CN115952240B - Financial data compliance examination method and device based on blockchain - Google Patents
Financial data compliance examination method and device based on blockchain Download PDFInfo
- Publication number
- CN115952240B CN115952240B CN202310248208.4A CN202310248208A CN115952240B CN 115952240 B CN115952240 B CN 115952240B CN 202310248208 A CN202310248208 A CN 202310248208A CN 115952240 B CN115952240 B CN 115952240B
- Authority
- CN
- China
- Prior art keywords
- financial
- financial data
- compliance
- data
- department
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 27
- 230000008859 change Effects 0.000 claims abstract description 125
- 238000012550 audit Methods 0.000 claims abstract description 66
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 51
- 238000012549 training Methods 0.000 claims abstract description 31
- 230000007246 mechanism Effects 0.000 claims abstract description 9
- 239000011159 matrix material Substances 0.000 claims description 37
- 238000012552 review Methods 0.000 claims description 29
- 238000011176 pooling Methods 0.000 claims description 22
- 238000010276 construction Methods 0.000 claims description 21
- 230000000977 initiatory effect Effects 0.000 claims description 16
- 238000013528 artificial neural network Methods 0.000 claims description 9
- 230000004913 activation Effects 0.000 claims description 8
- 238000012795 verification Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 7
- 238000011478 gradient descent method Methods 0.000 claims description 6
- 230000009467 reduction Effects 0.000 claims description 6
- 230000007480 spreading Effects 0.000 claims description 5
- 230000004044 response Effects 0.000 claims description 2
- 230000000875 corresponding effect Effects 0.000 description 11
- 238000007689 inspection Methods 0.000 description 9
- 238000011835 investigation Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000013474 audit trail Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000007792 addition Methods 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000001902 propagating effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of blockchain and compliance internal control, and discloses a blockchain-based financial data compliance examination method and device, wherein the method comprises the following steps: constructing a multi-department financial block chain, and uploading a financial change data block based on a consensus mechanism; acquiring a financial data index and a physical address of a financial change data block, and constructing a query index table; the method comprises the steps of arranging all nodes in a blockchain, and acquiring target financial data from all financial data change blocks by using a query index table; generating a financial data audit association cube model according to the target financial data; constructing a three-dimensional time sequence convolutional neural network, and training on a preset training set to obtain a financial data compliance examination model; and after developing the financial data audit-related cube model, inputting a financial data compliance examination model to obtain a compliance examination result of the target financial data. The invention realizes the automatic, high-efficiency and high-accuracy compliance examination of dynamic financial data in different time periods.
Description
Technical Field
The invention relates to the technical field of blockchain and internal control compliance, in particular to a blockchain-based financial data compliance examination method and device.
Background
The financial compliance is one of important links of enterprise internal control compliance, and the financial compliance has very important significance for managing and controlling enterprise operation risks. The traditional financial data compliance checking method generally performs compliance checking on annual financial data, and due to the large data volume of the annual financial data, the compliance checking time is long easily, and the compliance checking result is unified, so that the traditional method based on static financial data is difficult to effectively monitor and check the dynamic change process of the financial data in the face of frequent and detailed changes of the financial data and wrong or unreliable changes of the financial data.
To ensure the validity of the financial data, blockchain techniques may be employed to ensure that the financial details are not tamperable and not repudiatable. However, the current general blockchain technology does not support the blockquery index, and needs to search the blockdata in a full traversal mode, so that it is difficult to realize efficient on-chain query for frequently-changed financial detail data.
Disclosure of Invention
Based on the above, the invention aims to solve at least one technical problem in the background art, thereby providing a blockchain-based financial data compliance review method and device.
To solve the above problems, an embodiment of the present invention provides a blockchain-based financial data compliance review method, including:
constructing a multi-department financial block chain, and uploading a financial change data block based on a consensus mechanism;
acquiring a financial data index and a physical address of the financial change data block to construct a query index table;
deploying full nodes in the multi-department financial blockchain, and acquiring target financial data from all the financial data change blocks by utilizing the query index table;
generating a financial data audit association cube model according to the target financial data;
constructing a three-dimensional time sequence convolutional neural network, and training the three-dimensional time sequence convolutional neural network on a preset training set to obtain a financial data compliance examination model;
and after the financial data audit associated cube model is unfolded, inputting the financial data compliance examination model to obtain a compliance examination result of the target financial data.
Optionally, the constructing a three-dimensional time sequence convolutional neural network, and training the three-dimensional time sequence convolutional neural network on a preset training set to obtain a financial data compliance examination model includes:
constructing a three-dimensional time sequence convolution neural network comprising an input layer, a three-dimensional convolution layer, a pooling layer, a full-connection layer and an output layer;
acquiring a colluded-auditing associated cube model and a compliance examination real label of a financial data sample, and constructing a training set;
expanding the audit association cube model of the financial data sample to obtain a plurality of audit association matrixes in different time periods;
inputting a plurality of the audit association matrixes into the three-dimensional time sequence convolutional neural network, extracting independent characteristics of each audit association matrix and correlation characteristics between the audit association matrixes through the three-dimensional time sequence convolutional neural network, and obtaining an audit compliance probability value of a financial data sample according to the independent characteristics and the correlation characteristics;
obtaining a loss function according to the examination compliance probability value and the compliance examination real tag;
acquiring the gradient between the full-connection layer and the output layer in the three-dimensional time sequence convolutional neural network according to the loss function;
And (3) reversely spreading the gradient through a gradient descent method, and optimizing weight parameters between the full-connection layer and the output layer to obtain the financial data compliance examination model.
Optionally, the three-dimensional convolution layer includes a first convolution kernel of size 3*3 and a second convolution kernel of size 1×10, the number of the first convolution kernel and the second convolution kernel are both 10, and the pooling layer includes a maximum pooling kernel of size 2×2;
the extracting, by the three-dimensional time-sequence convolutional neural network, an independent feature of each of the audit association matrices and a correlation feature between the audit association matrices, and obtaining an audit compliance probability value of a financial data sample according to the independent feature and the correlation feature, including:
acquiring a plurality of correlation matrixes through an input layer of the three-dimensional time sequence convolutional neural network;
performing convolution operation on each of the correlation matrixes through a first convolution check of the three-dimensional convolution layer, extracting independent features, and performing convolution operation on values of the same data element in different time periods in the correlation matrixes through a second convolution check of the three-dimensional convolution layer, so as to extract correlation features;
performing dimension reduction processing on the independent features and the correlation features through the maximum pooling check of the pooling layer;
And performing network parameter fitting through the ReLU activation function and the Softmax activation function of the full connection layer, and outputting a checking compliance probability value through the output layer.
Optionally, the loss function adopts a mean square error loss function, specifically:
wherein ,
for the mean square error loss function->
Censoring compliance probability values for three-dimensional time sequence convolutional neural network output>
Inspecting the real label for compliance;
the gradient between the full connection layer and the output layer is as follows:
wherein ,
for step size->
Is input for the full connection layer.
Optionally, the uplink of the financial change data block based on the consensus mechanism includes:
each department node in the multi-department financial block chain detects whether the department financial data is changed;
when the financial data of a department is changed, generating a financial change data block containing a time stamp, a hash abstract, a digital signature, financial data change content and a financial data index by a department node which generates the financial data change, and generating a uplink request containing the financial change data block;
and the multi-department financial block chain receives and responds to the uplink request, performs consensus on the financial change data block, and newly adds and stores the financial change data block after the consensus is successful.
Optionally, the multi-department financial blockchain performs consensus on the financial change data block in response to the uplink request, and adds and stores the financial change data block after the consensus is completed, including:
respectively confirming a department node and other department nodes with financial data change in the multi-department financial block chain as a change initiating node and a consensus node;
each consensus node obtains a uplink request containing the financial change data block, performs identity verification on the change initiating node according to a digital signature in the financial change data block, and votes the change initiating node when the identity verification passes;
and the multi-department financial block chain acquires a voting result, and when the voting result meets a preset consensus success condition, the financial change data block is newly added to each consensus node.
Optionally, the deploying full nodes in the multi-department financial blockchain and obtaining target financial data from all the financial data change blocks using the query index table includes:
asymmetric encryption is carried out on the financial data index of the financial change data block on the multi-department financial block chain by adopting a public key, and all nodes with private keys are distributed in the multi-department financial block chain;
And when the full node receives a compliance examination request containing index information, acquiring a financial data index matched with the index information from the query index table, and accessing the physical address of the financial change data block according to the matched financial data index to acquire target financial data.
Optionally, the generating the financial data audit associated cube model according to the target financial data includes:
determining row and column indexes of the auditing association matrix according to field items in the target financial data;
acquiring a correlation coefficient between every two field items corresponding to the row and column indexes; the correlation coefficient includes a linear correlation coefficient and a pearson correlation coefficient;
filling the correlation coefficient into a row-column index corresponding position of the auditing association matrix to complete construction of the auditing association matrix;
and obtaining the audit trail association matrix of a plurality of time slices, and combining to obtain the financial data audit trail association cube model.
In addition, the embodiment of the invention also provides a financial data compliance examining device based on the blockchain, which comprises the following steps:
the block chain construction module is used for constructing a multi-department financial block chain and carrying out uplink on the financial change data block based on a consensus mechanism;
The index table construction module is used for acquiring the financial data index and the physical address of the financial change data block to construct a query index table;
the financial data query module is used for distributing all nodes in the multi-department financial block chain and acquiring target financial data from all financial data change blocks by utilizing the query index table;
the association model generation building module is used for generating a financial data audit association cube model according to the target financial data;
the compliance examination model construction module is used for constructing a three-dimensional time sequence convolutional neural network and training the three-dimensional time sequence convolutional neural network on a preset training set to obtain a financial data compliance examination model;
and the compliance examination module is used for expanding the financial data checking associated cube model and inputting the financial data compliance examination model to obtain a compliance examination result of the target financial data.
Optionally, the compliance censoring model building module includes:
the neural network construction submodule is used for constructing a three-dimensional time sequence convolutional neural network comprising an input layer, a three-dimensional convolutional layer, a pooling layer, a full-connection layer and an output layer;
The training set construction sub-module is used for acquiring a colluded auditing associated cube model and a compliance auditing real label of the financial data sample and constructing a training set;
the correlation model processing sub-module is used for expanding the correlation cube model of the financial data sample to obtain a plurality of correlation matrixes of different time periods;
the neural network training sub-module is used for inputting a plurality of the audit-taking incidence matrixes into the three-dimensional time sequence convolutional neural network, extracting independent characteristics of each audit-taking incidence matrix and correlation characteristics between the audit-taking incidence matrixes through the three-dimensional time sequence convolutional neural network, and obtaining an audit compliance probability value of a financial data sample according to the independent characteristics and the correlation characteristics; obtaining a loss function according to the examination compliance probability value and the compliance examination real tag; acquiring the gradient between the full-connection layer and the output layer in the three-dimensional time sequence convolutional neural network according to the loss function; and (3) reversely spreading the gradient through a gradient descent method, and optimizing weight parameters between the full-connection layer and the output layer to obtain the financial data compliance examination model.
The technical scheme provided by the embodiment of the invention has the following beneficial effects:
1) The financial data change is dynamically recorded through the multi-department financial block chain, so that the financial data cannot be tampered and counterfeited;
2) The financial change data blocks of all department nodes on the chain are accessed through the query index table so as to obtain target financial data, and the query and tracing efficiency of the financial data on the chain can be improved;
3) By generating the financial data audit association cube model, the correlation among field items in the financial data can be characterized, so that the non-compliance financial data mode can be conveniently mined;
4) And the three-dimensional time sequence convolutional neural network structure is adopted to carry out compliance examination by adopting the financial data compliance examination model, so that the automatic, high-efficiency and high-accuracy compliance examination of dynamic financial data in different time periods can be realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the present invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 illustrates a flow chart of a blockchain-based financial data compliance review method provided by an embodiment of the present invention;
FIG. 2 is a flowchart of step S50 in a blockchain-based financial data compliance review method provided by an embodiment of the present invention;
FIG. 3 is a schematic diagram of a block chain based financial data compliance review device according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a compliance review model building module in a blockchain-based financial data compliance review device according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the invention. Rather, they are merely examples of apparatus and methods consistent with aspects of the invention as detailed in the accompanying claims.
As shown in fig. 1, a flowchart of a blockchain-based financial data compliance review method according to an embodiment of the present invention includes the following steps:
S10, constructing a multi-department financial block chain, and uploading the financial change data block based on a consensus mechanism.
In step S10, a multi-department financial blockchain is built based on the federated chain, the multi-department financial blockchain including a plurality of department nodes, which may be represented as
，/>
A set of department nodes representing a multi-department financial blockchain,
representing a set of connected edges of the multi-department financial blockchain. Checking financial data recorded in ledger for each department node in multi-department financial blockchain +.>
Whether or not a change operation such as addition, modification, deletion occurs +.>
If a change operation occurs, namely +.>
The initiating department node of the financial data change generates a financial change data block->
Applying for uplink, and carrying out consensus with other department nodes on the link based on the Raft protocol
Multi-department financial blockchain is newly added and financial change data block is stored after consensus is successful
。
Preferably, in the step S10, the financial change data block is uplink based on a consensus mechanism, and specifically includes the following steps:
s101, detecting whether the department financial data is changed or not by each department node in the multi-department financial block chain.
S102, when the financial data of the department is changed, a department node which generates the financial change data block comprising a time stamp, a hash abstract, a digital signature, financial data change content and a financial data index generates a uplink request comprising the financial change data block.
In this embodiment, for any department node in the multi-department financial blockchain, if a change in financial data is detected, a time stamp is generated to be included
Hash digest->
Digital signature
Financial data change content->
Financial data index->
Financial change data block->
And generating a data block containing financial changes +.>
Is sent into the multi-department financial blockchain. Wherein, timestamp->
For recording financial data change times; hash abstract->
To use SHA-1 (Secure Hash Algorithm 1 ) to hash the entire financial change data block, the hashed signature is used to ensure that the financial change data block has not been tampered with; digital signature
Department private key is adopted for unique identification of department by adopting RSA algorithm (public key encryption algorithm)
Encrypted ciphertext identification, which can be used by other department nodes in the chain with the public key +.>
After decryption, verifying the identity authenticity of the department node generating the financial change data block (namely, the department node generating the financial data change); financial data change content->
For recording the entire content of the financial data change; financial data index->
Altering content for financial data >
Is a key word of (a).
And S103, the multi-department financial block chain receives and responds to the uplink request, performs consensus on the financial change data block, and newly adds and stores the financial change data block after the consensus is successful.
In this embodiment, the step S103 may include the steps of:
s1031, confirming the department nodes with financial data change and other department nodes in the multi-department financial block chain as change initiating nodes and consensus nodes respectively.
S1032, each consensus node obtains a uplink request containing the financial change data block, performs identity verification on the change initiating node according to the digital signature in the financial change data block, and votes on the change initiating node when the identity verification passes.
S1033, the multi-department financial block chain acquires a voting result, and when the voting result meets a preset consensus success condition, the financial change data block is newly added to each consensus node.
In step S1031, the department node where the financial data change occurs is a department node that generates a financial change data block or a department node that generates a uplink request.
In step S1032, after receiving the uplink request of the department node (i.e., the change initiating node) with the financial data change, the other department nodes (i.e., the consensus node) on the chain analyze the uplink request to obtain the financial change data block, then analyze the financial change data block to obtain the digital signature, decrypt the digital signature by using the public common key to obtain the department unique identifier of the change initiating node, and finally compare the department unique identifier of the change initiating node with the identifier information recorded in advance by the consensus node to verify the identity authenticity of the change initiating node, if the comparison result is the same, then determine that the identity verification passes, vote for the change initiating node, otherwise determine that the identity verification fails, and vote for the change initiating node.
In step S1033, the multi-department financial block chain counts the number of votes to obtain a voting result, and when it is detected that the number of votes exceeds half the number of nodes, it is determined that the voting result satisfies a consensus success condition, and the financial change data block is newly added to each consensus node. It can be appreciated that the multi-department financial blockchain constructed in this embodiment can dynamically record financial data changes, and ensure that the financial data cannot be tampered and forged.
S20, acquiring the financial data index and the physical address of the financial change data block to construct a query index table.
In step S20, first, a financial data index is parsed from the financial change data block, where the financial data index may be set according to keywords of financial data change content, and the keywords of financial data change content may be obtained by an existing keyword extraction tool. Then, the financial data index is correlated with the physical address of the financial change data block, and a lookup index table is constructed
Stored in a database of multi-department financial blockchains. It can be appreciated that the problem of low on-chain financial data query efficiency caused by complex and numerous details of the financial data on the blockchain can be solved by constructing the query index table.
S30, arranging all nodes in the multi-department financial block chain, and acquiring target financial data from all financial data change blocks by using the query index table.
In step S30, since the blockchain contains financial data for multiple departments, a problem is addressed in that the financial data may not have associated access rights, and thus, the financial data for the blockchain financial change data block is indexed 
Use of public key->
RSA encryption (i.e. asymmetric encryption) is performed>
And is deployed in the blockchain with the private key +.>
Is provided with a private key +.>
The full node of (a) may decrypt the financial data index of the financial change data block and thereby access the rights of all financial change data blocks on the chain. When compliance reviews are performed, the full node may query and normalize all of the financial change data blocks using the query index table.
Preferably, the step S30 includes the steps of:
s301, performing asymmetric encryption on the financial data index of the financial change data block on the multi-department financial blockchain by adopting a public key, and deploying a full node with a private key in the multi-department financial blockchain.
S302, when the full node receives a compliance review request containing index information, acquiring a financial data index matched with the index information from the query index table, and accessing the physical address of the financial change data block according to the matched financial data index to acquire target financial data.
In this embodiment, the multi-department financial blockchain includes a plurality of department nodes that store financial data and a full node that queries and reviews the financial data. After receiving a compliance inspection request containing index information, the full node analyzes the compliance inspection request to obtain index information, wherein the index information can be a keyword or an index name of a financial data index; and acquiring matched financial data indexes from the query index table according to the index information, marking the indexes as target financial data indexes, accessing financial change data blocks of corresponding department nodes on the chain by using the target financial data indexes, and analyzing financial data change contents from the financial change data blocks to serve as target financial data.
It will be appreciated that a full node of the multi-department financial blockchain may access the financial change data blocks of any of the department nodes on the chain to obtain at least one set of target financial data based on compliance review requests.
S40, generating a financial data audit association cube model according to the target financial data.
In step S40, based on the target financial numbers obtained from the nodes of each department on the chain, which are not tamperable and forgery-proofAccording to the data, generating a financial data audit association cube model, wherein the financial data audit association cube model is formed by audit association matrix
Time series of>
Composition, which can be expressed as
, wherein />
Indicate->
Checking the correlation matrix of each time segment, and +.>
. For the audit-based incidence matrix, the row index and the column index of the audit-based incidence matrix both represent field items in the target financial data in a data table format, and the numerical value of the corresponding position of the row and the column represents the correlation coefficient between every two field items, wherein the correlation coefficient comprises linear correlation and pearson correlation.
Preferably, the step S40 specifically includes the following steps:
s401, determining row and column indexes of the auditing association matrix according to field items in the target financial data.
S402, obtaining a correlation coefficient between every two field items corresponding to the row and column index.
S403, filling the correlation coefficient into a row-column index corresponding position of the auditing association matrix to complete construction of the auditing association matrix.
S404, acquiring the audit association matrix of a plurality of time slices, and combining according to the time sequence to obtain the financial data audit association cube model.
In step S401, the field items include, but are not limited to, payouts and avails of different enterprise operations, and the like.
In step S404, the time slot is set according to the change frequency of the financial data, and the higher the change frequency of the financial data is, the shorter the time slot is.
Further, the correlation coefficient includes a linear correlation coefficient and a pearson correlation coefficient, and at this time, the step S302 specifically includes the steps of:
s3021, calculating a linear correlation coefficient according to the numerical values of each row and each column of field items, where the linear correlation coefficient is calculated by:
wherein ,
is->
Numerical value of line field item,/->
Is->
Column field item value,/->
、/>
Is a constant value, and is used for the treatment of the skin,
is a linear correlation coefficient.
S3022, calculating a pearson correlation coefficient according to the numerical values of each row and each column of field items and the numerical average value of each row and each column of field items, where the pearson correlation coefficient is calculated by:
wherein ,
、/>
respectively +.>
Line and->
Column field item value,/->
、/>
Respectively +.>
Line and->
The value of the column field entry is +.>
An average value in the above.
And S3023, determining the linear correlation coefficient and the Pearson correlation coefficient as the correlation coefficient between the two field items.
Specifically, in the course of constructing the correlation matrix, firstly, according to the target financial data
The individual field items determine the +.>
Individual row index and->
Individual column index and initialize +.>
Checking the incidence matrix, and enabling the numerical value of each element in the checking incidence matrix to be a preset initial value; then obtaining the correlation coefficient between the field item corresponding to the row index and the field item corresponding to the column index through the calculation formula of the linear correlation coefficient and the calculation formula of the pearson correlation coefficient, and replacing the preset initial values of the corresponding elements of the row index and the column index in the initialized auditing correlation matrix with the correlation coefficient; and finally, after the numerical values of all elements in the initialized auditing association matrix are replaced, obtaining the final auditing association matrix.
It can be appreciated that when multiple sets of target financial data are obtained according to the compliance review request, a corresponding audit association matrix can be constructed for each set of target financial data and associated with the department unique identifier.
S50, constructing a three-dimensional time sequence convolutional neural network, and training the three-dimensional time sequence convolutional neural network on a preset training set to obtain a financial data compliance examination model.
In step S50, the preset training set includes the audit-related cube model of the financial data sample and the compliance audit-realistic tag of the financial data sample. The financial data sample is historical financial data recorded in an account book by each department node on the chain. The process of generating the associated cube model by the audit of the financial data sample can refer to the process of generating the associated cube model by the audit of the financial data sample, and specifically includes the steps S401 to S404. And obtaining the compliance examination real label of the financial data sample by a manual marking mode.
It should be noted that, the construction of the financial data compliance censoring model is only required before the compliance censoring is performed by applying the financial data compliance censoring model, that is, the step S50 may be performed before any of the steps S10 to S40, and fig. 1 is merely a flowchart of an alternative embodiment.
Preferably, as shown in fig. 2, the step S50 specifically includes the following steps:
s501, constructing a three-dimensional time sequence convolution neural network comprising an input layer, a three-dimensional convolution layer, a pooling layer, a full connection layer and an output layer.
S502, acquiring a colluded audit associated cube model and a compliance audit real label of a financial data sample, and constructing a training set.
And S503, expanding the audit association cube model of the financial data sample to obtain a plurality of audit association matrixes of different time periods.
S504, inputting a plurality of the audit association matrixes into the three-dimensional time sequence convolutional neural network, extracting independent characteristics of each audit association matrix and correlation characteristics between the audit association matrixes through the three-dimensional time sequence convolutional neural network, and obtaining an audit compliance probability value of a financial data sample according to the independent characteristics and the correlation characteristics.
S505, examining a real label according to the examination compliance probability value and the compliance, and obtaining a loss function; the loss function adopts a mean square error loss function, and can be expressed as:
wherein ,
for the mean square error loss function->
Censoring compliance probability values for three-dimensional time sequence convolutional neural network output>
The authentic tag is inspected for compliance.
S506, acquiring gradients between a full-connection layer and an output layer in the three-dimensional time sequence convolutional neural network according to the loss function; the gradient between the fully connected layer and the output layer can be expressed as:
wherein ,
for step size->
Is input for the full connection layer.
S507, reversely propagating the gradient through a gradient descent method, and optimizing weight parameters between the full-connection layer and the output layer to obtain the financial data compliance examination model.
It should be noted that, the construction of the three-dimensional time-series convolutional neural network and the construction of the training set may be performed simultaneously, or a certain step may be performed preferentially, that is, the step S501 and the step S501 may be performed simultaneously, or any step may be performed preferentially over another step, and fig. 2 is a flowchart of an alternative embodiment only.
Further, the three-dimensional convolution layer includes a first convolution kernel of size 3*3 and a second convolution kernel of size 1×10, the number of the first convolution kernel and the second convolution kernel is 10, and the pooling layer includes a maximum pooling kernel of size 2×2, where the step S504 includes the steps of:
s5031, obtaining a plurality of correlation matrixes through an input layer of the three-dimensional time sequence convolutional neural network.
S5032, performing convolution operation on each of the audit-related matrixes through a first convolution check of the three-dimensional convolution layer, extracting independent features, and performing convolution operation on values of the same data element in different time periods in the audit-related matrixes through a second convolution check of the three-dimensional convolution layer, so as to extract correlation features.
S5033, performing dimension reduction processing on the independent features and the correlation features through the maximum pooling check of the pooling layer.
S5034, performing network parameter fitting by using the ReLU activation function and Softmax activation function of the full connection layer, and outputting the inspection compliance probability value by using the output layer.
In this embodiment, the financial data compliance inspection model adopting the 3D time sequence convolutional neural network structure includes an input layer, a 3D convolutional layer, a pooling layer, a full connection layer and an output layer, wherein the input layer is used for obtaining a plurality of audit association matrices developed by the audit association cube model; the 3D convolution layer (namely a three-dimensional convolution layer) comprises two types of convolution kernels, wherein one type of convolution kernel is 10 convolution kernels with 3*3 sizes which are randomly generated and is used for carrying out convolution operation on the auditing association matrix in different time periods to extract independent features, and the other type of convolution kernel is 10 convolution kernels with 1 x 10 sizes which are randomly generated and is used for carrying out convolution operation on the values of the same data element in the auditing association matrix in different time periods to extract correlation features; the pooling layer adopts the largest pooling core with the size of 2 x 2 to carry out dimension reduction treatment for reserving the significance characteristics; the full-connection layer comprises a ReLU activation function and a Softmax activation function and is used for processing the feature information after dimension reduction; the output layer is used for outputting the examination compliance probability value. It can be appreciated that, the embodiment adopts the financial data compliance inspection model of the 3D time sequence convolutional neural network structure, can extract independent characteristics of the auditing association matrix in different time periods and correlation characteristics of the auditing association matrix in different time periods, automatically excavates non-compliance financial data modes, further judges that the financial data has the possibility of non-compliance, and realizes automatic, high-efficiency and high-accuracy compliance inspection of dynamic financial data in different time periods.
S60, after the financial data inspection association cube model is unfolded, inputting the financial data compliance inspection model to obtain a compliance inspection result of the target financial data.
Specifically, when the financial data compliance examination model is utilized to carry out compliance examination, a plurality of investigation correlation matrixes obtained by unfolding a financial data investigation correlation cube model are obtained through an input layer of the financial data compliance examination model, then independent features of each investigation correlation matrix and correlation features among the investigation correlation matrixes are extracted through a 3D convolution layer, and then after the independent features and the correlation features are subjected to dimension reduction through a pooling layer, a compliance probability value is output through a full connection layer and an output layer, so that a compliance examination result of target financial data is obtained.
In summary, the blockchain-based financial data compliance review method provided by the embodiment has the following beneficial effects:
1) The financial data change is dynamically recorded through the multi-department financial block chain, so that the financial data cannot be tampered and counterfeited;
2) The financial change data blocks of all department nodes on the chain are accessed through the query index table so as to obtain target financial data, and the query and tracing efficiency of the financial data on the chain can be improved;
3) By generating the financial data audit association cube model, the correlation among field items in the financial data can be characterized, so that the non-compliance financial data mode can be conveniently mined;
4) And the three-dimensional time sequence convolutional neural network structure is adopted to carry out compliance examination by adopting the financial data compliance examination model, so that the automatic, high-efficiency and high-accuracy compliance examination of dynamic financial data in different time periods can be realized.
In addition, as shown in fig. 3, the embodiment of the invention further provides a financial data compliance checking device based on a blockchain, which comprises:
the blockchain construction module 110 is configured to construct a multi-department financial blockchain and uplink a financial change data block based on a consensus mechanism;
an index table construction module 120, configured to obtain a financial data index and a physical address of the financial change data block, so as to construct a query index table;
a financial data query module 130 for deploying full nodes in the multi-department financial blockchain and obtaining target financial data from all of the financial data change blocks using the query index table;
the association model generation module 140 is configured to generate a financial data audit association cube model according to the target financial data;
The compliance review model construction module 150 is configured to construct a three-dimensional time-series convolutional neural network, and train the three-dimensional time-series convolutional neural network on a preset training set to obtain a financial data compliance review model;
and the compliance review module 160 is configured to input the compliance review model of the financial data after developing the associated cube model of the financial data, so as to obtain a compliance review result of the target financial data.
In some alternative embodiments, as shown in FIG. 4, the compliance review model building module 150 includes:
a neural network construction sub-module 151, configured to construct a three-dimensional time-sequential convolutional neural network including an input layer, a three-dimensional convolutional layer, a pooling layer, a full-connection layer, and an output layer;
a training set constructing sub-module 152 for acquiring the audit-taking associated cube model and the compliance audit real label of the financial data sample and constructing a training set;
the correlation model processing sub-module 153 is configured to expand the audit association cube model of the financial data sample to obtain a plurality of audit association matrices in different time periods;
the neural network training sub-module 154 is configured to input a plurality of the audit-taking correlation matrices into the three-dimensional time-sequence convolutional neural network, extract, through the three-dimensional time-sequence convolutional neural network, an independent feature of each of the audit-taking correlation matrices and a correlation feature between the audit-taking correlation matrices, and obtain an audit compliance probability value of a financial data sample according to the independent feature and the correlation feature; obtaining a loss function according to the examination compliance probability value and the compliance examination real tag; acquiring the gradient between the full-connection layer and the output layer in the three-dimensional time sequence convolutional neural network according to the loss function; and (3) reversely spreading the gradient through a gradient descent method, and optimizing weight parameters between the full-connection layer and the output layer to obtain the financial data compliance examination model.
It can be appreciated that the blockchain-based financial data compliance checking device provided in this embodiment is used to implement a blockchain-based financial data compliance checking method in the above embodiment, and has the beneficial effects of the corresponding method embodiments, which are not described herein.
It should be noted that in the description of the present invention, the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. Furthermore, in the description of the present invention, unless otherwise indicated, the meaning of "plurality" means at least two.
The present invention is not limited to the above embodiments, but is capable of modification and variation in detail, and other modifications and variations can be made by those skilled in the art without departing from the scope of the present invention.
Claims (8)
1. A blockchain-based financial data compliance review method, comprising:
constructing a multi-department financial block chain and uplink a financial change data block based on a consensus mechanism, comprising: each department node in the multi-department financial block chain detects whether the department financial data is changed;
when the financial data of a department is changed, generating a financial change data block containing a time stamp, a hash abstract, a digital signature, financial data change content and a financial data index by a department node which generates the financial data change, and generating a uplink request containing the financial change data block; the multi-department financial block chain receives and responds to the uplink request, carries out consensus on the financial change data block, and newly adds and stores the financial change data block after the consensus is successful;
acquiring a financial data index and a physical address of the financial change data block to construct a query index table;
deploying full nodes in the multi-department financial blockchain, and acquiring target financial data from all the financial data change blocks by utilizing the query index table;
generating a financial data audit associated cube model from the target financial data, comprising: determining row and column indexes of the auditing association matrix according to field items in the target financial data; acquiring a correlation coefficient between every two field items corresponding to the row-column index, wherein the correlation coefficient comprises a linear correlation coefficient and a pearson correlation coefficient; filling the correlation coefficient into a row-column index corresponding position of the auditing association matrix to complete construction of the auditing association matrix; acquiring the audit association matrixes of a plurality of time slices, and combining to obtain a financial data audit association cube model;
Constructing a three-dimensional time sequence convolutional neural network, and training the three-dimensional time sequence convolutional neural network on a preset training set to obtain a financial data compliance examination model;
and after the financial data audit association cube model is unfolded, inputting the financial data compliance examination model, extracting independent characteristics of each audit association matrix and correlation characteristics between the audit association matrices through the financial data compliance examination model, and obtaining a compliance examination result of the target financial data according to the independent characteristics and the correlation characteristics.
2. The blockchain-based financial data compliance review method of claim 1, wherein the constructing a three-dimensional time-series convolutional neural network and training the three-dimensional time-series convolutional neural network on a preset training set to obtain a financial data compliance review model comprises:
constructing a three-dimensional time sequence convolution neural network comprising an input layer, a three-dimensional convolution layer, a pooling layer, a full-connection layer and an output layer;
acquiring a colluded-auditing associated cube model and a compliance examination real label of a financial data sample, and constructing a training set;
expanding the audit association cube model of the financial data sample to obtain a plurality of audit association matrixes in different time periods;
Inputting a plurality of the audit association matrixes into the three-dimensional time sequence convolutional neural network, extracting independent characteristics of each audit association matrix and correlation characteristics between the audit association matrixes through the three-dimensional time sequence convolutional neural network, and obtaining an audit compliance probability value of a financial data sample according to the independent characteristics and the correlation characteristics;
obtaining a loss function according to the examination compliance probability value and the compliance examination real tag;
acquiring the gradient between the full-connection layer and the output layer in the three-dimensional time sequence convolutional neural network according to the loss function;
and (3) reversely spreading the gradient through a gradient descent method, and optimizing weight parameters between the full-connection layer and the output layer to obtain the financial data compliance examination model.
3. The blockchain-based financial data compliance review method of claim 2, wherein the three-dimensional convolution layer includes a first convolution kernel of size 3*3 and a second convolution kernel of size 1 x 10, the number of first and second convolution kernels each being 10, the pooling layer including a maximum pooling kernel of size 2 x 2;
the extracting, by the three-dimensional time-sequence convolutional neural network, an independent feature of each of the audit association matrices and a correlation feature between the audit association matrices, and obtaining an audit compliance probability value of a financial data sample according to the independent feature and the correlation feature, including:
Acquiring a plurality of correlation matrixes through an input layer of the three-dimensional time sequence convolutional neural network;
performing convolution operation on each of the correlation matrixes through a first convolution check of the three-dimensional convolution layer, extracting independent features, and performing convolution operation on values of the same data element in different time periods in the correlation matrixes through a second convolution check of the three-dimensional convolution layer, so as to extract correlation features;
performing dimension reduction processing on the independent features and the correlation features through the maximum pooling check of the pooling layer;
and performing network parameter fitting through the ReLU activation function and the Softmax activation function of the full connection layer, and outputting a checking compliance probability value through the output layer.
4. The blockchain-based financial data compliance review method of claim 2, wherein the loss function employs a mean square error loss function, specifically:
wherein ,
for the mean square error loss function->
Censoring compliance probability values for three-dimensional time sequence convolutional neural network output>
Inspecting the real label for compliance;
the gradient between the full connection layer and the output layer is as follows:
wherein ,
for step size->
Is input for the full connection layer.
5. The blockchain-based financial data compliance review method of claim 1, wherein the multi-department financial blockchain performs a consensus on the financial change data block in response to the uplink request, and newly adds and stores the financial change data block after the consensus is completed, comprising:
respectively confirming a department node and other department nodes with financial data change in the multi-department financial block chain as a change initiating node and a consensus node;
each consensus node obtains a uplink request containing the financial change data block, performs identity verification on the change initiating node according to a digital signature in the financial change data block, and votes the change initiating node when the identity verification passes;
and the multi-department financial block chain acquires a voting result, and when the voting result meets a preset consensus success condition, the financial change data block is newly added to each consensus node.
6. The blockchain-based financial data compliance review method of claim 1, wherein the deploying full nodes in the multi-department financial blockchain and using the lookup index table to obtain target financial data from all of the financial data change blocks comprises:
Asymmetric encryption is carried out on the financial data index of the financial change data block on the multi-department financial block chain by adopting a public key, and all nodes with private keys are distributed in the multi-department financial block chain;
and when the full node receives a compliance examination request containing index information, acquiring a financial data index matched with the index information from the query index table, and accessing the physical address of the financial change data block according to the matched financial data index to acquire target financial data.
7. A blockchain-based financial data compliance censoring device, comprising:
the block chain construction module is used for constructing a multi-department financial block chain and carrying out uplink on a financial change data block based on a consensus mechanism, and comprises the following steps: each department node in the multi-department financial block chain detects whether the department financial data is changed; when the financial data of a department is changed, generating a financial change data block containing a time stamp, a hash abstract, a digital signature, financial data change content and a financial data index by a department node which generates the financial data change, and generating a uplink request containing the financial change data block; the multi-department financial block chain receives and responds to the uplink request, carries out consensus on the financial change data block, and newly adds and stores the financial change data block after the consensus is successful;
The index table construction module is used for acquiring the financial data index and the physical address of the financial change data block to construct a query index table;
the financial data query module is used for distributing all nodes in the multi-department financial block chain and acquiring target financial data from all financial data change blocks by utilizing the query index table;
the association model generation module is used for generating a financial data audit association cube model according to the target financial data, and comprises the following steps: determining row and column indexes of the auditing association matrix according to field items in the target financial data; acquiring a correlation coefficient between every two field items corresponding to the row-column index, wherein the correlation coefficient comprises a linear correlation coefficient and a pearson correlation coefficient; filling the correlation coefficient into a row-column index corresponding position of the auditing association matrix to complete construction of the auditing association matrix; acquiring the audit association matrixes of a plurality of time slices, and combining to obtain a financial data audit association cube model;
the compliance examination model construction module is used for constructing a three-dimensional time sequence convolutional neural network and training the three-dimensional time sequence convolutional neural network on a preset training set to obtain a financial data compliance examination model;
And the compliance examination module is used for expanding the financial data checking association cube model, inputting the financial data compliance examination model, extracting the independent feature of each checking association matrix and the correlation feature between the checking association matrices through the financial data compliance examination model, and obtaining the compliance examination result of the target financial data according to the independent feature and the correlation feature.
8. The blockchain-based financial data compliance review device of claim 7, wherein the compliance review model building module comprises:
the neural network construction submodule is used for constructing a three-dimensional time sequence convolutional neural network comprising an input layer, a three-dimensional convolutional layer, a pooling layer, a full-connection layer and an output layer;
the training set construction sub-module is used for acquiring a colluded auditing associated cube model and a compliance auditing real label of the financial data sample and constructing a training set;
the correlation model processing sub-module is used for expanding the correlation cube model of the financial data sample to obtain a plurality of correlation matrixes of different time periods;
the neural network training sub-module is used for inputting a plurality of the audit-taking incidence matrixes into the three-dimensional time sequence convolutional neural network, extracting independent characteristics of each audit-taking incidence matrix and correlation characteristics between the audit-taking incidence matrixes through the three-dimensional time sequence convolutional neural network, and obtaining an audit compliance probability value of a financial data sample according to the independent characteristics and the correlation characteristics; obtaining a loss function according to the examination compliance probability value and the compliance examination real tag; acquiring the gradient between the full-connection layer and the output layer in the three-dimensional time sequence convolutional neural network according to the loss function; and (3) reversely spreading the gradient through a gradient descent method, and optimizing weight parameters between the full-connection layer and the output layer to obtain the financial data compliance examination model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310248208.4A CN115952240B (en) | 2023-03-15 | 2023-03-15 | Financial data compliance examination method and device based on blockchain |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310248208.4A CN115952240B (en) | 2023-03-15 | 2023-03-15 | Financial data compliance examination method and device based on blockchain |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115952240A CN115952240A (en) | 2023-04-11 |
| CN115952240B true CN115952240B (en) | 2023-06-02 |
Family
ID=85891444
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310248208.4A Active CN115952240B (en) | 2023-03-15 | 2023-03-15 | Financial data compliance examination method and device based on blockchain |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115952240B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115357623A (en) * | 2022-08-31 | 2022-11-18 | 普联软件股份有限公司 | Intelligent organization method, system and medium for multidimensional cube data |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109086325A (en) * | 2018-06-29 | 2018-12-25 | 阿里巴巴集团控股有限公司 | Data processing method and device based on block chain |
| CN108959228A (en) * | 2018-07-13 | 2018-12-07 | 众安信息技术服务有限公司 | Based on block chain creation, retrieval, the method and readable storage medium storing program for executing for editing data |
| CN109241192B (en) * | 2018-09-18 | 2021-06-15 | 百度在线网络技术（北京）有限公司 | Data modification and block verification method, device, equipment and medium for block chain |
| CN113255496A (en) * | 2021-05-17 | 2021-08-13 | 杭州西语科技有限公司 | Financial expense reimbursement management method based on block chain technology |
-
2023
- 2023-03-15 CN CN202310248208.4A patent/CN115952240B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115357623A (en) * | 2022-08-31 | 2022-11-18 | 普联软件股份有限公司 | Intelligent organization method, system and medium for multidimensional cube data |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115952240A (en) | 2023-04-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shen et al. | Privacy-preserving image retrieval for medical IoT systems: A blockchain-based approach | |
| JP5989183B2 (en) | Data signature generation method, data signature verification method, and management computer | |
| US20160283920A1 (en) | Authentication and verification of digital data utilizing blockchain technology | |
| JP5415795B2 (en) | Method and apparatus for verifying the integrity of an edited document | |
| US11321305B2 (en) | Utilizing independently stored validation keys to enable auditing of instrument measurement data maintained in a blockchain | |
| CN112491551A (en) | Data verification method and device based on block chain and electronic equipment | |
| US20210109917A1 (en) | System and Method for Processing a Database Query | |
| CN104798052B (en) | The matched method and apparatus of consumer | |
| US11361110B2 (en) | File verification method, file verification system and file verification server | |
| KR20120044002A (en) | Method for analysis and validation of online data for digital forensics and system using the same | |
| CN102045356A (en) | Cloud-storage-oriented trusted storage verification method and system | |
| CN111563243A (en) | Credible identity authentication platform based on WeChat applet | |
| CN104090835A (en) | eID (electronic IDentity) and spectrum theory based cross-platform virtual asset transaction audit method | |
| CN109636553B (en) | Credential management method, apparatus, computer device and storage medium | |
| CN107451459A (en) | The method and apparatus verified using picture validation code | |
| CN107171808A (en) | A kind of verification method and device of electronic record authenticity | |
| CN111339551A (en) | Data verification method and related device and equipment | |
| CN110753080A (en) | Block transmission method, device, equipment and readable storage medium | |
| CN110610094A (en) | College data increment treatment system based on block chain | |
| CN115952240B (en) | Financial data compliance examination method and device based on blockchain | |
| CN112947911A (en) | Interface script generation method, device, equipment and storage medium | |
| US9465858B2 (en) | Systems and methods for authenticating and aiding in indexing of and searching for electronic files | |
| CN112348041B (en) | Log classification and log classification training method and device, equipment and storage medium | |
| CN106960052B (en) | Credit investigation data acquisition method and system | |
| Eisenberg et al. | Building an Electronic Records Archive at the National Archives and Records Administration: Recommendations for a Long-Term Strategy |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |