is 1. It can effectively mix the information of each parameter, making the generated dynamic encryption key more secure.

[0029] The hashed message authentication code value calculated based on the Merkle tree hash root of historical transaction records, the basic key pre-assigned by the blockchain system, and the HMAC-SHA3-256 algorithm: The Merkle tree hash root of historical transaction records is a summary of all historical transaction records of digital collections. It can concisely summarize the entire transaction history information and has tamper-proof characteristics. The basic key pre-assigned by the blockchain system is a basic encryption parameter pre-set by the blockchain system, which increases the security of key generation. The HMAC-SHA3-256 algorithm is a hash message authentication code algorithm used to calculate the hashed message authentication code value. The algorithm combines the hash function (SHA3-256) and the key (composed of the Merkle tree hash root and the basic key) to generate a unique authentication code value that is closely related to the historical transaction record. This value fully considers the transaction history information of the digital collection in the process of generating the dynamic encryption key, so that the generated key can reflect the transaction flow characteristics of the collection, further enhancing the security and uniqueness of the key. For example, if the historical transaction records of a digital collection change, the Merkle tree hash root will change, causing the calculated hash message authentication code value to change, and the final generated dynamic encryption key will also be different.

[0030] Merkle tree hash root of historical transaction records: As mentioned earlier, it is a summary of historical transaction records and is directly involved in the calculation in the formula, ensuring that the dynamic encryption key is closely linked to the transaction history of the digital collection, reflecting the impact of transaction history on key generation.

[0031] The basic key pre-assigned by the blockchain system: As an encryption parameter pre-assigned by the system, it provides a fixed security foundation for key generation, and works together with other parameters to increase the complexity and security of the key.

[0032] Embodiment 3: In an embodiment of the present invention, a time-sensitive verification token is generated based on a dynamic encryption key, including: ; In the formula, Token is a time-sensitive verification token. It is a hash message authentication code value calculated based on the dynamic encryption key, the Merkle tree hash root of historical transaction records, the Unix timestamp, the current transaction amount, and the HMAC-SHA3-256 algorithm. is a dynamic encryption key, The Merkle tree hash root of historical transaction records. is the exclusive OR operator, The first 16-bit integer of the Unix timestamp. is the transaction amount of the current transaction, It is the time window identifier, which consists of the first 8 binary bits of the timestamp.

[0033] The formula integrates dynamic encryption keys, information related to historical transaction records, timestamps, and current transaction amounts to generate a verification token that changes over time and is closely associated with the transaction, ensuring the validity of the verification within a specific time frame and preventing delay attacks or tampering.

[0034] And the further explanation of each parameter in this embodiment is as follows: Time-sensitive verification token: This is the final result, which is used for time-sensitive verification in the subsequent verification process of digital collections. It combines multiple information, is time-sensitive and unique, and is the key identifier for determining whether the current verification is valid.

[0035] Hash message authentication code value calculated based on dynamic encryption key, Merkle tree hash root of historical transaction records, Unix timestamp, current transaction amount and HMAC-SHA3-256 algorithm: Using HMAC-SHA3-256 algorithm, the dynamic encryption key, Merkle tree hash root of historical transaction records, Unix timestamp and current transaction amount are integrated and calculated. HMAC-SHA3-256 algorithm can generate a unique hash value, which is closely related to all the input information. For example, as long as any of the information changes, such as the transaction amount changes or the timestamp is different, the generated hash message authentication code value will be different, thus ensuring the uniqueness of the verification token and its sensitivity to information changes.

[0036] Dynamic encryption key: As an important basic information for generating verification tokens, it has integrated many aspects of the creation, issuance and transaction history of digital collections. In the process of generating verification tokens, the dynamic encryption key further ensures the close connection between the verification token and the characteristics of the digital collection itself, increasing the reliability of verification.

[0037] Merkle tree hash root of historical transaction records: The Merkle tree hash root summarizes all historical transaction records of digital collectibles. Incorporating it into the formula makes the verification token linked to the transaction history of the digital collectibles. If the transaction history is tampered with, the Merkle tree hash root will change, which will in turn affect the generation of the verification token, thereby detecting the integrity and authenticity of the transaction history.

[0038] XOR operator: XOR operation is used to logically integrate different information, increasing the complexity and unpredictability of generating verification tokens. The characteristics of XOR operation make different information interact with each other, further strengthening the security of verification tokens.

[0039] The first 16 digits of the Unix timestamp: The Unix timestamp is the number of seconds from 00:00:00 UTC on January 1, 1970 to a specific time point. The first 16 digits of the timestamp are used to introduce a time factor into the verification token. Since time is constantly changing, the values ​​of this part of the verification token generated at different times are different, reflecting the timeliness of the verification token. For example, when verifying a digital collection at different times, the values ​​of this part of the verification token generated will be different over time.

[0040] The current transaction amount: The current transaction amount is included in the generation process of the verification token as an important transaction information. Different transaction amounts will result in the generation of different verification tokens, which makes the verification token not only related to time and historical transactions, but also closely related to the current specific transaction situation, further refining the pertinence of the verification.

[0041] Time window identifier (consisting of the first 8 binary digits of the timestamp): The time window identifier consists of the first 8 binary digits of the timestamp, which defines a specific time range. Within this time range, the verification token is considered valid. For example, if the verification token is used for verification outside the specified time window, the verification will fail because the time window identifier does not match the timestamp when it was generated, thereby effectively preventing delay attacks or tampering, and ensuring the timeliness and security of the verification.

[0042] Embodiment 4: In an embodiment of the present invention, S3: based on the hash operation result of the dynamic encryption key, the time series characteristics of the historical transaction data, the operation behavior data of the current trader and the time-sensitive verification token, a multi-chain verification result is generated, including: The hash operation result of the dynamic encryption key is checked for consistency with the genesis hash value stored in the smart contract to obtain the metadata integrity check result; the dynamic encryption key is hashed, and then the obtained hash operation result is compared with the genesis hash value stored in the smart contract to obtain the metadata integrity check result. The genesis hash value in the smart contract is a unique identifier generated by the hash algorithm in the initial state of the digital collection metadata, representing the original state of the metadata. By comparing the hash operation result of the dynamic encryption key with it, if the two are consistent, it means that the target digital collection metadata obtained from the blockchain has not been tampered with in the subsequent process, ensuring the integrity of the metadata. For example, if someone tries to tamper with the metadata, the dynamic encryption key will change, and its hash operation result will also be different from the genesis hash value, thereby detecting that the metadata has been modified.

[0043] The LSTM network based on the attention mechanism is used to analyze the time series characteristics of historical transaction data, generate transaction behavior feature vectors, and verify transaction behavior based on the transaction behavior feature vectors to obtain transaction behavior verification results; the long short-term memory network (LSTM) based on the attention mechanism is used to analyze the time series characteristics of historical transaction data. The LSTM network is good at processing time series data and can capture long-term dependencies in the data. The attention mechanism allows the network to pay more attention to some information related to the current task when processing sequence data. First, the historical transaction data is converted into a time series format to obtain transaction record information for multiple time steps. The LSTM network processes these transaction record information time step by time, and a hidden state is generated at each time step. The hidden state contains the key features of the transaction data up to the current time step. Next, the attention weight of the LSTM network on each time step in the historical transaction data at the current time step is calculated, and the historical transaction feature vector is weighted based on these attention weights to obtain the context vector. This context vector comprehensively considers the importance of transaction data at different time steps. After that, the output of the LSTM network based on the current transaction data is spliced ​​with the context vector to form a transaction behavior feature vector. Finally, the transaction behavior feature vector is input into the transaction behavior legality classification model. The model will determine whether the transaction behavior is legal based on the learned pattern, thereby obtaining the transaction behavior verification result. For example, if there are abnormal transaction time intervals and transaction amount mutations in the historical transaction data, the LSTM network combined with the attention mechanism can capture these features and identify possible abnormal transaction behaviors through the transaction behavior legality classification model.

[0044] Based on the operation behavior data of the current trader and the time-sensitive verification token, the identity association between the current trader and the genesis trader is verified to obtain the identity association verification result; first, the operation behavior data of the current trader is feature extracted to obtain the operation behavior feature vector of the current trader. This feature vector contains information such as the operation characteristics and habits of the current trader during the transaction process. Then, based on the similarity between the operation behavior feature vector of the current trader and the operation behavior feature vector of the genesis trader, as well as the verification result of the time-sensitive verification token, the identity association value between the current trader and the genesis trader is calculated, and this value is used as the identity association verification result. For example, if the operation behavior feature vector of the current trader has a high similarity with the genesis trader, and the time-sensitive verification token is verified, it means that the current trader and the genesis trader may have a strong identity association, otherwise the association is weak or there may be problems such as identity fraud.

[0045] Among them, the multi-chain verification results include metadata integrity verification results, transaction behavior verification results, and identity association verification results.

[0046] The verification results of the above three aspects, namely metadata integrity verification results, transaction behavior verification results, and identity relevance verification results, are integrated together to form a multi-chain verification result. This multi-dimensional verification method verifies digital collections from different angles, comprehensively evaluates the authenticity and legality of digital collections during the transaction process, and the relevance of the trader's identity, thereby improving the accuracy and reliability of digital collection verification.

[0047] Embodiment 5: In an embodiment of the present invention, the hash operation result of the dynamic encryption key is checked for consistency with the genesis hash value stored in the smart contract to obtain a metadata integrity check result, including: A hash operation is performed on the dynamic encryption key based on a preset encryption function to obtain a hash operation result; and a hash operation is performed on the dynamic encryption key using the preset encryption function. The hash function is unidirectional and unique, that is, given an input (here is the dynamic encryption key), it will generate a hash value of a fixed length (the hash operation result), and it is almost impossible to reversely derive the original dynamic encryption key from the hash value. Different inputs will produce completely different hash values ​​even if there are only slight differences. For example, common hash functions such as SHA-256 will output a 256-bit hash value regardless of the length and content of the input data. In this embodiment, the preset encryption function converts the dynamic encryption key into a specific hash operation result, which represents the characteristic information of the dynamic encryption key.

[0048] Determine whether the hash operation result of the dynamic encryption key is consistent with the genesis hash value stored in the smart contract, and obtain the metadata integrity check result. Compare the result obtained by the hash operation with the genesis hash value stored in the smart contract to determine the integrity of the metadata. A smart contract is an automatically executed contract stored on the blockchain. The genesis hash value is the hash value generated when the digital collection metadata is initially created. It is like a "fingerprint" of the metadata, which uniquely identifies the metadata in the initial state. If the hash operation result of the dynamic encryption key is exactly the same as the genesis hash value stored in the smart contract, it means that the target digital collection metadata obtained from the blockchain has not been tampered with since its creation, because any modification of the metadata may cause the dynamic encryption key to change, and then make the hash operation result different. At this time, the metadata integrity check result is passed. On the contrary, if the two are inconsistent, it means that the metadata may have been tampered with, and the metadata integrity check result is failed. This consistency check is a key step to ensure the authenticity and integrity of the digital collection metadata. It plays a fundamental role in the verification process of the digital collection and provides a reliable data basis for various subsequent analyses and decisions based on metadata.

[0049] Embodiment 6: In an embodiment of the present invention, an LSTM network based on an attention mechanism is used to analyze the time series characteristics of historical transaction data, generate a transaction behavior feature vector, and perform transaction behavior verification based on the transaction behavior feature vector to obtain a transaction behavior verification result, including: Convert historical transaction data into time series format to obtain transaction record information of multiple time steps; organize the collected historical transaction data into time series format, so that the transaction record information can be arranged in chronological order to obtain transaction records of multiple time steps. For example, each time step can represent a day, a week, or a time interval for a transaction, and each time step contains specific transaction information such as transaction amount, transaction parties, transaction time, etc. The time series format helps the LSTM network capture the patterns and trends of transaction behavior over time, because the LSTM network is good at processing data with time dependencies, laying the foundation for subsequent analysis.

[0050] Based on the LSTM network, all transaction record information is processed time by time step to obtain the hidden state of the LSTM network at each time step; the transaction record information of each time step is processed in sequence using the LSTM network. The LSTM network can effectively handle long-term dependency problems through the control of the input gate, forget gate, and output gate. When processing the transaction record of each time step, it combines the information of the previous time step to generate the hidden state of the current time step. The hidden state contains the key features of the transaction data up to the current time step. It retains the important information of the transaction data in the time series and provides a condensed representation containing time dimension features for subsequent analysis. For example, the hidden state may encode information such as the trend of changes in the transaction amount and the fluctuation of the transaction frequency over a period of time.

[0051] Calculate the attention weight of the LSTM network to the i-th time step in the historical trading data at time step t; the attention weight reflects the degree of attention of the current time step to the trading data of each time step in history. This calculation process usually involves complex operations on the hidden state of the LSTM network at the current time step and the feature vectors of each time step in the historical trading data. For example, the attention weight may be obtained by inputting the hidden state and the feature vectors of each time step into a model containing trainable parameters and processing them with an activation function. Through the attention mechanism, the LSTM network can pay more attention to the historical time step information related to the current trading behavior analysis, rather than treating the data of all time steps equally, thereby improving the accuracy of the extraction of trading behavior features. For example, if the trading behavior of the current time step is similar to the trading pattern of a specific time step in the past, the attention mechanism will give a higher weight to that time step.

[0052] Based on the attention weight of the LSTM network at the i-th time step in the historical transaction data at time step t, the historical transaction feature vector is weighted to obtain the context vector; based on the calculated attention weight, the historical transaction feature vector is weighted and summed to obtain the context vector. Specifically, the transaction feature vector of each time step is multiplied by the corresponding attention weight, and then the results of all time steps are added to generate a context vector that integrates the important information in the historical transaction data. The context vector integrates the key information related to the current time step in the historical transaction data. It not only considers the information on the time series, but also highlights the important time steps through the attention weight, which provides an important component for the subsequent generation of feature vectors that fully reflect the transaction behavior.

[0053] The output of the LSTM network based on the current transaction data is spliced ​​with the context vector to obtain the transaction behavior feature vector; the output of the LSTM network on the current transaction data contains the immediate features of the current transaction, while the context vector incorporates the relevant information of historical transactions. The transaction behavior feature vector formed by splicing the two fully reflects the characteristics of the current transaction behavior in the time series, including both the specific circumstances of the current transaction and the association with historical transactions. The transaction behavior feature vector generated in this way provides a rich and representative feature representation for subsequent transaction behavior verification, which can more accurately characterize the characteristics of the transaction behavior and help determine whether the transaction behavior is legal.

[0054] The transaction behavior feature vector is input into the pre-trained transaction behavior legality classification model to obtain the transaction behavior verification result. This classification model has learned a large number of characteristic patterns of legal and illegal transaction behaviors, and can judge the legality of the current transaction behavior based on the input transaction behavior feature vector. The model outputs the transaction behavior verification result, indicating whether the current transaction behavior is legal. For example, if the model determines that the current transaction behavior conforms to normal transaction patterns and rules, the verification result is legal; if the transaction behavior feature vector shows similar characteristics to known illegal transaction behaviors, such as abnormal transaction amount fluctuations, frequent abnormal counterparties, etc., the verification result is illegal. Through this series of steps, the historical transaction behavior of digital collections can be effectively analyzed and verified to ensure the legality of digital collection transactions.

[0055] Embodiment 7: In an embodiment of the present invention, calculating the attention weight of the LSTM network at time step t to the i-th time step in the historical transaction data includes: ; In the formula, is the attention weight of the LSTM network on the i-th time step in the historical transaction data at time step t; exp is an exponential function with the natural constant e as the base (it plays the role of amplifying or reducing the value in the formula. By exponentially processing the result of the exponential part, different input values ​​can be numerically different after the operation, thereby highlighting the difference in importance of different time steps. For example, if the result of the exponential part is large, after the exponential function operation, the corresponding attention weight will be relatively large, indicating that the information of this time step is more important to the current time step); v is a trainable parameter vector (this is the parameter vector that the model continuously adjusts during the training process. Its existence enables the model to automatically learn the relationship between each time step feature and the current time step based on the characteristics of historical trading data, and then determine the appropriate attention weight. In different data sets and task scenarios, the model will learn different weight values ​​to adapt to the judgment of the importance of information at different time steps); T is the number of time steps contained in the historical trading data (used as a normalization factor in the formula. It ensures that the sum of the attention weights over all time steps is 1, making the attention weights of different time steps comparable and balancing the attention paid to each time step overall. For example, if normalization is not performed, the weights of some time steps may be too large or too small, resulting in excessive attention or neglect of information in certain time periods); Tanh is the hyperbolic tangent activation function, which performs nonlinear transformation on the input and maps the input value to the range of -1 to 1. In this formula, it processes the combination of hidden state and feature vector after linear transformation, introduces nonlinear factors, and enables the model to learn more complex patterns. Because the time series relationship in trading behavior is often nonlinear, the hyperbolic tangent function can better capture this complex relationship, thereby more accurately calculating the attention weight; is used for the hidden state of the LSTM network at time step t-1 A trainable weight matrix that performs a linear transformation, is the hidden state of the LSTM network at time step t-1, which contains the key features of the transaction data up to the current time step, reflecting the evolution and accumulation of transaction behavior in the time series. When calculating the attention weight, the hidden state is combined with the feature vector of each time step, which enables the model to dynamically adjust the degree of attention to the information of each historical time step according to the progress of the current transaction behavior; The linear transformation of can adjust the weight of each dimension of information in the hidden state and highlight the important features related to the current time step; is the feature vector used for the i-th time step in the historical trading data A trainable weight matrix that performs a linear transformation, is the feature vector of the i-th time step in the historical transaction data, which contains the specific information of the transaction at that time step, such as the transaction amount, the two parties to the transaction, etc. The linear transformation of can enable the model to assign different weights to different feature dimensions according to its own learned patterns, so as to better combine with the hidden state.

[0056] is the feature vector used for the jth time step in the historical trading data A trainable weight matrix that performs a linear transformation, is the feature vector of the jth time step in the historical transaction data, representing the specific information of the transaction at that time step. The feature vectors of different time steps contain various transaction-related features such as transaction amount, transaction time, and transaction object. These features are the basic information for the model to judge transaction behavior.

[0057] The attention weight in this embodiment reflects the degree of attention of the current time step to the transaction data of each historical time step. Through this weight, the LSTM network can focus more on the historical information related to the current analysis when processing time series data, thereby improving the accuracy of extracting transaction behavior features.

[0058] Embodiment 8: In an embodiment of the present invention, based on the operation behavior data of the current trader and the time-sensitive verification token, the identity association between the current trader and the creation trader is verified to obtain the identity association verification result, including: The feature extraction of the current trader's operation behavior data is performed to obtain the current trader's operation behavior feature vector; these operation behavior data may include information such as the time pattern of transaction initiation, the setting habits of transaction amount, and the selected trading platform function. For example, by analyzing the time when the current trader initiates multiple transactions, the preference characteristics of trading in a specific time period of a day or a week are extracted; from the transaction amount data, the characteristics such as whether the transaction amount is inclined to integer amount transactions and the fluctuation range of the transaction amount are extracted. These features are encoded and combined in a certain way to form a vector, which represents the operation behavior pattern of the current trader. The operation behavior feature vector provides a data basis for the subsequent comparison of the behavior similarity between the current trader and the genesis trader. Each trader's operation behavior has a certain uniqueness. By extracting the feature vector, this uniqueness can be quantified and characterized for comparative analysis with the genesis trader.

[0059] Based on the similarity between the operation behavior feature vector of the current trader and the operation behavior feature vector of the genesis trader and the verification result based on the time-sensitive verification token, the identity association value between the current trader and the genesis trader is calculated as the identity association verification result.

[0060] In this embodiment, the similarity can be calculated in a variety of ways, such as cosine similarity, Euclidean distance, etc. If the values ​​of the operation behavior feature vectors of the two are close in multiple dimensions, then the calculated similarity will be high, indicating that the current trader and the creation trader are similar in operation behavior mode. For example, if the creation trader often trades with similar amounts in a specific time period, and the current trader also shows similar operation characteristics, then the similarity of their operation behavior feature vectors will be high.

[0061] In this embodiment, the time-sensitive verification token is generated in the previous step, and it is closely related to the transaction time, transaction history and other information of the digital collection. The time-sensitive verification token is verified to determine whether it is within the valid time range and whether it matches the transaction-related information. If the verification passes, it means that the current transaction is reliable in terms of consistency of time and related information.

[0062] In this embodiment, the identity association value between the current trader and the original trader is calculated by combining the similarity between the operational behavior feature vectors of the current trader and the original trader and the verification result of the time-sensitive verification token. For example, a calculation rule can be set. When the similarity reaches a certain threshold and the time-sensitive verification token is verified, a higher identity association value is given, indicating that the current trader and the original trader have a stronger identity association; conversely, if the similarity is low or the verification token verification fails, the identity association value is low, indicating that the identity association between the two is weak. This identity association value is the identity association verification result, which comprehensively evaluates the identity connection between the current trader and the original trader from the two aspects of operational behavior and time verification, and provides an important basis for judging the legitimacy of the current transaction and the authenticity of the trader's identity.

[0063] Embodiment 9: In an embodiment of the present invention, S4: weighted voting is performed on the multi-chain verification results based on a consensus mechanism to obtain a final verification result, including: Establish the credit rating of each chain in the multi-chain verification results based on historical verification results; review the historical verification results in the multi-chain verification process, and set the credit rating for each chain in the multi-chain verification results based on these past records. For example, if a chain has given accurate verification results that are consistent with the actual situation in multiple historical verifications, then its credit rating will be relatively high; conversely, if a chain frequently has erroneous or inaccurate results in historical verifications, its credit rating will be low.

[0064] Based on the credit rating of each chain in the multi-chain verification result, the relative contribution of each chain in the multi-chain verification result is determined; based on the credit rating of each chain, their relative contribution in the multi-chain verification result is further determined. The chain with a high credit rating will also have a higher relative contribution, which means that in the process of forming the final verification result, the verification result of the chain should be given a greater weight. For example, the credit rating can be converted into a relative contribution through a certain mathematical mapping relationship. For example, the credit rating is set to be proportional to the relative contribution. The relative contribution of a chain with a credit rating of A is twice that of a chain with a credit rating of B. The relative contribution clarifies the difference in the importance of each chain in the verification process, so that when the multi-chain verification results are integrated, the weight can be reasonably allocated according to the reliability and past performance of each chain, avoiding the irrationality that may be caused by simple averaging, thereby integrating multi-chain verification information more scientifically.

[0065] Based on the consensus mechanism and the relative contribution of each chain in the multi-chain verification results, weighted voting is performed on the multi-chain verification results to obtain the final verification results. Based on the consensus mechanism and the relative contribution of each chain, weighted voting is performed on the multi-chain verification results. The consensus mechanism ensures the fairness and consistency of the entire voting process. It stipulates how to reach a consensus among multiple participants. In this process, the verification results of each chain are assigned corresponding weights according to the relative contribution of each chain, and then a comprehensive calculation is performed. For example, suppose there are three chains, the relative contribution of chain A is 0.5, chain B is 0.3, and chain C is 0.2. Their verification results are "passed", "failed", and "passed" respectively. When weighted voting is performed, the total weight of "passed" is 0.5+0.2=0.7, and the total weight of "failed" is 0.3. According to the pre-set rules (such as weight exceeding 0.5 is judged as passed), the final verification result of "passed" is finally obtained.

[0066] This weighted voting method effectively integrates the results of multi-chain verification and avoids the one-sidedness of a single verification. By considering the credit rating and relative contribution of each chain, the final verification result is more authoritative and credible, and can more accurately reflect the true status of digital collections, providing more reliable protection for the transaction security and authenticity of digital collections.

[0067] In this embodiment, the credit rating is a quantitative assessment of the reliability of each chain's verification, which reflects the chain's performance in previous verification tasks. By establishing a credit rating, it provides an important reference for the subsequent determination of the weight of each chain in the final verification result, which helps to distinguish the verification quality and credibility of different chains.

[0068] Embodiment 10: In an embodiment of the present invention, a verification system for a verification method of a digital collection is provided, referring to Figure 2 ,include: A two-way data acquisition module is used to obtain the metadata of the target digital collection from the blockchain, and simultaneously obtain the historical transaction records of the target digital collection and the operation behavior data of the current trader; A key and token generation module, used to generate a dynamic encryption key based on the creation hash and release timestamp in the metadata of the target digital collection and the historical transaction record, and to generate a time-sensitive verification token based on the dynamic encryption key; A multi-chain verification module is used to generate multi-chain verification results based on the hash operation results of dynamic encryption keys, the time series characteristics of historical transaction data, the operation behavior data of the current trader, and the time-sensitive verification token; The final verification module is used to perform weighted voting on the multi-chain verification results based on the consensus mechanism to obtain the final verification results.

[0069] The digital collection verification system collects metadata, historical transaction records and current trader behavior data on the blockchain in real time through a two-way data acquisition module, and uses the key and token generation module to generate dynamic encryption keys and time-sensitive verification tokens based on the creation hash, issuance timestamp and historical transactions. The multi-chain verification module combines key hash operations, historical transaction time series analysis, behavioral feature comparison and token verification to generate multi-chain verification results. Finally, the final verification module uses a weighted voting consensus mechanism to integrate multi-chain results to form an authoritative verification conclusion, thereby realizing full-process verification of the authenticity, transaction compliance and identity relevance of digital collections.

[0070] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. A method for verifying digital collections, characterized in that: include: S1: Obtain the metadata of the target digital collection from the blockchain, and simultaneously obtain the historical transaction records of the target digital collection and the operation behavior data of the current trader; S2: Generate a dynamic encryption key based on the creation hash and release timestamp in the metadata of the target digital collection and the historical transaction record, and generate a time-sensitive verification token based on the dynamic encryption key; S3: Generates multi-chain verification results based on the hash operation results of the dynamic encryption key, the time series characteristics of historical transaction data, the operation behavior data of the current trader, and the time-sensitive verification token; S4: Perform weighted voting on the multi-chain verification results based on the consensus mechanism to obtain the final verification result.

2. The method for verifying digital collections according to claim 1, characterized in that: Generate dynamic encryption keys based on the creation hash and release timestamp in the metadata of the target digital collection and historical transaction records, including: ; In the formula, is a dynamic encryption key, is a shared key generated based on the elliptic curve Diffie-Hellman algorithm, is the creator's public key, Is the release timestamp, is the exclusive OR operator, The hash message authentication code value calculated by the Merkle tree hash root based on historical transaction records and the pre-assigned basic key of the blockchain system and the HMAC-SHA3-256 algorithm. The Merkle tree hash root of historical transaction records. The base key pre-assigned for the blockchain system.

3. The digital collection verification method according to claim 1, characterized in that: Generates a time-sensitive verification token based on a dynamic encryption key, including: ; In the formula, Token is a time-sensitive verification token. It is a hash message authentication code value calculated based on the dynamic encryption key, the Merkle tree hash root of historical transaction records, the Unix timestamp, the current transaction amount, and the HMAC-SHA3-256 algorithm. is a dynamic encryption key, The Merkle tree hash root of historical transaction records. is the exclusive OR operator, The first 16-bit integer of the Unix timestamp. is the transaction amount of the current transaction, It is the time window identifier, which consists of the first 8 binary bits of the timestamp.

4. The method for verifying digital collections according to claim 1, characterized in that: S3: Generates multi-chain verification results based on the hash operation results of the dynamic encryption key, the time series characteristics of historical transaction data, the operation behavior data of the current trader, and the time-sensitive verification token, including: Perform a consistency check between the hash operation result of the dynamic encryption key and the genesis hash value stored in the smart contract to obtain the metadata integrity check result; The LSTM network based on the attention mechanism is used to analyze the time series characteristics of historical transaction data, generate transaction behavior feature vectors, and verify the transaction behavior based on the transaction behavior feature vectors to obtain the transaction behavior verification results; Based on the current trader's operation behavior data and time-sensitive verification token, verify the identity association between the current trader and the genesis trader, and obtain the identity association verification result; Among them, the multi-chain verification results include metadata integrity verification results, transaction behavior verification results, and identity association verification results.

5. The method for verifying digital collections according to claim 4, characterized in that: The hash operation result of the dynamic encryption key is checked for consistency with the genesis hash value stored in the smart contract to obtain the metadata integrity check result, including: Performing a hash operation on the dynamic encryption key based on a preset encryption function to obtain a hash operation result; Determine whether the hash operation result of the dynamic encryption key is consistent with the genesis hash value stored in the smart contract, and obtain the metadata integrity verification result.

6. The verification method of digital collection according to claim 4, characterized in that: The LSTM network based on the attention mechanism is used to analyze the time series characteristics of historical transaction data, generate transaction behavior feature vectors, and perform transaction behavior verification based on the transaction behavior feature vectors to obtain transaction behavior verification results, including: Convert historical transaction data into time series format to obtain transaction record information for multiple time steps; Based on the LSTM network, all transaction record information is processed time-by-time to obtain the hidden state of the LSTM network at each time step; Calculate the attention weight of the LSTM network at time step t to the i-th time step in the historical transaction data; The historical transaction feature vector is weighted based on the attention weight of the LSTM network on the i-th time step in the historical transaction data at time step t to obtain the context vector; The output of the LSTM network based on the current transaction data is concatenated with the context vector to obtain the transaction behavior feature vector; The transaction behavior feature vector is input into the transaction behavior legitimacy classification model to obtain the transaction behavior verification result.

7. The method for verifying digital collectibles according to claim 6, characterized in that: Calculate the attention weight of the LSTM network at time step t for the i-th time step in the historical transaction data, including: ; In the formula, is the attention weight of the LSTM network to the i-th time step in the historical transaction data at time step t, exp is an exponential function with the natural constant e as the base, v is a trainable parameter vector, T is the number of time steps contained in the historical transaction data, tanh is the hyperbolic tangent activation function, is used for the hidden state of the LSTM network at time step t-1 A trainable weight matrix that performs a linear transformation, is the hidden state of the LSTM network at time step t-1, is the feature vector used for the i-th time step in the historical trading data A trainable weight matrix that performs a linear transformation, is the feature vector of the i-th time step in the historical transaction data, is the feature vector used for the jth time step in the historical trading data A trainable weight matrix that performs a linear transformation, is the feature vector of the jth time step in the historical transaction data.

8. The method for verifying digital collections according to claim 4, characterized in that: Based on the current trader's operation behavior data and time-sensitive verification token, verify the identity association between the current trader and the genesis trader, and obtain the identity association verification results, including: Extract features from the current trader's operation behavior data to obtain the current trader's operation behavior feature vector; Based on the similarity between the operation behavior feature vector of the current trader and the operation behavior feature vector of the genesis trader and the verification result based on the time-sensitive verification token, the identity association value between the current trader and the genesis trader is calculated as the identity association verification result.

9. The method for verifying digital collections according to claim 1, characterized in that: S4: Perform weighted voting on the multi-chain verification results based on the consensus mechanism to obtain the final verification results, including: Establish the credit rating of each chain in the multi-chain verification results based on historical verification results; Determine the relative contribution of each chain in the multi-chain verification results based on the credit rating of each chain in the multi-chain verification results; Based on the consensus mechanism and the relative contribution of each chain in the multi-chain verification results, a weighted vote is performed on the multi-chain verification results to obtain the final verification result.

10. The verification system of the digital collection verification method according to any one of claims 1 to 9, characterized in that: include: A two-way data acquisition module is used to obtain the metadata of the target digital collection from the blockchain, and simultaneously obtain the historical transaction records of the target digital collection and the operation behavior data of the current trader; A key and token generation module, used to generate a dynamic encryption key based on the creation hash and release timestamp in the metadata of the target digital collection and the historical transaction record, and to generate a time-sensitive verification token based on the dynamic encryption key; A multi-chain verification module is used to generate multi-chain verification results based on the hash operation results of dynamic encryption keys, the time series characteristics of historical transaction data, the operation behavior data of the current trader, and the time-sensitive verification token; The final verification module is used to perform weighted voting on the multi-chain verification results based on the consensus mechanism to obtain the final verification results.

Images