CN113535848A - Block chain-based credit investigation grade determination method, device, equipment and storage medium - Google Patents
Block chain-based credit investigation grade determination method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN113535848A CN113535848A CN202110699625.1A CN202110699625A CN113535848A CN 113535848 A CN113535848 A CN 113535848A CN 202110699625 A CN202110699625 A CN 202110699625A CN 113535848 A CN113535848 A CN 113535848A
- Authority
- CN
- China
- Prior art keywords
- matrix
- credit investigation
- data
- credit
- block chain
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/27—Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/64—Protecting data integrity, e.g. using checksums, certificates or signatures
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Security & Cryptography (AREA)
- Software Systems (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Computing Systems (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Bioethics (AREA)
- Health & Medical Sciences (AREA)
- Algebra (AREA)
- Computer Hardware Design (AREA)
- General Health & Medical Sciences (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention provides a block chain-based credit investigation grade determination method, a block chain-based credit investigation grade determination device and a block chain-based storage medium, wherein the method comprises the steps of acquiring credit investigation data of a target enterprise in a block chain system, carrying out credit investigation on a credit investigation data packet, and converting the credit investigation index into a test matrix; multiplying the test matrix by a predetermined target weight matrix to obtain a target score matrix; the target score matrix is normalized, the category of a row of corresponding credit grades with the highest score is determined to be used as the credit investigation grade of a target enterprise, so that the safety and the accuracy of data transmission of credit investigation indexes are guaranteed due to the characteristic that a block chain system cannot be tampered, meanwhile, due to the target weight matrix and a calculation model of the normalization process, the process of determining the credit investigation grade by the credit investigation indexes can be automatically completed, the grade determination process is free of manual participation, and the accuracy and the reliability of a credit investigation grade determination result are effectively improved.
Description
Technical Field
The present invention relates to the field of block chain and credit investigation data processing technologies, and in particular, to a block chain-based credit investigation level determination method, apparatus, device, and storage medium.
Background
The block chain has the technical characteristics of decentralization, distributed trust, timestamp, asymmetric encryption and the like, can ensure the safety and controllability of credit data sharing and effectively ensure the data privacy, and has good application scenes in the fields of data transaction, data sharing and the like of credit investigation industries, such as banks, insurance and the like. The open, shared, transparent, reciprocal ideas of the blockchain are highly consistent with the value sought by the credit investigation industry all the time, and have been successfully applied to the fields of finance, digital currency, supply chain and the like. The blockchain technology is considered by many scholars to be capable of solving the problems of data trust, privacy protection, data right confirmation and the like in credit investigation information sharing by virtue of the characteristics of distrust, no tampering, data source tracing and the like. Meanwhile, as the current credit investigation service scene is increasingly complex and the credit investigation data amount is increasingly increased, various theoretical systems, such as a Bayesian method and a logistic regression algorithm designed by the comprehensive statistical theory, have been developed gradually by the credit investigation evaluation method; designing a bp neural network and an SVM model by using a machine learning theory; the Credit Metrics model and the KVM model, which perform Credit evaluation based on market value, are used to evaluate Credit.
In general industries, credit investigation data are stored in a block chain network, so that the credit investigation data cannot be tampered randomly, accuracy and safety of the credit investigation data are further guaranteed, and then a credit investigation data demand party acquires the credit investigation data through the block chain network.
However, after the data demand party acquires the credit investigation data, the credit investigation grade evaluation is mostly performed on the related enterprises corresponding to the credit investigation data in a manual mode, so that the accuracy and reliability of the evaluation result are relatively low.
Disclosure of Invention
The invention provides a block chain-based credit investigation grade determination method, a block chain-based credit investigation grade determination device, block chain-based credit investigation grade determination equipment and a block chain-based memory medium, which are used for overcoming the defects of low accuracy and reliability of manual credit investigation grade assessment results in the prior art and realizing automatic credit investigation grade assessment.
The invention provides a credit investigation grade determining method based on a block chain, which comprises the following steps:
acquiring credit investigation data of a target enterprise in a block chain system, wherein the credit investigation data comprises credit investigation indexes;
converting the credit investigation index into a test matrix;
multiplying the test matrix by a predetermined target weight matrix to obtain a target score matrix;
and carrying out normalization processing on the target score matrix, and determining the category of the credit level corresponding to a column with the highest score as the credit investigation level of the target enterprise.
According to the credit investigation grade determination method based on the block chain provided by the invention, before acquiring credit investigation data of a target enterprise in the block chain system, the method further comprises the following steps:
acquiring any training data set;
converting the training data set into a sample data matrix, and marking the credit level corresponding to each row of data in the sample data matrix as a marking probability array;
obtaining a sample probability matrix based on the sample data matrix according to the initial weight matrix and the Softmax function;
determining a loss function according to the sample probability matrix and the mark probability array, and obtaining a loss matrix according to the loss function;
judging whether the loss matrix is converged;
if not, updating the initial weight matrix based on a gradient descent algorithm and a preset learning rate;
and if the convergence is achieved, determining the initial weight matrix corresponding to the convergence as a target weight matrix.
According to the credit investigation level determination method based on the block chain, the obtaining of the sample probability matrix based on the sample data matrix according to the initial weight matrix and the Softmax function includes:
multiplying the sample data matrix by an initial weight matrix to obtain a sample score matrix;
and carrying out normalization processing on each row of data in the sample score matrix by using a Softmax function to obtain a sample probability matrix.
According to the credit investigation grade determining method based on the block chain, the loss function comprises a first loss function and/or a second loss function;
the first loss function is a balanced cross entropy loss function;
the second loss function is a combined function that balances the combination of the cross-entropy loss function and the center loss function.
According to the credit investigation grade determination method based on the block chain provided by the invention, before acquiring credit investigation data of a target enterprise in the block chain system, the method further comprises the following steps:
setting different roles for each node in the block chain system;
and configuring different operation authorities for each different role so as to realize the initialization of the block chain system.
According to the credit investigation grade determination method based on the block chain provided by the invention, before acquiring credit investigation data of a target enterprise in the block chain system, the method further comprises the following steps:
and storing credit investigation data of any enterprise through the initialized blockchain system, wherein the credit investigation data of any enterprise is uploaded to the initialized blockchain system by a data provider according to a preset format.
According to the credit investigation grade determining method based on the block chain provided by the invention, the credit investigation data of the target enterprise is acquired in the block chain system, and the method comprises the following steps:
acquiring a demand instruction of a data demander, wherein the demand instruction carries an enterprise identifier and a credit investigation index;
and inquiring credit investigation data of a target enterprise in the block chain system according to the enterprise identification.
The invention also provides a credit investigation grade determining device based on the block chain, which comprises the following components:
the acquisition module is used for acquiring credit investigation data of a target enterprise in a block chain system, wherein the credit investigation data comprises credit investigation indexes;
the conversion module is used for converting the credit investigation index into a test matrix;
the determining module is used for multiplying the test matrix and a predetermined target weight matrix to obtain a target score matrix; and carrying out normalization processing on the target score matrix, and determining the category of the credit level corresponding to a column with the highest score as the credit investigation level of the target enterprise.
The invention also provides an electronic device, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the program to realize the steps of the block chain-based credit rating determination method.
The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when being executed by a processor, carries out the steps of the block chain based credit rating determination method as set forth in any of the above.
According to the block chain-based credit investigation grade determination method, device, equipment and storage medium, the credit investigation data of a target enterprise are acquired in a block chain system, the credit investigation indexes are collected by a credit investigation data packet, and the credit investigation indexes are converted into a test matrix; multiplying the test matrix by a predetermined target weight matrix to obtain a target score matrix; the target score matrix is normalized, the category of a row of corresponding credit grades with the highest score is determined to be used as the credit investigation grade of a target enterprise, so that the safety and the accuracy of data transmission of credit investigation indexes are guaranteed due to the characteristic that a block chain system cannot be tampered, meanwhile, due to the target weight matrix and a calculation model of the normalization process, the process of determining the credit investigation grade by the credit investigation indexes can be automatically completed, the grade determination process is free of manual participation, and the accuracy and the reliability of a credit investigation grade determination result are effectively improved.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed for the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a block chain-based credit rating determination method according to the present invention;
FIG. 2 is a schematic flow chart of the determination of the target weight matrix in FIG. 1 according to the present invention;
fig. 3 is a second schematic flowchart of the block chain-based credit assessment level determination method according to the present invention;
fig. 4 is a schematic structural diagram of a block chain-based credit rating determination apparatus according to the present invention.
Fig. 5 is a schematic structural diagram of an electronic device provided in the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The block chain-based credit investigation level determination method, apparatus, device and storage medium of the present invention are described below with reference to fig. 1 to 5.
Fig. 1 is a schematic flow chart of a block chain-based credit rating determination method according to the present invention.
As shown in fig. 1, one of the flow diagrams of the method for determining a credit rating based on a block chain according to this embodiment includes the following steps:
101. and acquiring credit investigation data of the target enterprise in the block chain system, wherein the credit investigation data comprises credit investigation indexes.
Specifically, the data demand side sends a credit investigation data acquisition instruction carrying an enterprise identifier and a credit investigation index according to the actual demand of the data demand side, queries a target enterprise through the enterprise identifier after receiving the credit investigation data acquisition instruction, and then acquires the required credit investigation index from the credit investigation data corresponding to the target enterprise.
When credit grading is needed to be carried out on credit investigation customers according to credit investigation indexes, the data demander node takes out data subjected to credit investigation grading from the block chain world state database according to actual requirements. When the node calls the data to obtain the intelligent contract, the credit investigation data in the world state database is generally obtained through key query, and the block chain can also adopt a CouchDB complex state database to meet the query requirements of rich query, non-key query and the like.
102. And converting the credit investigation index into a test matrix.
Converting the credit investigation index into a Test matrix, wherein the Test matrix is a matrix with 1 row and M columns and is a Test1×MWherein M is a natural number and represents the number of credit investigation indexes. For example, 3 credit investigation indexes, the test matrix is 1 row and 3 columns, 24 credit investigation indexes, and the test matrix is 1 row and 24 columns.
103. And multiplying the test matrix by a predetermined target weight matrix to obtain a target score matrix.
The target Weight matrix is a predetermined matrix, for example, the target Weight matrix is a matrix of M rows and K columns, denoted as WeightM×KAnd multiplying the test matrix by a predetermined target weight matrix to obtain a target score matrix with 1 row and K columns, wherein K represents the credit level. The score of each credit indicator at the credit rating is then obtained. For example, the test matrix has 1 row and 3 columns, the credit rating is A, B, C, D, the target weight matrix is a 3 row and 4 column matrix, wherein the first column has a credit rating of a, the second column has a credit rating of B, the third column has a credit rating of C, and the fourth column has a credit rating of D.
104. And carrying out normalization processing on the target score matrix, and determining the category of the credit grade corresponding to a column with the highest score as the credit investigation grade of the target enterprise.
The target score matrix can be subjected to normalization processing through a Softmax regression algorithm, and the function is to map data to a range from 0 to 1 for processing, so that the data processing is more convenient and faster. And simultaneously, determining a column with the highest score, wherein the category of the credit grade corresponding to the column with the highest score is the credit evaluation grade of the target enterprise. According to the above example, the target score matrix is a matrix with 1 row and 4 columns, for example, the normalized matrix is [0.6, 0.7, 0.8, 0.9], and then the credit rating category corresponding to 0.9 is the credit rating of the target enterprise, i.e. D level.
Fig. 2 is a schematic flow chart of the determination of the target weight matrix in fig. 1 according to the present invention.
As shown in fig. 2, the process schematic diagram for determining the target weight matrix in fig. 1 provided in this embodiment includes the following steps, wherein the process of determining the target weight matrix in steps 201 and 207 is performed before the embodiment of 101 and 104.
201. Any training data set is acquired.
For example, the obtained training data set has N pieces of data, each piece of data has M indexes, and the credit rating has K classes, where N, M, K are all natural numbers.
202. And converting the training data set into a sample data matrix, and marking the credit level corresponding to each row of data in the sample data matrix as a marking probability array.
After N pieces of Data are obtained, the obtained Data are converted into a sample Data matrix, wherein the format of the sample Data matrix is a matrix with N rows and M columns, and the sample Data matrix is marked as DataN×M. Meanwhile, marking the credit type corresponding to each row of data as LabelN,LabeliE {1,2,3, … …, K }, it is noted that labeled is probability, i.e., the probability value that each row of data is of a certain class.
203. And obtaining a sample probability matrix based on the sample data matrix according to the initial weight matrix and the Softmax function.
Specifically, the initial Weight values of the initial Weight matrix are all 1, and the initial Weight matrix can be marked as WeightM×K. Multiplying the sample Data matrix by the initial weight matrix to obtain a sample score matrix, namely the sample Data matrix Data with N rows and M columnsN×MAnd M rows and K columns of initial Weight matrix WeightM×KMultiplying to obtain N rows and K columnsThe sample score Matrix of (2) is denoted as MatrixN×KAnd the data of 1 to K columns in the sample score matrix respectively represent the scores of the credit data under the current weight in K levels.
Normalizing each row of data in the sample scoring matrix by using a Softmax function to obtain a sample probability matrix which is recorded as SoftmaxN×KThat is, the score corresponding to each row of data is converted into a probability, and the probability matrix in which each data is a probability is obtained by converting the score corresponding to each row of data in the manner of formula 1.
204. And determining a loss function according to the sample probability matrix and the marked probability array, and obtaining a loss matrix according to the loss function.
The calculation probability matrix is obtained through the steps, the probability matrix is obtained through calculation according to the preset calculation rule, and before the training data are obtained, probability marking is carried out on each row of data. At this time, the pre-marked mark probability array is compared with the sample probability matrix to determine the Loss function, and then the Loss matrix with N rows and K columns is generated according to the Loss function and is marked as LossN×K。
Specifically, the loss function includes a first loss function and/or a second loss function; the first loss function is a balanced cross entropy loss function; the second loss function is a combined function that balances the combination of the cross-entropy loss function and the center loss function. Equation 2 is a process of calculating each element in the loss matrix according to the balanced cross entropy loss function, and equation 3 is a process of calculating each element in the loss matrix according to a combined function of the balanced cross entropy loss function and the central loss function.
When j is LabeliWhen y isj1, otherwise yj＝0；LabeliThe true value of the sample, if Labeli＝2，y21, other (y)1,y3…yk) Is 0, y ═ 0,1,0, … …,0]I.e. y is the one-hot matrix of Label.
For example, assuming that K is 3, i.e. there are 3 categories, the sample score Matrix is calculated to be Matrixi＝[2,3,4]
Wherein, Labeli2, i.e. the sample itself is of class 2, then y is 0,1,0]
SoftmaxiRegression gave [0.0903, 0.2447, 0.665 [)]
Computing Loss matrix Lossi＝[0，ln(0.2447)，0]
DataiIs a feature of the ith sample,
c is to maintain a class center for each class, each class center having m features, β being a constant.
Another way of calculating the loss matrix can be determined according to the calculation of equation 3. The loss matrix calculated according to the formula 2 adopts an inter-class competition mechanism, so that the inter-class separability can be well considered, and the requirement on hardware is low. The loss matrix calculated by adopting the mode of the formula 3 is set for each feature, the data of the same type are continuously converged in the class in the training process, the classification accuracy is further enhanced, and the calculation result is more accurate.
205. And judging whether the loss matrix is converged.
206. And if not, updating the initial weight matrix based on a gradient descent algorithm and a preset learning rate.
And according to the loss function obtained by calculation and the thought of gradient reduction, obtaining the gradient of the loss function in the parameter direction, and further updating the initial weight matrix according to a preset learning rate, wherein alpha is the learning rate.
Equation 4 is an updated initial weight matrix calculation formula,
for the gradient descent algorithm, the gradient of the loss matrix in each weight direction, w, is calculatedkIs weightkFor short. The calculation process of the gradient descent algorithm is as shown in formula 5:
if the loss matrix does not converge, step 203-206 is repeated until the obtained loss matrix satisfies the convergence condition.
207. And if the convergence is achieved, determining the initial weight matrix corresponding to the convergence as a target weight matrix. .
And when the loss matrix meets the convergence condition, taking the initial weight matrix corresponding to the condition meeting the convergence condition as a target weight matrix, thereby completing the determination of the target weight matrix. In the process of determining the target weight matrix, the cyclic condition is whether the loss matrix converges or not.
Fig. 3 is a second flowchart of the block chain-based credit rating determination method according to the present invention.
As shown in fig. 3, the second flowchart of the credit rating determination method based on the block chain provided in this embodiment specifically includes the following steps:
301. different roles are set for each node in the blockchain system.
302. And configuring different operation authorities for each different role so as to realize the initialization of the block chain system.
The block chain system has the function of providing the authenticity guarantee of the full-flow data for credit investigation data transmission. The nodes on the block chain provide identity authentication service for the nodes through the CA authentication technology and the asymmetric encryption technology; distributed trust is realized through a decentralized and consensus mechanism, and the information isolated island problem is solved; the data right determination problem is solved through the traceability performance of the data right determination system; the risk of data leakage can be reduced by its non-tamperable nature.
According to the description of the overall architecture, different authorities are configured for role information of nodes on the credit blockchain system, and therefore initialization of the blockchain system is completed. And meanwhile, determining common parameter configuration information such as the longest block interval, the maximum transaction number of the blocks, the maximum byte number of the blocks and the like, and writing the common parameter configuration information into the created blocks. The block chain mainly has two roles, and the data provider node is responsible for collecting credit investigation data information and carrying out preprocessing and data uplink work on the collected credit investigation data; the data demander node can acquire data from the block chain according to information such as data numbers and the like, and can call a Softmax regression model for credit evaluation.
303. And storing credit investigation data of any enterprise through the initialized blockchain system, wherein the credit investigation data of any enterprise is uploaded to the initialized blockchain system by a data provider according to a preset format.
When a data provider node successfully acquires credit investigation information in a certain field, the data cannot be directly uploaded to a block chain, the data also needs to be processed according to a certain format and is convenient to store in the block chain, non-quantized indexes can be processed into quantized indexes in the data preprocessing process, continuous data can be processed into discrete data, or all credit investigation indexes are completely converted into shaping data, and the processing can reduce the storage space occupied by the data in the block chain on one hand and facilitate the subsequent processing of credit investigation data by a credit investigation model on the other hand.
When the data processing is finished, a credit investigation data uploading task is carried out, a data provider node needs to endow each credit investigation data with a unique number for uniquely identifying the credit investigation data, then a data uploading intelligent contract is called, and the data is stored in a world state database in a Key-Value pair (Key-Value) mode.
The specific steps of 304-307 are explicitly explained in the steps 101-104 of the above embodiment, and therefore, they are not specifically explained in this embodiment, and can be understood by referring to each other.
For example, a credit investigation scene of a power grid material supply chain type enterprise is used for providing credit investigation and evaluation services for the power grid material supply chain type enterprise, and credit investigation data credit levels participating in evaluation are classified into A, B, C, D categories in total. Standing in the angle of a power grid enterprise, selecting credit assessment indexes from the angles of power grid safety construction, stable operation and emergency guarantee requirements as guidance, integrating quality, efficiency, regulation, benefit, green, innovation and the like, wherein 24 credit assessment indexes are selected in total, namely ' enterprise registered capital ', ' registered employees ', ' equipment fault rate ', ' selective inspection qualification rate ', ' product satisfaction ', ' supply completion rate ', ' supply accuracy ', ' supply punctuality ', ' open purchase rate ', ' purchase regulation ', ' standardized material application rate ', ' selective inspection coverage rate ', ' electric charge settlement condition ', ' electric power contract execution condition ', ' enterprise annual service quality contract rate ', ' clean self-regulation index ', ' safety accident occurrence times ', ' professional health management system certification ', ' emergency power management The method comprises the following steps of ' illegal electricity utilization behavior ', ' whether environmental protection certification passes ' or not ', ' waste material disposal rate ', ' management innovation number ' and ' standard popularization degree '. Each credit investigation data has 26 parameters including ID of credit investigation data, 24 credit investigation indexes and credit rating of credit investigation data. Some of the 24 selected credit investigation indexes are qualitative indexes, and some of the 24 selected credit investigation indexes are quantitative indexes, so that the 24 selected credit investigation indexes are converted into shaping data for the convenience of subsequent processing and training of the data. The standard promotional index was processed as in table 1.
TABLE 1
The uploaded data content is stored in a world state database of the block chain in a key value pair mode, the credit ID is used as a main key to uniquely identify the credit data, 24 credit indexes and 1Marshal () converts individual credit levels to data in json format. A world state database adopted by the block chain network is CouchDB, and the world state database can support the realization of rich query, interval query and other work on data in json format, so that the mode of acquiring the node data has more selectivity; meanwhile, the method can also realize field-level security by screening and shielding various attributes in the transaction so as to realize data protection. When credit investigation evaluation is needed, acquiring credit investigation Data of the power grid material supply chain type enterprises from the block chain system, respectively storing 24 credit indexes and 1 credit grade into a Data matrix and a Label matrix, and putting the Data matrix and the Label matrix into a target Weight matrix Weight for the power grid material supply chain type enterprises to be trained through a Softmax regression algorithm to obtain the credit investigation of the power grid material supply chain type enterprises24×4。
In the credit investigation determination method, a block chain network data transmission mode is used for replacing a network point-to-point data transmission mode in a traditional credit investigation model, a block chain architecture and a Softmax regression algorithm are introduced, and the credit investigation data transmission full-process credibility and accurate credit investigation results are realized. Different roles are designed for the block chain link points, and corresponding functions are completed through intelligent contracts, so that credit investigation data in a block chain network are transmitted orderly and reliably. By means of a Softmax regression algorithm, existing credit investigation data is fully utilized, a target weight matrix which can be effectively used for credit investigation evaluation is obtained through model training, the influence of human subjective factors in the credit investigation evaluation process is reduced, and the scientificity and accuracy of credit investigation evaluation are improved.
Based on the same general inventive concept, the invention also protects a credit investigation grade determination device based on the block chain.
Fig. 4 is a schematic structural diagram of a block chain-based credit rating determination apparatus according to the present invention.
As shown in fig. 4, the credit rating determination apparatus based on a block chain according to this embodiment includes:
the acquisition module 10 is configured to acquire credit investigation data of a target enterprise in a block chain system, where the credit investigation data includes credit investigation indexes;
the conversion module 20 is used for converting the credit investigation indicator into a test matrix;
a determining module 30, configured to multiply the test matrix with a predetermined target weight matrix to obtain a target score matrix; and carrying out normalization processing on the target score matrix, and determining the category of the credit grade corresponding to a column with the highest score as the credit investigation grade of the target enterprise.
The credit investigation level determination device based on the block chain provided by the embodiment obtains credit investigation data of a target enterprise, credit investigation indexes of a credit investigation data packet and converts the credit investigation indexes into a test matrix in the block chain system; multiplying the test matrix by a predetermined target weight matrix to obtain a target score matrix; the target scoring matrix is normalized through a Softmax regression algorithm, the category of a list of corresponding credit levels with the highest score is determined to serve as the credit investigation level of a target enterprise, so that the safety and the accuracy of data transmission of credit investigation indexes are guaranteed due to the characteristic that a block chain system cannot be tampered, meanwhile, due to the calculation models of the target weighting matrix and the Softmax regression algorithm, the process of determining the credit investigation levels through the credit investigation indexes can be automatically completed, the level determination process is free of manual participation, and the accuracy and the reliability of credit investigation level determination results are effectively improved.
Further, the apparatus for determining a credit rating based on a block chain in this embodiment further includes a target weight matrix determining module, configured to:
acquiring any training data set;
converting the training data set into a sample data matrix, and marking the credit level corresponding to each row of data in the sample data matrix as a marking probability array;
obtaining a sample probability matrix based on the sample data matrix according to the initial weight matrix and the Softmax function;
determining a loss function according to the sample probability matrix and the mark probability array, and obtaining a loss matrix according to the loss function;
judging whether the loss matrix is converged;
if not, updating the initial weight matrix based on a gradient descent algorithm and a preset learning rate;
and if the convergence is achieved, determining the initial weight matrix corresponding to the convergence as a target weight matrix.
Further, the target weight matrix determining module in the block chain-based credit rating determining apparatus of this embodiment is specifically configured to:
multiplying the sample data matrix by an initial weight matrix to obtain a sample score matrix;
and carrying out normalization processing on each row of data in the sample score matrix by using a Softmax function to obtain a sample probability matrix.
Further, the loss function in this embodiment includes a first loss function and/or a second loss function;
the first loss function is a balanced cross entropy loss function;
the second loss function is a combined function that balances the combination of the cross-entropy loss function and the center loss function.
Further, the apparatus for determining a credit rating based on a block chain in this embodiment further includes: an initialization module to:
setting different roles for each node in the block chain system;
and configuring different operation authorities for each different role so as to realize the initialization of the block chain system.
Further, the credit investigation level determination device based on the block chain in this embodiment further includes a data uploading module, configured to:
and storing credit investigation data of any enterprise through the initialized blockchain system, wherein the credit investigation data of any enterprise is uploaded to the initialized blockchain system by a data provider according to a preset format.
Further, the obtaining module 10 in this embodiment is specifically configured to:
acquiring a demand instruction of a data demander, wherein the demand instruction carries an enterprise identifier and a credit investigation index;
and inquiring credit investigation data of a target enterprise in the block chain system according to the enterprise identification.
Embodiments relating to apparatus parts have been specifically described in corresponding method embodiments, which may be understood by reference to each other.
Fig. 5 is a schematic structural diagram of an electronic device provided in the present invention.
As shown in fig. 5, the electronic device may include: a processor (processor)510, a communication Interface (Communications Interface)520, a memory (memory)530, and a communication bus 540, wherein the processor 510, the communication Interface 520, and the memory 530 communicate with each other via a communication bus 840. The processor 510 may call logic instructions in the memory 530 to perform a blockchain based credit level determination method comprising: acquiring credit investigation data of a target enterprise in a block chain system, wherein the credit investigation data comprises credit investigation indexes; converting the credit investigation index into a test matrix; multiplying the test matrix by a predetermined target weight matrix to obtain a target score matrix; and carrying out normalization processing on the target scoring matrix through a Softmax regression algorithm, and determining the category of the credit grade corresponding to a column with the highest score as the credit investigation grade of the target enterprise.
Furthermore, the logic instructions in the memory 530 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
In another aspect, the present invention further provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, which when executed by a computer, enable the computer to perform the block chain-based credit rating determination method provided by the above methods, the method comprising: acquiring credit investigation data of a target enterprise in a block chain system, wherein the credit investigation data comprises credit investigation indexes; converting the credit investigation index into a test matrix; multiplying the test matrix by a predetermined target weight matrix to obtain a target score matrix; and carrying out normalization processing on the target scoring matrix through a Softmax regression algorithm, and determining the category of the credit grade corresponding to a column with the highest score as the credit investigation grade of the target enterprise.
In yet another aspect, the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, the computer program being implemented by a processor to execute the above-mentioned methods for determining a credit rating based on a blockchain, the method comprising: acquiring credit investigation data of a target enterprise in a block chain system, wherein the credit investigation data comprises credit investigation indexes; converting the credit investigation index into a test matrix; multiplying the test matrix by a predetermined target weight matrix to obtain a target score matrix; and carrying out normalization processing on the target scoring matrix through a Softmax regression algorithm, and determining the category of the credit grade corresponding to a column with the highest score as the credit investigation grade of the target enterprise.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A credit investigation grade determination method based on a block chain is characterized by comprising the following steps:
acquiring credit investigation data of a target enterprise in a block chain system, wherein the credit investigation data comprises credit investigation indexes;
converting the credit investigation index into a test matrix;
multiplying the test matrix by a predetermined target weight matrix to obtain a target score matrix;
and carrying out normalization processing on the target score matrix, and determining the category of the credit level corresponding to a column with the highest score as the credit investigation level of the target enterprise.
2. The method according to claim 1, wherein before obtaining credit investigation data of a target enterprise in the blockchain system, the method further comprises:
acquiring any training data set;
converting the training data set into a sample data matrix, and marking the credit level corresponding to each row of data in the sample data matrix as a marking probability array;
obtaining a sample probability matrix based on the sample data matrix according to the initial weight matrix and the Softmax function;
determining a loss function according to the sample probability matrix and the mark probability array, and obtaining a loss matrix according to the loss function;
judging whether the loss matrix is converged;
if not, updating the initial weight matrix based on a gradient descent algorithm and a preset learning rate;
and if the convergence is achieved, determining the initial weight matrix corresponding to the convergence as a target weight matrix.
3. The method according to claim 2, wherein obtaining a sample probability matrix based on the sample data matrix according to an initial weight matrix and a Softmax function comprises:
multiplying the sample data matrix by an initial weight matrix to obtain a sample score matrix;
and carrying out normalization processing on each row of data in the sample score matrix by using a Softmax function to obtain a sample probability matrix.
4. The method for determining a credit rating based on a block chain according to claim 2, wherein the loss function comprises a first loss function and/or a second loss function;
the first loss function is a balanced cross entropy loss function;
the second loss function is a combined function that balances the combination of the cross-entropy loss function and the center loss function.
5. The method according to claim 1, wherein before obtaining credit investigation data of a target enterprise in the blockchain system, the method further comprises:
setting different roles for each node in the block chain system;
and configuring different operation authorities for each different role so as to realize the initialization of the block chain system.
6. The method according to claim 5, wherein before obtaining credit investigation data of a target enterprise in the blockchain system, the method further comprises:
and storing credit investigation data of any enterprise through the initialized blockchain system, wherein the credit investigation data of any enterprise is uploaded to the initialized blockchain system by a data provider according to a preset format.
7. The method according to claim 1, wherein the obtaining credit investigation data of a target enterprise in the blockchain system comprises:
acquiring a demand instruction of a data demander, wherein the demand instruction carries an enterprise identifier and a credit investigation index;
and inquiring credit investigation data of a target enterprise in the block chain system according to the enterprise identification.
8. An apparatus for determining credit rating based on a block chain, comprising:
the acquisition module is used for acquiring credit investigation data of a target enterprise in a block chain system, wherein the credit investigation data comprises credit investigation indexes;
the conversion module is used for converting the credit investigation index into a test matrix;
the determining module is used for multiplying the test matrix and a predetermined target weight matrix to obtain a target score matrix; and carrying out normalization processing on the target score matrix, and determining the category of the credit level corresponding to a column with the highest score as the credit investigation level of the target enterprise.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program performs the steps of the block chain based credit rating determination method according to any of claims 1 to 7.
10. A non-transitory computer readable storage medium, having stored thereon a computer program, when being executed by a processor, the computer program implementing the steps of the method for determining a block chain based credit rating according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110699625.1A CN113535848A (en) | 2021-06-23 | 2021-06-23 | Block chain-based credit investigation grade determination method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110699625.1A CN113535848A (en) | 2021-06-23 | 2021-06-23 | Block chain-based credit investigation grade determination method, device, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113535848A true CN113535848A (en) | 2021-10-22 |
Family
ID=78096517
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110699625.1A Pending CN113535848A (en) | 2021-06-23 | 2021-06-23 | Block chain-based credit investigation grade determination method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113535848A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114974593A (en) * | 2022-07-11 | 2022-08-30 | 江苏优创生物医学科技有限公司 | Enterprise health management evaluation method and system based on block chain |
| CN115082190A (en) * | 2022-06-21 | 2022-09-20 | 京东科技信息技术有限公司 | Credit investigation index testing method and device, storage medium and electronic equipment |
| CN117235773A (en) * | 2023-11-15 | 2023-12-15 | 山东顺国电子科技有限公司 | Intelligent data classification management method, system and storage medium for big data center |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107506941A (en) * | 2017-09-09 | 2017-12-22 | 杭州数立方征信有限公司 | A kind of enterprise in charge of construction's credit assessment method and system based on big data technology |
| CN108596443A (en) * | 2018-04-02 | 2018-09-28 | 广东电网有限责任公司 | A kind of Electricity customers method for evaluating credit rating based on multi-dimensional data |
| CN110414983A (en) * | 2019-08-06 | 2019-11-05 | 北京软通智慧城市科技有限公司 | Reference information processing method, device, equipment and storage medium based on block chain |
| CN110599098A (en) * | 2019-09-17 | 2019-12-20 | 上海天地汇供应链科技有限公司 | Logistics industry-based block chain credit investigation management method and system |
| CN111754044A (en) * | 2020-06-30 | 2020-10-09 | 深圳前海微众银行股份有限公司 | Employee behavior auditing method, device, equipment and readable storage medium |
| CN112685766A (en) * | 2020-12-15 | 2021-04-20 | 广西大学 | Enterprise credit investigation management method and device based on block chain, computer equipment and storage medium |
-
2021
- 2021-06-23 CN CN202110699625.1A patent/CN113535848A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107506941A (en) * | 2017-09-09 | 2017-12-22 | 杭州数立方征信有限公司 | A kind of enterprise in charge of construction's credit assessment method and system based on big data technology |
| CN108596443A (en) * | 2018-04-02 | 2018-09-28 | 广东电网有限责任公司 | A kind of Electricity customers method for evaluating credit rating based on multi-dimensional data |
| CN110414983A (en) * | 2019-08-06 | 2019-11-05 | 北京软通智慧城市科技有限公司 | Reference information processing method, device, equipment and storage medium based on block chain |
| CN110599098A (en) * | 2019-09-17 | 2019-12-20 | 上海天地汇供应链科技有限公司 | Logistics industry-based block chain credit investigation management method and system |
| CN111754044A (en) * | 2020-06-30 | 2020-10-09 | 深圳前海微众银行股份有限公司 | Employee behavior auditing method, device, equipment and readable storage medium |
| CN112685766A (en) * | 2020-12-15 | 2021-04-20 | 广西大学 | Enterprise credit investigation management method and device based on block chain, computer equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 蒋宏成: "基于BSC 的中小企业信用多级模糊综合评价", 《财会通讯》, 20 April 2016 (2016-04-20) * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115082190A (en) * | 2022-06-21 | 2022-09-20 | 京东科技信息技术有限公司 | Credit investigation index testing method and device, storage medium and electronic equipment |
| CN114974593A (en) * | 2022-07-11 | 2022-08-30 | 江苏优创生物医学科技有限公司 | Enterprise health management evaluation method and system based on block chain |
| CN117235773A (en) * | 2023-11-15 | 2023-12-15 | 山东顺国电子科技有限公司 | Intelligent data classification management method, system and storage medium for big data center |
| CN117235773B (en) * | 2023-11-15 | 2024-02-02 | 山东顺国电子科技有限公司 | Intelligent data classification management method, system and storage medium for big data center |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108564286B (en) | Artificial intelligent financial wind-control credit assessment method and system based on big data credit investigation | |
| CN113535848A (en) | Block chain-based credit investigation grade determination method, device, equipment and storage medium | |
| CN108898476A (en) | A kind of loan customer credit-graded approach and device | |
| CN109670104A (en) | Information-pushing method, unit and storage medium based on machine learning | |
| CN110147823A (en) | A kind of air control model training method, device and equipment | |
| CN109740622A (en) | Image labeling task crowdsourcing method and system based on the logical card award method of block chain | |
| CN112053061A (en) | Method and device for identifying surrounding label behaviors, electronic equipment and storage medium | |
| CN113014566B (en) | Malicious registration detection method and device, computer readable medium and electronic device | |
| CN110503565A (en) | Behaviorist risk recognition methods, system, equipment and readable storage medium storing program for executing | |
| CN112116103A (en) | Method, device and system for evaluating personal qualification based on federal learning and storage medium | |
| CN111950889A (en) | Client risk assessment method and device, readable storage medium and terminal equipment | |
| CN111882140A (en) | Risk evaluation method, model training method, device, equipment and storage medium | |
| CN113570260A (en) | Task allocation method, computer-readable storage medium and electronic device | |
| CN108345979B (en) | Service testing method and device | |
| CN117540803A (en) | Decision engine configuration method and device based on large model, electronic equipment and medium | |
| CN108492169A (en) | Risk Modeling method and system based on credit card approval scene are realized | |
| CN110704803A (en) | Target object evaluation value calculation method and device, storage medium and electronic device | |
| CN112000862B (en) | Data processing method and device | |
| CN113837803A (en) | Second-hand mobile phone recycling price prediction algorithm based on multi-model fusion | |
| CN113011966A (en) | Credit scoring method and device based on deep learning | |
| CN115982646B (en) | Management method and system for multisource test data based on cloud platform | |
| CN117196630A (en) | Transaction risk prediction method, device, terminal equipment and storage medium | |
| CN117056323A (en) | Data quality assessment method based on data access behaviors | |
| CN111353728A (en) | Risk analysis method and system | |
| CN117422321A (en) | Patent value evaluation method, device, electronic equipment and storage medium |
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 |