A credit determination method, apparatus and device

Abstract

Embodiments of the present specification provide a credit determination method, device and equipment, wherein the method comprises: obtaining an initial rough set; wherein the initial rough set contains a plurality of original feature information of institutions; performing information reduction on the initial rough set by using a discernible matrix reduction algorithm to obtain an evidence information set; each feature information after information reduction serves as a separate evidence; obtaining basic probability assignment values of a recognition framework corresponding to each evidence in the evidence information set; and performing uncertain reasoning by using evidence theory according to the basic probability assignment values of the recognition framework corresponding to each evidence to obtain a representation value of each institution; wherein the representation value is used to represent the credit of the institution. In the embodiments of the present specification, the credit of each institution can be efficiently and accurately determined in the case that the original data amount of each institution is small, and the data exists uncertain or inaccurate knowledge expression.

Description

TECHNICAL FIELD

[0001] Embodiments of the present specification relate to the technical field of artificial intelligence, and particularly relate to a credit degree determination method, device and equipment. BACKGROUND

[0002] Based on the original information, making accurate identification and judgment on small and micro enterprises is a key problem in risk control calculation research. In the prior art, the credit score of an enterprise is usually predicted by training a machine learning model. However, since the amount of data available for modeling of small and micro enterprises is small, the model trained by small samples is prone to overfitting of small samples and underfitting of the target task, and therefore, the method of training a machine learning model cannot accurately determine the credit score of an enterprise.

[0003] At present, no effective solution has been proposed for the above problems. SUMMARY

[0004] Embodiments of the present specification provide a credit degree determination method, device and equipment to solve the problem that the credit score of an enterprise cannot be accurately determined in the prior art.

[0005] The embodiment of the present specification provides a credit degree determination method, which comprises: obtaining an initial rough set; wherein the initial rough set contains original feature information of a plurality of institutions; performing information reduction on the initial rough set by using a discernible matrix reduction algorithm to obtain an evidence information set; wherein each feature information after information reduction serves as a separate evidence; obtaining basic probability assignment values of a recognition framework corresponding to each evidence in the evidence information set; and obtaining a representation value of each institution by using evidence theory for uncertain reasoning according to the basic probability assignment values of the recognition framework corresponding to each evidence; wherein the representation value is used to represent the credit degree of the institution.

[0006] The embodiment of the present specification also provides a credit degree determination device, which comprises: a first obtaining module configured to obtain an initial rough set; wherein the initial rough set contains original feature information of a plurality of institutions; an information reduction module configured to perform information reduction on the initial rough set by using a discernible matrix reduction algorithm to obtain an evidence information set; wherein each feature information after information reduction serves as a separate evidence; a second obtaining module configured to obtain basic probability assignment values of a recognition framework corresponding to each evidence in the evidence information set; and a processing module configured to obtain a representation value of each institution by using evidence theory for uncertain reasoning according to the basic probability assignment values of the recognition framework corresponding to each evidence; wherein the representation value is used to represent the credit degree of the institution.

[0007] The embodiment of the present specification further provides a credit degree determination device, comprising a processor and a memory for storing processor executable instructions, the processor executes the instructions to realize the steps of any one of the method embodiments of the present specification.

[0008] The embodiment of the present specification further provides a computer readable storage medium, which stores computer instructions, the instructions are executed to realize the steps of any one of the method embodiments of the present specification.

[0009] The embodiment of the present specification provides a credit degree determination method, an initial rough set containing original feature information of a plurality of institutions can be acquired, since the initial rough set contains redundant information, and there may be inaccurate and repeated submission of data, therefore, the information reduction of the initial rough set can be performed by using a discernible matrix reduction algorithm to obtain an evidence information set, wherein each feature information after information reduction can be used as a separate evidence. Further, the basic probability assignment value of the corresponding recognition framework of each evidence in the evidence information set can be acquired, and the representation value of each institution can be obtained by using evidence theory for uncertain reasoning according to the basic probability assignment value of the corresponding recognition framework of each evidence, wherein the representation value can be used to represent the credit degree of the institution. Therefore, the credit degree of each institution can be determined efficiently and accurately in the case that the original data amount of each institution is small, and the data contains uncertain or inaccurate knowledge. BRIEF DESCRIPTION OF DRAWINGS

[0010] The drawings described herein are used to provide further understanding of the embodiments of the present specification, and form a part of the embodiments of the present specification, and do not constitute a limitation to the embodiments of the present specification. In the drawings:

[0011] Figure 1 is a step schematic diagram of the credit degree determination method provided by the embodiments of the present specification;

[0012] Figure 2 is a structure schematic diagram of the credit degree determination device provided by the embodiments of the present specification;

[0013] Figure 3 is a structure schematic diagram of the credit degree determination device provided by the embodiments of the present specification. DETAILED DESCRIPTION

[0014] The principles and spirit of the embodiments of the present specification will be described below with reference to a number of exemplary embodiments. It should be understood that these embodiments are given only to enable those skilled in the art to better understand and implement the embodiments of the present specification, and in no way limit the scope of the embodiments of the present specification. On the contrary, these embodiments are provided to make the disclosure of the embodiments of the present specification more thorough and complete, and to fully convey the scope of the disclosure to those skilled in the art.

[0015] Those skilled in the art understand that the embodiments of the present specification can be implemented as a system, device, method or computer program product. Therefore, the embodiments of the present specification can be embodied in the form of a complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0016] Although the processes described below include a number of operations that appear in a particular order, it should be clear that these processes can include more or fewer operations, and that the operations can be executed in a different order (for example, using a parallel processor or multi-threaded environment).

[0017] Please refer to Figure 1 The embodiments of the present specification can provide a credit determination method. The credit determination method can be used to efficiently and accurately determine the credit of each institution when the amount of original data of each institution is small, and the data exists uncertain or inaccurate knowledge expression. The credit determination method can include the following steps.

[0018] S101: Obtain an initial rough set; wherein the initial rough set contains original feature information of a plurality of institutions.

[0019] In the embodiments of the present specification, an initial rough set can be obtained. The initial rough set can contain original feature information of a plurality of institutions. Rough set theory is a mathematical tool for handling uncertainty. In many actual systems, there are different degrees of uncertainty factors. The collected data often contains noise, inaccuracy or incompleteness. Rough set can effectively handle uncertain or inaccurate knowledge expression, inconsistent information analysis, reasoning based on uncertain and incomplete knowledge, data simplification under the premise of preserving information, etc.

[0020] In the embodiments of the present specification, the plurality of institutions can be institutions whose credit is to be determined. The specific institutions can be determined according to actual conditions, and the embodiments of the present specification do not limit this.

[0021] In the embodiment, the initial rough set can include the original acquired rough data. The original feature information of the institution can be information for representing the institution attributes and operation conditions. In some embodiments, the original feature information of the institution can include enterprise registration supplementary information, enterprise registration change information, shareholder registration information, unit insurance information, enterprise legal person, enterprise financial data, enterprise tax information, enterprise tax registration information, and enterprise accumulation fund deposit information. It can be understood that the original feature information can also include other information, such as institution name, enterprise business income, and enterprise basic information. The specific information can be determined according to actual conditions, and the embodiments of the specification are not limited in this regard.

[0022] In the embodiment, the initial rough set can be obtained by pulling from a preset database or by using a rule extraction method to mine from a text description in combination with a corpus. It can be understood that the initial rough set can also be obtained by other possible methods, such as searching the initial rough set in a webpage according to certain search conditions. The specific method can be determined according to actual conditions, and the embodiments of the specification are not limited in this regard. The preset database can be a database for storing historical data and data submitted or generated by users in real time.

[0023] In the embodiment, the initial rough set can be obtained by pulling from a preset database or by using a rule extraction method to mine from a text description in combination with a corpus. It can be understood that the initial rough set can also be obtained by other possible methods, such as searching the initial rough set in a webpage according to certain search conditions. The specific method can be determined according to actual conditions, and the embodiments of the specification are not limited in this regard. The preset database can be a database for storing historical data and data submitted or generated by users in real time.

[0024] In the embodiment, the initial rough set can include the original acquired rough data. The original feature information of the institution can be information for representing the institution attributes and operation conditions. In some embodiments, the original feature information of the institution can include enterprise registration supplementary information, enterprise registration change information, shareholder registration information, unit insurance information, enterprise legal person, enterprise financial data, enterprise tax information, enterprise tax registration information, and enterprise accumulation fund deposit information. It can be understood that the original feature information can also include other information, such as institution name, enterprise business income, and enterprise basic information. The specific information can be determined according to actual conditions, and the embodiments of the specification are not limited in this regard.

[0025] In the embodiment, the attribute core core(C) is usually the first step of the reduction, considering the uniqueness of the attribute core. The value of 1 in the discernibility matrix M means that the feature attribute set cannot be classified by other discriminant attributes. Therefore, the elements with a value of 1 are the attribute core and should be retained.

[0026] In the embodiment, the discernibility matrix can be calculated, and the attribute set after reduction can be obtained by calculation. The discernibility matrix is used to find all combinations that do not contain the core attributes, and the combinations are expressed as standard data sets. By finding combinations that do not contain the core attributes, the data set can be simplified.

[0027] In the embodiment, the discernibility matrix reduction algorithm can be used to reduce information of the initial rough set in the following steps: input: an information system S={U, R, V, f}, wherein S is a decision table, U is a domain, R is an attribute set, R=C∪D, C is an attribute set, D is a decision set, V represents a value range of an attribute, and f is an information function; output: a reduced re.

[0028] 1. re=Ф, core=Ф, count(ai=0), (i=1, 2, …, n), wherein ai is the ith attribute, core is a core attribute set, Ф is a reduced set, and n is an integer.

[0029] 2. A discernibility matrix M is calculated, and the same elements in the matrix are merged, and an attribute with only one element is added to the core, that is, core=core∪{ai}, and n++.

[0030] 3. redu=core, all items containing the core in M are deleted, count(ai) is calculated, and |M| is calculated.

[0031] 4. if |M|≠0

[0032] 5. An attribute with the largest attribute frequency is added to the redu, n++ is calculated, all items containing ai in M are deleted, count(ai) is recalculated, and |M| is calculated.

[0033] 6. end if

[0034] 7. if POS redu-ai(D) ≠POS C (D), it is indicated that ai is necessary in C; wherein POS C (D) represents a C positive region of D, POS redu-ai(D) redu-ai(D) positive region of D, and a set of all necessary attributes of C is referred to as a core core(C) of C.

[0035] 8. ai is retained, or ai is deleted, and n-- is calculated.

[0036] 9. end if

[0037] 10. The reduced redu is output.

[0038] In the embodiment, other possible ways can also be used to reduce information of the rough set, for example, information reduction based on information entropy. The specific method can be determined according to actual conditions, and the embodiment of the present application does not limit the same.

[0039] In the embodiment, each feature information in the reduced feature information set can be taken as an attribute, and each attribute can be taken as a separate piece of evidence. The decision table is an important concept in rough set theory, and can organize chaotic and disordered information into a clear and ordered knowledge structure. The reduced feature information set can be referred to as a decision table or an evidence information set. The actual situation can be determined, and the embodiments of the present specification are not limited in this regard.

[0040] In the embodiment, the process of deleting redundant condition attributes and rules without changing the classification ability of the original decision table is referred to as reduction, but it should be noted that the reduction of the decision table is for discrete information and data.

[0041] S103: Obtain the basic probability assignment value of the recognition framework corresponding to each piece of evidence in the evidence information set.

[0042] In the embodiment, an important step of the D-S evidence theory synthesis formula is to calculate the basic probability assignment value of the recognition framework corresponding to each piece of evidence, so the basic probability assignment value of the recognition framework corresponding to each piece of evidence in the evidence information set can be obtained.

[0043] In the embodiment, the basic probability assignment value represents a reasoning of the credibility of the target. In some embodiments, the probability assignment value can be determined according to the target type and the environmental weighting coefficient, obtained by using statistical evidence, obtained by using target speed and acceleration, determined by using target identity, etc. It can be understood that the basic probability assignment value can also be determined in other possible ways, for example, by determining the strength of the decision rule to obtain the basic probability assignment value, by obtaining a reasonable basic probability assignment value from past risk rule experience, etc. The actual situation can be determined, and the embodiments of the present specification are not limited in this regard.

[0044] In the embodiment, different probability assignment functions have a certain influence on the result of information fusion, and the probability assignment function can be determined according to the characteristics of the research object as needed.

[0045] S104: According to the basic probability assignment value of the recognition framework corresponding to each piece of evidence, the evidence theory is used to perform uncertain reasoning to obtain the representation value of each agency; wherein the representation value is used to represent the credit degree of the agency.

[0046] In the embodiment, the evidence theory can be used to perform uncertain reasoning to obtain the representation value of each agency according to the basic probability assignment value of the recognition framework corresponding to each piece of evidence; wherein the representation value can be used to represent the credit degree of the agency, so as to be applied to enterprise access and credit.

[0047] In the embodiment, the basic probability assignment values of the recognition framework corresponding to each piece of evidence of each institution can be fused by using the evidence theory, and the basic probability assignment values fused finally by each institution can be obtained. In some embodiments, the basic probability assignment values fused finally by each institution can be directly used as the representation values of each institution, or the values obtained by preprocessing the basic probability assignment values fused finally by each institution can be used as the representation values of each institution. The preprocessing can include normalization processing, calculation according to a specified formula, etc., and the specific implementation can be determined according to actual conditions, which are not limited in the embodiments of the present disclosure.

[0048] In the embodiment, the representation value can be a numerical value greater than 0, for example, 10 points, 60 points, 92 points, etc. In some embodiments, the representation value can also be a grade, for example, low credit, good credit, high credit, etc. It can be understood that other possible ways of representing the representation value can also be used, for example, A, B, C, etc., and the specific implementation can be determined according to actual conditions, which are not limited in the embodiments of the present disclosure.

[0049] In the embodiment, if the decision table after information reduction (evidence information set) is used as the input of the D-S evidence theory, each attribute in the reduced decision table can correspond to the evidence in the D-S evidence theory, so that the rough set and the D-S evidence theory can be combined to perform uncertain reasoning.

[0050] In the embodiment, the D-S evidence theory is an inaccurate reasoning theory, also known as Dempster-Shafer evidence theory, and has the ability to process uncertain information. As an uncertain reasoning method, the main feature of the evidence theory is that it satisfies a weaker condition than the Bayesian probability theory and has the ability to directly express “uncertainty” and “unknown”.

[0051] In the embodiment, in the D-S evidence theory, a complete set composed of mutually incompatible basic propositions (assumptions) is called a recognition framework, which represents all possible answers to a problem, but only one of them is correct. A subset of the framework is called a proposition. The degree of trust assigned to each proposition is called a basic probability assignment (BPA, also known as m function), and m(A) is a basic belief number, which reflects the degree of belief in A. Let m1 and m2 be basic probability assignment functions derived from two independent evidence sources (sensors), and the Dempster combination rule can calculate a new basic probability assignment function reflecting the fused information generated by the joint action of the two evidences.

[0052] From the above description, it can be seen that the embodiments of the present specification achieve the following technical effects: the initial rough set containing the original feature information of multiple institutions can be obtained. Since the initial rough set may have redundant information, and there may be and there are data inaccuracies and repeated submissions, etc., the discernible matrix reduction algorithm can be used to reduce the information of the initial rough set to obtain an evidence information set, wherein each feature information after information reduction can be used as a separate evidence. Further, the basic probability assignment value of the identification framework corresponding to each evidence in the evidence information set can be obtained, and the representation value of each institution can be obtained by using evidence theory for uncertain reasoning according to the basic probability assignment value of the identification framework corresponding to each evidence, wherein the representation value can be used to represent the credit degree of the institution. Thus, the credit degree of each institution can be determined efficiently and accurately in the case that the original data amount of each institution is small and the data has uncertain or inaccurate knowledge expression.

[0053] In one embodiment, before the initial rough set is reduced in information by using the discernible matrix reduction algorithm to obtain the evidence information set, the initial rough set can also include: discretizing the initial rough set to obtain a first rough set. Further, the first rough set can be de-redundant to obtain a second rough set. The second rough set can be cleaned to obtain a target rough set.

[0054] In the present embodiment, since the original data obtained cannot be directly used for rough set reduction, the original data can be discretized. The steps of discretization can include: 1) by calculating, a data interval is taken out, usually Xi, Xi+1; 2) from the data interval, it is judged whether Xi, Xi+1 are equal and whether d(Xi) and d(Xi+1) are comparable, if not equal, go to the next step, if equal, continue to select appropriate Xi, Xi+1 until the appropriate interval is found, wherein d(Xi) represents the first step of rough data; 3) take the middle value of a(Xi) and a(Xi+1) as the most appropriate discretization point. Considering the characteristics of rough set information system, the data is discretized according to the result breakpoint subset, wherein a(Xi) represents the point close to the discretization point.

[0055] In the present embodiment, removing redundant data can include removing or merging some highly related features to reduce the dimension of data. In some embodiments, business analysis can also be used to remove information irrelevant to evidence reasoning and business. Among them, the variables with low correlation to the label analyzed by WOE value can be used as information irrelevant to business. WOE (Weight of Evidence) is the weight of evidence, which is the logarithm of the proportion of good and bad customers under a certain value of a character variable or a certain segment of a continuous variable.

[0056] In the embodiment, the data cleaning refers to the last procedure of finding and correcting the identifiable errors in the data file, including checking data consistency, processing invalid values and missing values, etc. Therefore, the second rough set can be subjected to data cleaning to obtain a target rough set. The data cleaning can include: simple deletion, taking the enterprise tax credit rating feature as an example, the data of missing enterprise tax credit rating and credit rating M level can be deleted; filling the median, the median of feature statistics is used as the value filled in the missing value; filling the mode, the value with the largest frequency of occurrence is used as the value filled in the missing value according to feature statistics; filling the mean, the mean is used as the value filled in the missing value according to feature statistics.

[0057] In the embodiment, the initial rough set can also be preprocessed in other possible ways, for example, normalization processing, etc. The specific preprocessing method can be determined according to actual conditions, which is not limited in the embodiments of the present specification.

[0058] In one embodiment, the de-redundancy operation on the first rough set to obtain the second rough set can include: determining the redundant features in the first rough set by using a target analysis algorithm; wherein the target analysis algorithm includes histogram distribution analysis, similarity analysis or principal component analysis. Further, the redundant features in the first rough set can be removed or combined to obtain the second rough set.

[0059] In the embodiment, before data de-redundancy, the three-dimensional redundant features in the first rough set can be identified, and the redundant features in the first rough set can be determined by using a target analysis algorithm. The target analysis algorithm can include histogram distribution analysis, similarity analysis, principal component analysis, etc. It can be understood that other possible ways can also be used to identify the redundant features, and the specific method can be determined according to actual conditions, which is not limited in the embodiments of the present specification.

[0060] In the embodiment, the histogram is a statistical report chart, which represents the data distribution by a series of vertical stripes or line segments with different heights. The histogram can be used to analyze the regularity of the data, and the distribution state of the data can be intuitively observed, so that the overall distribution of the data can be easily judged. Principal component analysis (PCA) is a statistical process that uses an orthogonal transformation to convert a set of possibly correlated variables (entities, each entity has different numerical values) into a set of linearly uncorrelated variables called principal components. It can be completed by eigenvalue decomposition of the data covariance (or correlation) matrix or singular value decomposition of the data matrix, usually after the normalization step of the initial data.

[0061] In the embodiment, the identified redundant features in the first rough set can be removed, and in some embodiments, the identified redundant features can also be merged. The actual situation can be determined, and the embodiments of the present specification are not limited in this regard.

[0062] In one embodiment, the initial rough set is reduced in information by using a discernible matrix reduction algorithm to obtain an evidence information set, which can include reducing the target rough set in information by using a discernible matrix reduction algorithm to obtain an evidence information set.

[0063] In the embodiment, the original data obtained cannot be directly used for rough set reduction, so the preprocessed target rough set can be reduced in information by using a discernible matrix reduction algorithm to obtain an evidence information set.

[0064] In the embodiment, the above-mentioned evidence information set can be recorded in the form of a table, or in the form of an image or text, and the actual situation can be determined, and the embodiments of the present specification are not limited in this regard.

[0065] In one specific embodiment, the reduced evidence information set can be as shown in Table 1.

[0066] Table 1

[0067]

[0068] In one embodiment, according to the basic probability assignment value of the identification framework corresponding to each piece of evidence, the representation value of each agency can be obtained by using evidence theory for uncertain reasoning, which can include fusing the basic probability assignment value of the identification framework corresponding to each piece of evidence of each agency by using evidence theory to obtain the evidence fusion result of each agency. Further, the representation value of each agency can be obtained according to the evidence fusion result of each agency.

[0069] In the embodiment, the cornerstone of the evidence theory is the fusion rule. The fusion rule can include: two pieces of evidence m i and m j of an agency can be taken to determine whether the conflict factor of m i and m j is greater than a preset threshold value. If the conflict factor of m i and m j is less than or equal to the preset threshold value, m i and m j are merged, and the merged new evidence is fused with other evidence of the agency; if the conflict factor of m i and m j is greater than the preset threshold value, it means that m i and m jThe combination of the two can be put aside for the time being, and another evidence m is taken n , and m i , and m n are combined. In turn, until the last two evidences are fused, it can be determined whether the conflict factor of the last two evidences is greater than a preset threshold value, and if not, the two evidences can be directly combined to obtain the final evidence fusion result; if it is determined to be yes, the two evidences are fused by using the mean weighted method.

[0070] In this embodiment, specifically, the basic probability assignment values of the two evidences can be fused. The above-mentioned preset threshold value is a value greater than 0, for example, 0.9, 0.99, 1, 1.1, and specifically, it can be determined according to actual conditions, and the embodiments of the present application do not limit this.

[0071] In this embodiment, the basic probability assignment value finally fused by each agency can be directly used as the representation value of each agency, or the value of the basic probability assignment value finally fused by each agency after preprocessing can be used as the representation value of each agency. The above-mentioned preprocessing can include normalization processing, calculation according to a specified formula, or reasonable model probability superposition, and specifically, it can be determined according to actual conditions, and the embodiments of the present application do not limit this.

[0072] In some embodiments, the basic probability assignment value can be obtained by determining the decision rule strength, for example: different regions are given scores, and the scores given to the southeast coastal provinces are higher than those given to the central region. Of course, the determination method of the basic probability assignment value is not limited to the above examples, and those skilled in the art can also make other changes under the inspiration of the technical essence of the embodiments of the present application, as long as the functions and effects achieved are the same or similar to those of the embodiments of the present application, which should be covered within the protection scope of the embodiments of the present application.

[0073] In one embodiment, the mean weighted method can be used to resolve the conflict of the evidence that generates a paradox.

[0074] In this embodiment, the mean weighted method can be used to resolve the conflict of the evidence that generates a paradox according to the following formula:

[0075]

[0076] , wherein, is the basic probability assignment value of the new evidence after fusion; m1 is evidence 1; m2 is evidence 2; m1(A) is the basic probability assignment value of evidence 1; and m2(A) is the basic probability assignment value of evidence 2.

[0077] In an embodiment, the characteristic information can include enterprise registration supplementary information, enterprise registration change information, shareholder registration information, unit insurance participation information, enterprise legal person, enterprise financial data, enterprise tax information, enterprise tax registration information, enterprise provident fund payment information, and the like.

[0078] In the embodiment, the characteristic information of the above-mentioned institution can be related information for representing the attribute and operation of the institution. In some embodiments, the characteristic information of the above-mentioned institution can include enterprise registration supplementary information, enterprise registration change information, shareholder registration information, unit insurance participation information, enterprise legal person, enterprise financial data, enterprise tax information, enterprise tax registration information, enterprise provident fund payment information, and the like. It can be understood that the above-mentioned original characteristic information can also include other information, for example, institution name, enterprise business income, enterprise basic information, and the like. The specific information can be determined according to the actual situation, and the embodiments of the present specification are not limited thereto.

[0079] Based on the same inventive concept, the embodiments of the present specification also provide a credit degree determination device, as described in the following embodiments. Since the principle of solving the problem of the credit degree determination device is similar to that of the credit degree determination method, the implementation of the credit degree determination device can refer to the implementation of the credit degree determination method, and the repeated parts will not be described herein. The term "unit" or "module" used below can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware or a combination of software and hardware is also possible and is contemplated. Figure 2 is a structural block diagram of the credit degree determination device of the embodiments of the present specification, as shown in Figure 2 The structure can include a first acquisition module 201, an information reduction module 202, a second acquisition module 203, and a processing module 204, which will be described below.

[0080] The first acquisition module 201 can be used to acquire an initial rough set; wherein the initial rough set contains a plurality of original characteristic information of institutions;

[0081] The information reduction module 202 can be used to reduce the information of the initial rough set by using a discernible matrix reduction algorithm to obtain an evidence information set; wherein each characteristic information after information reduction is used as a separate evidence;

[0082] The second acquisition module 203 can be used to acquire the basic probability assignment value of the identification framework corresponding to each evidence in the evidence information set;

[0083] The processing module 204 can be configured to obtain the representation value of each institution by using evidence theory to perform uncertain reasoning according to the basic probability assignment value of the identification framework corresponding to each piece of evidence; wherein the representation value is used to represent the credit degree of the institution.

[0084] From the above description, it can be seen that the embodiments of the present specification achieve the following technical effects: the initial rough set containing the original feature information of multiple institutions can be obtained. Since the initial rough set may contain redundant information, and may contain inaccurate and repeated submission data, the information reduction algorithm of the identifiable matrix can be used to reduce the information of the initial rough set to obtain the evidence information set. Each feature information after information reduction can be used as a separate piece of evidence. Further, the basic probability assignment value of the identification framework corresponding to each piece of evidence in the evidence information set can be obtained, and the representation value of each institution can be obtained by using evidence theory to perform uncertain reasoning according to the basic probability assignment value of the identification framework corresponding to each piece of evidence. The representation value can be used to represent the credit degree of the institution. Therefore, the credit degree of each institution can be determined efficiently and accurately in the case that the original data amount of each institution is small and the data contains uncertain or inaccurate knowledge.

[0085] The embodiments of the present specification also provide an electronic device, which can specifically refer to Figure 3 The electronic device provided by the credit degree determination method provided by the embodiments of the present specification can specifically include an input device 31, a processor 32, and a memory 33. The input device 31 can be specifically configured to input an initial rough set. The processor 32 can be specifically configured to obtain the initial rough set, wherein the initial rough set contains the original feature information of multiple institutions; reduce the information of the initial rough set by using an identifiable matrix reduction algorithm to obtain an evidence information set; wherein each feature information after information reduction is used as a separate piece of evidence; obtain the basic probability assignment value of the identification framework corresponding to each piece of evidence in the evidence information set; and obtain the representation value of each institution by using evidence theory to perform uncertain reasoning according to the basic probability assignment value of the identification framework corresponding to each piece of evidence; wherein the representation value is used to represent the credit degree of the institution. The memory 33 can be specifically configured to store the representation value of each institution and other parameters.

[0086] In the embodiment, the input device can be one of the main devices for information exchange between the user and the computer system. The input device can include a keyboard, a mouse, a camera, a scanner, a light pen, a handwriting input board, a voice input device, etc. The input device is used to input raw data and programs for processing the data into the computer. The input device can also obtain data transmitted by other modules, units, devices. The processor can be implemented in any appropriate manner. For example, the processor can take the form of a microprocessor or a processor and a computer readable medium storing computer readable program code (such as software or firmware) executable by the (micro) processor, logic gates, switches, application specific integrated circuits (ASIC), programmable logic controllers, and embedded microcontrollers, etc. The memory can be a memory device used to store information in modern information technology. The memory can include multiple levels, and in a digital system, as long as it can store binary data, it can be a memory; in an integrated circuit, a circuit without a physical form with a storage function is also called a memory, such as RAM, FIFO, etc.; in a system, a storage device with a physical form is also called a memory, such as a memory stick, a TF card, etc.

[0087] In the embodiment, the functions and effects of the electronic device can be explained in comparison with other embodiments, and will not be repeated here.

[0088] The embodiment of the present application also provides a computer storage medium based on a credit determination method, which stores computer program instructions. When the computer program instructions are executed, the credit determination method can be implemented. The method comprises the following steps: obtaining an initial rough set; wherein the initial rough set contains a plurality of original feature information of an institution; performing information reduction on the initial rough set by using a discernible matrix reduction algorithm to obtain an evidence information set; wherein each feature information after information reduction is used as a separate evidence; obtaining a basic probability assignment value of an identification framework corresponding to each evidence in the evidence information set; obtaining a representation value of each institution by using evidence theory for uncertain reasoning according to the basic probability assignment value of the identification framework corresponding to each evidence; wherein the representation value is used to represent the credit of the institution.

[0089] In the present embodiment, the storage medium includes, but is not limited to, a Random Access Memory (RAM), a Read-Only Memory (ROM), a Cache, a Hard Disk Drive (HDD), or a Memory Card. The storage medium can be used to store computer program instructions. The network communication unit can be an interface configured according to a standard set by a communication protocol, and used to perform network connection communication.

[0090] In the present embodiment, the functions and effects realized by the program instructions stored in the computer storage medium can be explained in comparison with other embodiments, and will not be described here.

[0091] Obviously, those skilled in the art should understand that each module or each step of the above-described embodiments of the present specification can be realized by a general computing device, which can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, and alternatively, each module or each step can be realized by program codes executable by a computing device, so that each module or each step can be stored in a storage device and executed by a computing device, and in some cases, the steps shown or described can be executed in an order different from that described herein, or each module or step can be manufactured as an individual integrated circuit module, or a plurality of modules or steps can be manufactured as a single integrated circuit module. Thus, the embodiments of the present specification are not limited to any particular combination of hardware and software.

[0092] Although the embodiments of the present specification provide the method operation steps as described in the above embodiments or flowcharts, more or fewer operation steps can be included in the methods based on convention or without creative labor. In steps without necessary causality in logic, the execution order of the steps is not limited to the execution order provided by the embodiments of the present specification. When the methods are executed by actual devices or terminal products, the methods can be executed in sequence or in parallel (for example, in a parallel processor or a multi-thread processing environment) according to the method sequence shown in the embodiments or the drawings.

[0093] It should be understood that the above description is intended for illustration only and not for limitation. Many implementations and many applications other than the examples provided would be apparent to those skilled in the art from the above description. The scope of the embodiments of the present specification should therefore not be determined with reference to the above description, but instead should be determined with reference to the appended claims along with their full scope of equivalents.

[0094] The above merely provides preferred embodiments of the present application, and is not intended to limit the present application. Based on the above teachings, one of ordinary skill in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Therefore, the general principles defined herein are intended to be broad in scope and claims are intended to encompass all such changes and variations.

Claims

1. A method of determining creditworthiness, characterized by, The method comprises the following steps: acquiring an initial rough set, discretizing the initial rough set to obtain a first rough set, wherein the initial rough set contains a plurality of original feature information of an institution; determining redundant features in the first rough set by using a target analysis algorithm, wherein the target analysis algorithm comprises histogram distribution analysis, similarity analysis or principal component analysis; removing or merging the redundant features in the first rough set to obtain a second rough set; performing data cleaning on the second rough set to obtain a target rough set; performing information reduction on the target rough set by using a discernible matrix reduction algorithm to obtain an evidence information set, wherein each feature information after information reduction is used as a separate evidence; obtaining a basic probability assignment value of a recognition framework corresponding to each evidence in the evidence information set by using a statistical evidence; obtaining a representation value of each institution by using evidence theory for uncertain reasoning according to the basic probability assignment value of the recognition framework corresponding to each evidence, wherein the representation value is used to represent the credit degree of the institution. The evidence theory comprises a fusion rule, and the fusion rule comprises the following steps: taking two evidences of any institution, and judging whether a conflict factor of the two evidences is greater than a preset threshold value; if the conflict factor of the two evidences is less than or equal to the preset threshold value, fusing the two evidences, and fusing a new evidence after fusion with other evidences of the institution; if the conflict factor of the two evidences is greater than the preset threshold value, taking another evidence, and fusing one of the two evidences with the another evidence; repeating the step of fusing until determining whether a conflict factor of the last two evidences is greater than the preset threshold value when fusing the last two evidences of the institution; if no, fusing the last two evidences to obtain an evidence fusion result of the institution; if yes, fusing the last two evidences by using a mean weighted method, wherein the step of fusing the last two evidences by using the mean weighted method comprises the following steps: performing conflict resolution on the evidences producing paradoxes by using the mean weighted method according to the following formula: wherein, is the basic probability assignment of the fused new evidence; m1 is evidence 1; m2 is evidence 2; m1(A) is the basic probability assignment of evidence 1; and m2(A) is the basic probability assignment of evidence 2.

2. The method of claim 1, wherein, obtaining a representation value of each institution by using evidence theory for uncertain reasoning according to the basic probability assignment value of the recognition framework corresponding to each evidence, comprising the following steps: fusing the basic probability assignment value of the recognition framework corresponding to each evidence of each institution by using evidence theory to obtain an evidence fusion result of each institution; obtaining a representation value of each institution according to the evidence fusion result of each institution.

3. The method of claim 1, wherein, The feature information comprises enterprise registration supplementary information, enterprise registration change information, shareholder registration information, unit insurance participation information, enterprise legal person, enterprise financial data, enterprise tax information, enterprise tax registration information and enterprise provident fund payment information.

4. An apparatus for determining creditworthiness, characterized by The method comprises the following steps: The first obtaining module is configured to obtain an initial rough set, discretize the initial rough set, and obtain a first rough set; the initial rough set contains a plurality of original feature information of an organization; a target analysis algorithm is used to determine redundant features in the first rough set; the target analysis algorithm includes histogram distribution analysis, similarity analysis, or principal component analysis; the redundant features in the first rough set are removed or merged to obtain a second rough set; the second rough set is cleaned to obtain a target rough set; The information reduction module is configured to use a discernible matrix reduction algorithm to reduce information of the target rough set to obtain an evidence information set; each feature information after information reduction is used as a separate evidence; The second obtaining module is configured to use statistical evidence to obtain a probability distribution value, and obtain a basic probability distribution value of an identification framework corresponding to each evidence in the evidence information set; The processing module is configured to use evidence theory to perform uncertain reasoning to obtain a representation value of each organization according to the basic probability distribution value of the identification framework corresponding to each evidence; the representation value is used to represent a credit degree of the organization; The evidence theory includes a fusion rule, and the fusion rule includes: Two evidences of any organization are taken, and it is determined whether a conflict factor of the two evidences is greater than a preset threshold; If the conflict factor of the two evidences is less than or equal to the preset threshold, the two evidences are fused, and a new evidence after fusion is fused with other evidences of the organization; If the conflict factor of the two evidences is greater than the preset threshold, another evidence is taken, and one of the two evidences is fused with the other evidence; The fusion step is repeated until it is determined whether a conflict factor of the last two evidences of the organization is greater than the preset threshold when the last two evidences are fused; If not, the last two evidences are fused to obtain an evidence fusion result of the organization; If yes, the last two evidences are fused by using a mean weighted method, and the mean weighted method includes: According to the following formula, the mean weighted method is used to resolve the conflict of the evidences that produce paradoxes: wherein, is the basic probability assignment of the fused new evidence; m1 is evidence 1; m2 is evidence 2; m1(A) is the basic probability assignment of evidence 1; and m2(A) is the basic probability assignment of evidence 2.

5. A credit degree determining apparatus characterized by comprising: A processor and a memory for storing processor-executable instructions are included, and the processor executes the instructions to implement the steps of the method in any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that, A computer instruction is stored thereon, and the instruction is executed to implement the steps of the method in any one of claims 1 to 3.

Images