CN112529681A - Credit risk transfer method based on credit subject correlation strength - Google Patents
Credit risk transfer method based on credit subject correlation strength Download PDFInfo
- Publication number
- CN112529681A CN112529681A CN202110180079.0A CN202110180079A CN112529681A CN 112529681 A CN112529681 A CN 112529681A CN 202110180079 A CN202110180079 A CN 202110180079A CN 112529681 A CN112529681 A CN 112529681A
- Authority
- CN
- China
- Prior art keywords
- credit
- layer
- training
- method based
- association strength
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Databases & Information Systems (AREA)
- Animal Behavior & Ethology (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Finance (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- General Business, Economics & Management (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a credit risk transfer method based on the association strength of credit main bodies, which comprises the steps of constructing a credit relation network among the credit main bodies by using a credit knowledge mapping technology; searching one or more nodes with the most influence as key points according to the ant lion algorithm, and constructing a credit transfer path of a credit subject based on the association strength; when the credit body changes, searching the credit transmission path through a depth-first algorithm, and calculating the influence degree of the relevant nodes by the current body; calculating to obtain the influence degree after superposition by utilizing an irrelevance strategy, obtaining a plurality of important key paths, and constructing a directed acyclic graph; based on the directed acyclic graph, a heuristic method optimizes credit transfer time, transfers credit risks step by step and reduces the influence degree of the credit risks on relevant mechanisms. The invention transfers the credit risk step by step, and reduces the influence degree of the credit risk on related mechanisms.
Description
Technical Field
The invention relates to the technical field of credit risk transfer calculation, in particular to a credit risk transfer method based on credit principal association strength.
Background
The credit risk is not isolated but has a transfer effect, the transfer of the credit risk is the effect of combining macroscopic economy and microscopic economy, and the macroscopic angle refers to the total factors of slowing down the growth of the macroscopic economy, reducing the market investment risk preference, tightening up the financing environment and the like; the transfer of credit risk is from a microscopic point of view, and mainly refers to the credit risk transfer effect caused by a series of microscopic economic connections of production, investment, financing and other activities between enterprises.
Just because there are the complicated and intricacies between the credit main bodies in social production life, the transmission of credit risk in the formed social network is multipath, and at present, there is no effective method to solve the problem of transmission of credit risk and reduce the influence degree of credit risk to relevant organizations.
Disclosure of Invention
This section is for the purpose of summarizing some aspects of embodiments of the invention and to briefly introduce some preferred embodiments. In this section, as well as in the abstract and the title of the invention of this application, simplifications or omissions may be made to avoid obscuring the purpose of the section, the abstract and the title, and such simplifications or omissions are not intended to limit the scope of the invention.
The present invention has been made in view of the above-mentioned conventional problems.
Therefore, the invention provides a credit risk transfer method based on the association strength of credit subjects, which can reduce the influence degree of credit risk on relevant mechanisms.
In order to solve the technical problems, the invention provides the following technical scheme: constructing a credit relation network between credit subjects by using a credit knowledge graph technology; searching one or more nodes with the most influence as key points according to the ant lion algorithm, and constructing a credit transfer path of a credit subject based on the association strength; when the credit body changes, searching the credit transmission path through a depth-first algorithm, and calculating the influence degree of the relevant nodes by the current body; calculating to obtain the influence degree after superposition by utilizing an irrelevance strategy, obtaining a plurality of important key paths, and constructing a directed acyclic graph; and calculating a credit transfer critical path with minimized time based on the directed acyclic graph, transferring the credit risk step by step, and reducing the influence degree of the credit risk on related mechanisms.
As a preferable scheme of the credit risk transferring method based on the association strength of the credit body, the method comprises the following steps: constructing the credit relationship network comprises the steps of preprocessing collected credit subject data, identifying entities, extracting relationships, aligning the entities and generating a knowledge graph; converting the data into texts by using an identification technology, and then converting the texts into knowledge; for the structured data stored in the relational database, directly converting and extracting entities and relations through D2R; and generating user-defined word segmentation for the text generated by the unstructured database according to the professional corpus of the credit subject, and performing word segmentation on the text.
As a preferable scheme of the credit risk transferring method based on the association strength of the credit body, the method comprises the following steps: the method also comprises the steps of utilizing the CRF + bidirectional LSTM to carry out entity identification, and carrying out relation extraction according to the SDP + LSTM to form preliminary knowledge; carrying out disambiguation, synonym loading and changing processing and coreference resolution on the relation between the entities; and forming correct knowledge after quality evaluation and storing the correct knowledge into the knowledge graph to generate the credit relationship network.
As a preferable scheme of the credit risk transferring method based on the association strength of the credit body, the method comprises the following steps: obtaining the credit delivery path may include obtaining the credit delivery path,
wherein N is the number of measurement points, L is the number of sensors, i.e. APiI =1, 2, …, L, when credit risk transfer occurs at RPjAt point in time, sensor APiMeasured associated signal strength of
And τ is the number of measurements.
As a preferable scheme of the credit risk transferring method based on the association strength of the credit body, the method comprises the following steps: calculating the influence degree, wherein a deep learning model is required to be constructed, and the deep learning model comprises a CNN layer and a coding layer; the CNN layer comprises a convolution layer, a pooling layer and a full-connection layer; the coding layer comprises a convolution coding layer and a convolution decoding layer.
As a preferable scheme of the credit risk transferring method based on the association strength of the credit body, the method comprises the following steps: the deep learning model needs to be trained, a training set is input into the deep learning model, and the coding layer carries out unsupervised pre-training on the training set to obtain the preliminary characteristics of the training set; initializing the convolutional layer by using the extracted features to obtain a training sample label; the CNN layer sequentially performs convolution, pooling and full-connection processing on the training set after pre-training; the coding layer carries out convolutional coding and decoding processing on the training set subjected to convolutional processing, and output data are fed back to the convolutional layer; and after finishing the full connection processing, the CNN layer outputs a recognition result and judges whether the recognition result is consistent with the training sample label.
As a preferable scheme of the credit risk transferring method based on the association strength of the credit body, the method comprises the following steps: if the recognition result is consistent with the training sample label, stopping iterative training, and finishing the deep learning model training; and if the recognition result is inconsistent with the training sample label, continuing iterative training until the judgment result is consistent, stopping training, and outputting the deep learning model.
As a preferable scheme of the credit risk transferring method based on the association strength of the credit body, the method comprises the following steps: constructing the directed acyclic graph, including analyzing the key paths according to the sequence of influence degree from high to low; and adding the key points in the analyzed key path to obtain the credit transfer key path.
The invention has the beneficial effects that: the method constructs a structure diagram through a knowledge map technology, finds key points and key paths, calculates the influence degree, optimizes credit transfer time, transfers credit risks step by step and reduces the influence degree of the credit risks to relevant mechanisms.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise. Wherein:
FIG. 1 is a flowchart illustrating a credit risk delivery method based on the association strength of credit body according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating an influence weight of a credit risk transfer method based on the association strength of a credit subject according to an embodiment of the present invention;
fig. 3 is a schematic diagram of node collapse by selecting a node with the largest influence by using a crawler algorithm in combination with a depth-first search and a pruning strategy according to the credit risk delivery method based on credit subject association strength according to an embodiment of the present invention;
fig. 4 is a schematic diagram illustrating an output of an experimental comparison curve of a credit risk delivery method based on the association strength of a credit subject according to an embodiment of the present invention.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, specific embodiments accompanied with figures are described in detail below, and it is apparent that the described embodiments are a part of the embodiments of the present invention, not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making creative efforts based on the embodiments of the present invention, shall fall within the protection scope of the present invention.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways than those specifically described and will be readily apparent to those of ordinary skill in the art without departing from the spirit of the present invention, and therefore the present invention is not limited to the specific embodiments disclosed below.
Furthermore, reference herein to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one implementation of the invention. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.
The present invention will be described in detail with reference to the drawings, wherein the cross-sectional views illustrating the structure of the device are not enlarged partially in general scale for convenience of illustration, and the drawings are only exemplary and should not be construed as limiting the scope of the present invention. In addition, the three-dimensional dimensions of length, width and depth should be included in the actual fabrication.
Meanwhile, in the description of the present invention, it should be noted that the terms "upper, lower, inner and outer" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of describing the present invention and simplifying the description, but do not indicate or imply that the referred device or element must have a specific orientation, be constructed in a specific orientation and operate, and thus, cannot be construed as limiting the present invention. Furthermore, the terms first, second, or third are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
The terms "mounted, connected and connected" in the present invention are to be understood broadly, unless otherwise explicitly specified or limited, for example: can be fixedly connected, detachably connected or integrally connected; they may be mechanically, electrically, or directly connected, or indirectly connected through intervening media, or may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
Example 1
Referring to fig. 1 to 4, a first embodiment of the present invention provides a credit risk delivery method based on the association strength of a credit principal, including:
s1: and constructing a credit relation network between credit subjects by using a credit knowledge graph technology. It should be noted that constructing the credit relationship network includes:
preprocessing collected credit subject data, identifying entities, extracting relationships, aligning the entities and generating a knowledge graph;
converting the data into texts by using an identification technology, and then converting the texts into knowledge;
for the structured data stored in the relational database, directly converting and extracting entities and relations through D2R;
generating self-defined word segmentation for the text generated by the unstructured database according to the professional corpus of credit subjects, and performing word segmentation on the text;
performing entity identification by using the CRF + bidirectional LSTM, and performing relation extraction according to the SDP + LSTM to form preliminary knowledge;
carrying out disambiguation, synonym loading and changing processing and coreference resolution on the relation between the entities;
and after quality evaluation is carried out, correct knowledge is formed and stored in a knowledge graph to generate a credit relationship network.
S2: and searching one or more most influential nodes as key points according to the ant lion algorithm, and constructing a credit transfer path of the credit subject based on the association strength. It should be noted that, obtaining the credit transfer path includes:
wherein N is the number of measurement points, L is the number of sensors, i.e. APiI =1, 2, …, L, when credit risk transfer occurs at RPjAt point in time, sensor APiMeasured associated signal strength of
And τ is the number of measurements.
S3: when the credit body changes, the credit transmission path is searched through a depth-first algorithm, and the influence degree of the relevant nodes by the current body is calculated. It should be further noted that, the calculation of the influence degree requires the construction of a deep learning model, which includes:
a CNN layer and a coding layer;
the CNN layer comprises a convolution layer, a pooling layer and a full-connection layer;
the encoding layer includes a convolutional encoding layer and a convolutional decoding layer.
Further, the deep learning model needs to be trained, including:
inputting the training set into a deep learning model, and carrying out unsupervised pre-training on the training set by a coding layer to obtain the preliminary characteristics of the training set;
initializing the convolutional layer by using the extracted features to obtain a training sample label;
the CNN layer sequentially performs convolution, pooling and full-connection processing on the pre-trained training set;
the coding layer carries out convolutional coding and decoding processing on the training set subjected to the convolutional processing, and output data are fed back to the convolutional layer;
after the CNN layer finishes the full connection processing, outputting an identification result and judging whether the identification result is consistent with a training sample label;
if the recognition result is consistent with the training sample label, stopping iterative training, and finishing deep learning model training;
and if the recognition result is inconsistent with the label of the training sample, continuing the iterative training until the judgment result is consistent, stopping the training, and outputting a deep learning model.
S4: and calculating to obtain the influence degree after superposition by using an irrelevance strategy, obtaining a plurality of important key paths, and constructing a directed acyclic graph. Is provided with
For the degree of incidence of the node i, the influence degree of the multipoint multipath composite on the node i is defined as:
it should be further noted that in this step, constructing a directed acyclic graph includes:
analyzing the critical path according to the sequence of influence degree from high to low;
and adding key points in the analyzed key path to obtain a credit transfer key path.
S5: based on the directed acyclic graph, the heuristic method optimizes the credit transfer time, transfers the credit risk step by step and reduces the influence degree of the credit risk on relevant mechanisms.
Preferably, in order to facilitate understanding of technical difficulties solved by the technical solution by non-skilled persons, the present embodiment makes the following detailed description:
(1) a description of a credit risk delivery issue;
in a credit principal relationship network (directed acyclic graph) G (V, E), a set of nodes in V generation, each node represents a credit individual or credit organization, E represents a set of edges, each edge represents the relationship (mother-child relationship, equity relationship, supply relationship and the like) among credit principal bodies, each credit principal body calculates and obtains the credit value of the credit principal body according to a credit evaluation model, the existing credit related field empirical evidence research shows that the transmission of credit risk of the credit principal bodies is mainly related to the association density of the credit principal relationship network and the association degree among the credit principal bodies, and for a certain node i in V(some credit principal), suppose
In order to achieve the goal of the method,
for the reason of the out-degree, the credit changes due to the economic macro or micro factors and the subjective or objective factors of the main body, and the adjacent associated nodes are subjected to the change
The weight of the influence is
，
，
The larger the value, the higher the degree of influence; on the other hand, paths formed by the nodes
Associated node of
The degree of influence of the node i gradually decreases, and the calculation of the influence degree of the multipoint multipath composite on a certain node j in G (V, E) is defined as:
wherein, each neighbor node of the node j has an accumulative influence on the node j, and when the influence is accumulated
Exceeds a certain threshold
(obtained by the deep learning method described above), the credit for node j may change.
Referring to fig. 2, in the network with influence weight, node 3 is influenced by nodes 1 and 2, and is defined according to the calculation of the influence degree of multipoint multipath composite on a certain node j in G (V, E)
。
(2) Searching for a technical route with respect to a credit principal associated network influence maximization node;
the searching strategy of the influence maximization node of the credit subject associated network is divided into two steps: firstly, a heuristic method obtains an initial node set with potential influence
(ii) a Secondly, selecting the node with the largest influence by adopting a crawler algorithm in combination with a depth-first search and pruning strategy.
Firstly, a heuristic method obtains an initial node set with potential influence
;
Out degree of node i
Reflecting the position and the influence of the credit subject in the credit subject gateway network, the larger the out degree is, the more nodes are directly influenced by the out degree; degree of entry of node i
Reflecting the tightness degree of the connection of the nodes with other people, the higher the degree of entrance, the more easily influenced by other nodes, the degree of entrance and the degree of exit reflect the structural complexity of the credit subject gateway network, the degree centrality of the nodes, and the heuristic method is considered to obtain the initial node set with potential influence
The method comprises the following steps:
the global impact factor for node i is defined as:
，
,
the algorithm framework of the heuristic method is as follows:
inputting: g (V, E), node set
，
，
，
And (3) outputting: initial set of nodes with potential impact
1、
2、for each i，
3. Calculating its global influence factor
4、endfor
5. To each node according to
In descending order of
6. Selecting the first a% of nodes from the obtained mixture and adding the selected nodes into the obtained mixture
In (1) obtaining
。
Secondly, selecting a node with the largest influence by adopting a crawler algorithm in combination with a depth-first search and pruning strategy;
node set with potential influence based on initial
For each node i in the set, if its out degree is
Is provided with
Only the initial crawler is used, the node i is used as an initial node, depth-first search is adopted, and the crawler searches related nodes on related paths in sequence
Calculating and storing the influence degree of each node j under the single path of the node i in the graph
And simultaneously calculating the composite influence degree of each crawler on the node j under the multipath
。
Referring to fig. 3, in order to avoid the repeated calculation and increase the searching speed, if the affected degree of the node j is transmitted through a certain path
Then, using pruning strategy, delete all its incoming edges and update the set
，
The influence of the nodes is gradually weakened in the process of path transmission, when a certain credit subject is impacted by a plurality of nodes in the network, collapse can be generated, the composite influence degree of the node j is 0.75, the credit risk is high, and therefore 5 layers of exploration is considered in depth-first search, the process is repeated until no new node is added to the depth-first search
In (1).
The algorithm framework for selecting the node with the largest influence is as follows:
inputting: g (V, E), initial node set
，
And (3) outputting: node set with maximum influence
1、 for each i，
3. According to its degree of play
Is provided with a plurality of reptiles
4. Calculating the complexity of each node j influenced by i by adopting a depth-first search strategy
5. Any node j in for search path
6、 if 
7. Deleting all incoming edges of j by adopting pruning algorithm
8、 
9、 endif
10、endfor
11. After updating
Repeating the steps until no new node is added.
Preferably, in order to better verify and explain the technical effects adopted in the method of the present invention, the embodiment selects to perform a comparative test with the conventional credit association transfer method and the method of the present invention, and compares the test results with a scientific demonstration means to verify the actual effects of the method of the present invention.
In order to verify that the method of the present invention has higher influence elimination performance and transmission safety compared with the conventional method, the present embodiment respectively performs real-time measurement and comparison on the simulated credit system by using the conventional method and the method of the present invention.
And (3) testing environment: (1) inputting the running state quantity of the credit main body into PSAPAC software for simulation running, and respectively simulating the running of credit risk transmission;
(2) taking the historical running state detection cases from 1 month to 12 months in 2019 as experimental data, selecting 100 groups for normalization processing and training, and finally determining 10 groups of data with higher training degree and better normalization as a test set;
(3) the experimental data are analyzed by two methods respectively, 10 groups of data are tested by each method, the state identification quantity of each group of data is calculated, and the error is calculated by comparing with the state quantity input by simulation.
Referring to fig. 4, a schematic diagram of a final result output curve of experimental comparison is shown, in which a solid line is a curve output by the present invention, and a dotted line is a curve output by a conventional method, according to the schematic diagram of fig. 4, it can be intuitively seen that the solid line and the dotted line gradually pull apart with increasing time, the dotted line always presents an unstable fluctuation trend and always has a lower safety increase than the solid line, although the solid line slightly fluctuates, the solid line basically tends to be stable, and the dotted line is always kept at a rising position of the dotted line, thereby verifying the real effect of the method of the present invention.
It should be noted that the above-mentioned embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, which should be covered by the claims of the present invention.
Claims (8)
1. A credit risk transfer method based on the association strength of credit body is characterized in that: comprises the steps of (a) preparing a mixture of a plurality of raw materials,
constructing a credit relation network between credit subjects by using a credit knowledge graph technology;
searching one or more nodes with the most influence as key points according to the ant lion algorithm, and constructing a credit transfer path of a credit subject based on the association strength;
when the credit body changes, searching the credit transmission path through a depth-first algorithm, and calculating the influence degree of the relevant nodes by the current body;
calculating to obtain the influence degree after superposition by utilizing an irrelevance strategy, obtaining a plurality of important key paths, and constructing a directed acyclic graph;
and calculating a credit transfer critical path with minimized time based on the directed acyclic graph, transferring the credit risk step by step, and reducing the influence degree of the credit risk on related mechanisms.
2. The credit risk delivery method based on credit body association strength as claimed in claim 1, wherein: constructing the network of credit relationships includes,
preprocessing collected credit subject data, identifying entities, extracting relationships, aligning the entities and generating a knowledge graph;
converting the data into texts by using an identification technology, and then converting the texts into knowledge;
for the structured data stored in the relational database, directly converting and extracting entities and relations through D2R;
and generating user-defined word segmentation for the text generated by the unstructured database according to the professional corpus of the credit subject, and performing word segmentation on the text.
3. The credit risk delivery method based on credit body association strength as claimed in claim 2, wherein: also comprises the following steps of (1) preparing,
performing entity identification by using the CRF + bidirectional LSTM, and performing relation extraction according to the SDP + LSTM to form preliminary knowledge;
carrying out disambiguation, synonym loading and changing processing and coreference resolution on the relation between the entities;
and forming correct knowledge after quality evaluation and storing the correct knowledge into the knowledge graph to generate the credit relationship network.
4. A credit risk delivery method based on credit body association strength according to claim 1 or 3, characterized in that: obtaining the credit delivery path may include obtaining the credit delivery path,
wherein N is the number of measurement points, L is the number of sensors, i.e. APiI =1, 2, …, L, when credit risk transfer occurs at RPjAt point in time, sensor APiMeasured associated signal strength of
And τ is the number of measurements.
5. The credit risk delivery method based on credit body association strength as claimed in claim 4, wherein: calculating the influence degree, wherein a deep learning model is required to be constructed, and the deep learning model comprises a CNN layer and a coding layer;
the CNN layer comprises a convolution layer, a pooling layer and a full-connection layer;
the coding layer comprises a convolution coding layer and a convolution decoding layer.
6. The credit risk delivery method based on credit body association strength as claimed in claim 5, wherein: the deep learning model is required to be trained, including,
inputting a training set into the deep learning model, and carrying out unsupervised pre-training on the training set by the coding layer to obtain the preliminary characteristics of the training set;
initializing the convolutional layer by using the extracted features to obtain a training sample label;
the CNN layer sequentially performs convolution, pooling and full-connection processing on the training set after pre-training;
the coding layer carries out convolutional coding and decoding processing on the training set subjected to convolutional processing, and output data are fed back to the convolutional layer;
and after finishing the full connection processing, the CNN layer outputs a recognition result and judges whether the recognition result is consistent with the training sample label.
7. The credit risk delivery method based on credit body association strength as claimed in claim 6, wherein: also comprises the following steps of (1) preparing,
if the recognition result is consistent with the training sample label, stopping iterative training, and finishing the deep learning model training;
and if the recognition result is inconsistent with the training sample label, continuing iterative training until the judgment result is consistent, stopping training, and outputting the deep learning model.
8. The credit risk delivery method based on credit body association strength as claimed in claim 7, wherein: constructing the directed acyclic graph, including,
analyzing the critical path according to the sequence of influence degree from high to low;
and adding the key points in the analyzed key path to obtain the credit transfer key path.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110180079.0A CN112529681B (en) | 2021-02-08 | 2021-02-08 | Credit risk transfer method based on credit subject correlation strength |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110180079.0A CN112529681B (en) | 2021-02-08 | 2021-02-08 | Credit risk transfer method based on credit subject correlation strength |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112529681A true CN112529681A (en) | 2021-03-19 |
| CN112529681B CN112529681B (en) | 2021-05-07 |
Family
ID=74975593
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110180079.0A Active CN112529681B (en) | 2021-02-08 | 2021-02-08 | Credit risk transfer method based on credit subject correlation strength |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112529681B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112927082A (en) * | 2021-03-22 | 2021-06-08 | 中国工商银行股份有限公司 | Credit risk prediction method, apparatus, device, medium, and program product |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109598509A (en) * | 2018-10-17 | 2019-04-09 | 阿里巴巴集团控股有限公司 | The recognition methods of risk clique and device |
| CN109858740A (en) * | 2018-12-21 | 2019-06-07 | 中化资本有限公司 | Appraisal procedure, device, computer equipment and the storage medium of business risk |
| CN111383102A (en) * | 2020-03-27 | 2020-07-07 | 北京明略软件***有限公司 | Financial credit risk identification method, model construction method and device |
| CN111652704A (en) * | 2020-06-09 | 2020-09-11 | 唐松 | Financial credit risk assessment method based on knowledge graph and graph deep learning |
| CN112214614A (en) * | 2020-10-16 | 2021-01-12 | 民生科技有限责任公司 | Method and system for mining risk propagation path based on knowledge graph |
-
2021
- 2021-02-08 CN CN202110180079.0A patent/CN112529681B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109598509A (en) * | 2018-10-17 | 2019-04-09 | 阿里巴巴集团控股有限公司 | The recognition methods of risk clique and device |
| CN109858740A (en) * | 2018-12-21 | 2019-06-07 | 中化资本有限公司 | Appraisal procedure, device, computer equipment and the storage medium of business risk |
| CN111383102A (en) * | 2020-03-27 | 2020-07-07 | 北京明略软件***有限公司 | Financial credit risk identification method, model construction method and device |
| CN111652704A (en) * | 2020-06-09 | 2020-09-11 | 唐松 | Financial credit risk assessment method based on knowledge graph and graph deep learning |
| CN112214614A (en) * | 2020-10-16 | 2021-01-12 | 民生科技有限责任公司 | Method and system for mining risk propagation path based on knowledge graph |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112927082A (en) * | 2021-03-22 | 2021-06-08 | 中国工商银行股份有限公司 | Credit risk prediction method, apparatus, device, medium, and program product |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112529681B (en) | 2021-05-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110516085B (en) | Image text mutual retrieval method based on bidirectional attention | |
| CN106777274B (en) | A kind of Chinese tour field knowledge mapping construction method and system | |
| CN111931506B (en) | Entity relationship extraction method based on graph information enhancement | |
| CN109299262A (en) | A kind of text implication relation recognition methods for merging more granular informations | |
| CN109635291A (en) | A kind of recommended method of fusion score information and item contents based on coorinated training | |
| CN113704546B (en) | Video natural language text retrieval method based on space time sequence characteristics | |
| CN110825881A (en) | Method for establishing electric power knowledge graph | |
| CN106354710A (en) | Neural network relation extracting method | |
| CN108563703A (en) | A kind of determination method of charge, device and computer equipment, storage medium | |
| CN108630199A (en) | A kind of data processing method of acoustic model | |
| CN113505200B (en) | Sentence-level Chinese event detection method combined with document key information | |
| CN111651558B (en) | Hyperspherical surface cooperative measurement recommendation device and method based on pre-training semantic model | |
| CN106970910A (en) | A kind of keyword extracting method and device based on graph model | |
| CN109857457B (en) | Function level embedding representation method in source code learning in hyperbolic space | |
| CN112988917B (en) | Entity alignment method based on multiple entity contexts | |
| CN106649272A (en) | Named entity recognizing method based on mixed model | |
| CN106598950A (en) | Method for recognizing named entity based on mixing stacking model | |
| CN110232395A (en) | A kind of fault diagnosis method of electric power system based on failure Chinese text | |
| CN105868187B (en) | The construction method of more translation Parallel Corpus | |
| CN113239148B (en) | Scientific and technological resource retrieval method based on machine reading understanding | |
| CN114048286B (en) | Automatic fact verification method integrating graph converter and common attention network | |
| CN112966091A (en) | Knowledge graph recommendation system fusing entity information and heat | |
| CN110990718A (en) | Social network model building module of company image improving system | |
| CN107092605A (en) | A kind of entity link method and device | |
| CN112687374A (en) | Psychological crisis early warning method based on text and image information joint calculation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |