CN112488820A - Model training method and default prediction method based on noctilucent remote sensing data - Google Patents
Model training method and default prediction method based on noctilucent remote sensing data Download PDFInfo
- Publication number
- CN112488820A CN112488820A CN202011300431.1A CN202011300431A CN112488820A CN 112488820 A CN112488820 A CN 112488820A CN 202011300431 A CN202011300431 A CN 202011300431A CN 112488820 A CN112488820 A CN 112488820A
- Authority
- CN
- China
- Prior art keywords
- time window
- target area
- noctilucent
- remote sensing
- default
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 75
- 238000012549 training Methods 0.000 title claims abstract description 49
- 238000012502 risk assessment Methods 0.000 claims abstract description 15
- 238000003860 storage Methods 0.000 claims description 10
- 238000012417 linear regression Methods 0.000 claims description 6
- 238000007637 random forest analysis Methods 0.000 claims description 6
- 238000013135 deep learning Methods 0.000 claims description 5
- 238000005516 engineering process Methods 0.000 abstract description 3
- 230000000694 effects Effects 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000036541 health Effects 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000012954 risk control Methods 0.000 description 3
- 238000004590 computer program Methods 0.000 description 2
- 238000005034 decoration Methods 0.000 description 2
- 235000019800 disodium phosphate Nutrition 0.000 description 2
- 238000012854 evaluation process Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000003331 infrared imaging Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000003915 air pollution Methods 0.000 description 1
- 239000005441 aurora Substances 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- -1 lightning Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000013058 risk prediction model Methods 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
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
- G06Q40/03—Credit; Loans; Processing thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Finance (AREA)
- Medical Informatics (AREA)
- Mathematical Physics (AREA)
- Accounting & Taxation (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Technology Law (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application provides a model training method and default prediction method based on noctilucent remote sensing data, which are applied to the technical field of financial science and technology, wherein the device comprises the following steps: determining noctilucence remote sensing data of a target region preset time window, and determining noctilucence derivative characteristic indexes of the target region preset time window based on the noctilucence remote sensing data of the preset time window; and determining default probability through a default prediction model based on the noctilucence derivative characteristic indexes of the preset time window of the target area, and taking the default probability as a macroscopic credit risk assessment result. The assessment of the macroscopic credit risk is carried out according to the noctilucent remote sensing data, and the hysteresis of the assessment of the macroscopic credit risk based on official statistical data is avoided, so that the timeliness of the assessment of the macroscopic credit risk is improved.
Description
Technical Field
The application relates to the technical field of financial science and technology, in particular to a model training method and a default prediction method based on noctilucent remote sensing data.
Background
In the credit risk evaluation process under the credit scene, the business health condition of an enterprise or the stability of the individual economic income is often the main factor influencing the credit risk, but the business health condition or the stability of the individual economic income is greatly influenced by the external macroscopic environment, and particularly, the business health condition or the stability of the individual economic income is obviously different in different regions and industries, so that the current macroscopic credit risk of a client needs to be accurately evaluated. In the past, the consideration of the current macroscopic economic state of a client is mainly based on traditional official statistical data, but the traditional official statistical data does not measure the macroscopic economic activities very accurately due to high data statistical difficulty, high cost and long period, and particularly for developing countries, due to the fact that government statistical mechanisms are imperfect, statistical calibers in different regions are inconsistent, a large proportion of economic activities are not brought into statistical categories, and statistical results such as domestic price indexes and actual domestic total production values of the countries have large measurement errors. And official published data still has the poor problem of timeliness, when the macroscopic environment worsens, because can not in time obtain reliable information, often can cause the increase of credit product whole default rate, bring huge loss for financial institution.
With the improvement of remote sensing technology and the continuous improvement of image resolution, satellite remote sensing data is widely proved to have high correlation with the macroscopic economy of human society. The remote sensing data has the advantages of space-time continuity, independence, objectivity and the like, is used for monitoring regional forest coverage, air pollution, infrastructure investment and the like, and becomes one of common data sources for measuring the macroscopic economic state. How to use the noctilucent remote sensing data to serve credit scenes is a main problem to be solved by the method, and the method can quickly and accurately make intelligent response to external environment changes in the macroscopic credit risk evaluation process and improve the accuracy and stability of the overall macroscopic credit evaluation system.
Disclosure of Invention
The application provides a model training method and a default prediction method based on noctilucent remote sensing data, which are used for solving the problem of poor timeliness of predicting macroscopic economic activities based on official statistical data and improving the timeliness of the predicted macroscopic economic activities. The technical scheme adopted by the application is as follows:
in a first aspect, a default prediction model training method based on noctilucent remote sensing data is provided, and comprises the following steps:
determining historical noctilucent remote sensing data of each time window of a target area, and determining noctilucent derivative characteristic indexes of each time window based on the historical noctilucent remote sensing data of each time window;
determining default rates of all time windows of a target area;
and training a default prediction model based on the default rate of each time window of the target area and the noctilucence derivative characteristic indexes of each time window of the target area.
Optionally, determining a default rate of each time window of the target area includes:
acquiring debt data of a plurality of target objects in a preset time period of a target area;
and determining default rates of the time windows based on the debt item data of the target objects in the target area within the preset time period.
Optionally, the method further comprises:
acquiring boundary vector data of a plurality of target areas and integral noctilucent remote sensing data;
and obtaining the noctilucent remote sensing data of each target area based on the boundary vector data of each target area and the whole noctilucent remote sensing data.
Optionally, the noctilucent derivative characteristic index comprises a noctilucent total intensity index, a Moran index, a week-to-weekend noctilucent intensity contrast index, and a first and second half noctilucent intensity contrast index.
Optionally, the method further comprises:
determining the proportional increase of the noctilucent derivative characteristic index;
training a default prediction model based on default rates of all time windows of a target area and noctilucence derivative characteristic indexes of all time windows of the target area, wherein the training of the default prediction model comprises the following steps:
and training a default prediction model based on the default rate of each time window of the target area and the comparably increasing of the noctilucent derivative characteristic indexes of each time window of the target area.
Optionally, training the default prediction model based on the default rate of each time window of the target area and the proportional increase of the noctilucent derivative characteristic indicator of each time window of the target area includes:
and training the default prediction model by using a gradient lifting method, a linear regression method, a ridge regression method, an SVM method, a random forest method and a deep learning method.
In a second aspect, an embodiment of the present application provides a default prediction method based on noctilucent remote sensing data, including:
determining noctilucence remote sensing data of a target region preset time window, and determining noctilucence derivative characteristic indexes of the target region preset time window based on the noctilucence remote sensing data of the preset time window;
and determining the default probability through the default prediction model of any one of the first aspect based on the noctilucent derivative characteristic indexes of the predetermined time window of the target area, and taking the default probability as a macroscopic credit risk assessment result.
Optionally, the method further comprises:
acquiring boundary vector data of a plurality of target areas and integral noctilucent remote sensing data;
and obtaining the noctilucent remote sensing data of each target area based on the boundary vector data of each target area and the whole noctilucent remote sensing data.
Optionally, the noctilucent derivative characteristic index comprises a noctilucent total intensity index, a Moran index, a week-to-weekend noctilucent intensity contrast index, and a first and second half noctilucent intensity contrast index.
Optionally, the method further comprises:
determining the proportional increase of the noctilucent derivative characteristic index;
determining the default probability based on the noctilucence derivative characteristic indexes of the predetermined time window of the target area through the default prediction model of any one of the first aspect, including:
and determining the default probability by any default prediction model of the first aspect based on the comparably increasing noctilucent derived characteristic indexes of the target area in the predetermined time window.
In a third aspect, an embodiment of the present application provides a default prediction model training device based on noctilucent remote sensing data, including:
the first determination module is used for determining historical noctilucent remote sensing data of each time window of the target area and determining noctilucent derivative characteristic indexes of each time window based on the historical noctilucent remote sensing data of each time window;
the second determining module is used for determining the default rate of each time window of the target area;
and the training module is used for training the default prediction model based on the default rate of each time window of the target area and the noctilucence derivative characteristic indexes of each time window of the target area.
Optionally, the second determining module is further configured to obtain debt item data of a plurality of target objects in a predetermined time period of the target area; and determining the default rate of each time window based on the debt item data of the plurality of target objects in the target area within the predetermined time period.
Optionally, the apparatus further comprises:
the first acquisition module is used for acquiring boundary vector data of a plurality of target areas and integral noctilucent remote sensing data;
and the first obtaining module is used for obtaining the noctilucent remote sensing data of each target area based on the boundary vector data of each target area and the whole noctilucent remote sensing data.
Optionally, the noctilucent derivative characteristic index comprises a noctilucent total intensity index, a Moran index, a week-to-weekend noctilucent intensity contrast index, and a first and second half noctilucent intensity contrast index.
Optionally, the apparatus further comprises:
the third determining module is used for determining the same-proportion growth of the noctilucent derivative characteristic indexes;
and the training module is also used for training a default prediction model based on the default rate of each time window of the target area and the proportional increase of the noctilucent derivative characteristic indexes of each time window of the target area.
Optionally, the training module is further configured to train the default prediction model using a gradient boosting method, a linear regression method, a ridge regression method, an SVM method, a random forest method, or a deep learning method.
In a fourth aspect, the present application provides a default prediction apparatus based on noctilucent remote sensing data, including:
the fourth determination module is used for determining the noctilucent remote sensing data of the target area preset time window and determining noctilucent derivative characteristic indexes of the target area preset time window based on the noctilucent remote sensing data of the preset time window;
and the prediction module is used for determining default probability through any default prediction model shown in the first aspect on the basis of the noctilucent derivative characteristic indexes of the preset time window of the target area, and taking the default probability as a macroscopic credit risk assessment result.
Optionally, the apparatus further comprises:
the second acquisition module is used for acquiring boundary vector data of a plurality of target areas and integral noctilucent remote sensing data;
and the second obtaining module is used for obtaining the noctilucent remote sensing data of each target area based on the boundary vector data of each target area and the whole noctilucent remote sensing data.
Optionally, the noctilucent derivative characteristic index comprises a noctilucent total intensity index, a Moran index, a week-to-weekend noctilucent intensity contrast index, and a first and second half noctilucent intensity contrast index.
Optionally, the apparatus further comprises:
a fifth determining module, configured to determine a proportional increase of the noctilucent derived feature indicator;
the prediction module is further used for determining the default probability through any default prediction model shown in the first aspect based on the geometric increase of the noctilucent derivative characteristic indexes of the target area in the predetermined time window.
In a fifth aspect, an electronic device is provided, which includes:
one or more processors;
a memory;
one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to: the method shown in the first and second aspects is performed.
In a sixth aspect, a computer-readable storage medium is provided for storing computer instructions that, when executed on a computer, cause the computer to perform the method illustrated in the first and second aspects.
Compared with the prior art that macroscopic credit risk is carried out through official statistical data, the noctilucent remote sensing data of the target area in the preset time window are determined, and noctilucent derivative characteristic indexes of the target area in the preset time window are determined based on the noctilucent remote sensing data of the preset time window; and determining default probability through a default prediction model based on the noctilucence derivative characteristic indexes of the preset time window of the target area, and taking the default probability as a macroscopic credit risk assessment result. The assessment of the macroscopic credit risk is carried out according to the noctilucent remote sensing data, and the hysteresis quality of the assessment of the macroscopic credit risk based on official statistical data is avoided, so that the timeliness of the assessment of the macroscopic credit risk is improved.
Additional aspects and advantages of the present application will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the present application.
Drawings
The foregoing and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a schematic flow chart of a default prediction model training method based on noctilucent remote sensing data according to an embodiment of the present application;
FIG. 2 is a schematic flow chart of a default prediction method based on noctilucent remote sensing data according to an embodiment of the present application;
FIG. 3 is a schematic structural diagram of a default prediction model training device based on noctilucent remote sensing data according to an embodiment of the present application;
FIG. 4 is a schematic structural diagram of a default prediction apparatus based on noctilucent remote sensing data according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present application;
fig. 6 is a flowchart illustrating a default prediction model training method based on noctilucent remote sensing data according to an embodiment of the present application.
Detailed Description
Reference will now be made in detail to the embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary only for the purpose of explaining the present application and are not to be construed as limiting the present application.
As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. As used herein, the term "and/or" includes all or any element and all combinations of one or more of the associated listed items.
To make the objects, technical solutions and advantages of the present application more clear, embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
The following describes the technical solutions of the present application and how to solve the above technical problems with specific embodiments. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. Embodiments of the present application will be described below with reference to the accompanying drawings.
Example one
The embodiment of the application provides a default prediction model training method based on noctilucent remote sensing data, as shown in fig. 1, comprising the following steps:
step S101, determining historical noctilucent remote sensing data of each time window of a target area, and determining noctilucent derivative characteristic indexes of each time window based on the historical noctilucent remote sensing data of each time window; the target area can be one or a plurality of areas, wherein the target area can be an area divided according to administrative areas such as provinces, cities and counties; or areas of a specific industry such as a scientific park, and the like, so that the risk prediction model of the specific industry can be trained; the combination of administrative areas and scientific and technological parks can also be realized; wherein the time window may be a period of one month, one week, etc.
The noctilucent remote sensing data can be noctilucent data collected by a visible light infrared imaging radiometer (VIIRS) sensor carried by a Suomi national grade orbit partnership Satellite (SNPP), and can also be noctilucent remote sensing data collected by other modes capable of realizing the functions of the application.
Step S102, determining default rates of all time windows of a target area; specifically, debt data of a plurality of enterprises can be acquired, default conditions of the enterprises are determined, and then default rates within a certain time period are determined statistically.
Step S103, training default prediction models based on default rates of all time windows of the target area and noctilucent derived characteristic indexes of all time windows of the target area.
For the embodiment of the application, when economic situations in domestic and foreign regions are considered by a traditional risk assessment model, official statistical macro data are often used, and the problems of poor timeliness, inconsistent statistical skills, low accuracy and the like exist. The invention innovatively uses more objective and more timely noctilucent remote sensing data to evaluate the macroscopic credit risk, and provides the processing logic and more effective characteristic indexes of the noctilucent data in the field of risk control. Based on the economic and financial theory, a more scientific and timely macroscopic credit risk evaluation system is constructed, intelligent response to external environment changes can be rapidly and accurately made, and the overall wind control capability of the financial institution is improved. The embodiment of the application provides a default prediction model training method based on noctilucent remote sensing data, and a default prediction model is trained based on default rates of all time windows of a target area and noctilucent derivative characteristic indexes of all time windows of the target area, so that macroscopic risk prediction through official statistical data is avoided, timeliness of macroscopic credit risk prediction is improved, and losses of financial institutions are avoided.
Optionally, determining a default rate of each time window of the target area includes:
acquiring debt data of a plurality of target objects in a preset time period of a target area;
and determining default rates of the time windows based on the debt item data of the target objects in the target area within the preset time period.
Optionally, the method further comprises:
acquiring boundary vector data of a plurality of target areas and integral noctilucent remote sensing data;
and obtaining the noctilucent remote sensing data of each target area based on the boundary vector data of each target area and the whole noctilucent remote sensing data.
Optionally, the noctilucent derivative characteristic index comprises a noctilucent total intensity index, a Moran index, a week-to-weekend noctilucent intensity contrast index, and a first and second half noctilucent intensity contrast index. Thereby reducing or eliminating the influence caused by the difference of the noctilucence of the first and the second midnight in the middle and at the end of the week.
Optionally, the method further comprises:
determining the proportional increase of the noctilucent derivative characteristic index;
training a default prediction model based on default rates of all time windows of a target area and noctilucence derivative characteristic indexes of all time windows of the target area, wherein the training of the default prediction model comprises the following steps:
and training a default prediction model based on the default rate of each time window of the target area and the comparably increasing of the noctilucent derivative characteristic indexes of each time window of the target area.
Optionally, training the default prediction model based on the default rate of each time window of the target area and the proportional increase of the noctilucent derivative characteristic indicator of each time window of the target area includes:
and training the default prediction model by using a gradient lifting method, a linear regression method, a ridge regression method, an SVM method, a random forest method and a deep learning method.
Illustratively, the present application provides a model training method, which mainly includes the following steps:
acquiring loan application and debt data of the small and micro enterprises in the last three years, boundary vector data of province and conference cities, and Chinese day noctilucence data of a visible light infrared imaging radiometer (VIIRS) sensor carried by a Suomi national grade orbit partnership Satellite (SNPP);
calculating the default rate result of average debt item as the macroscopic credit risk index of each province in each month by taking the month as a time window
Cutting out noctilucent image data of each province and meeting city based on administrative region boundary vector data of each province and meeting city to obtain SNPP-VIIRS data of each province and meeting city every day in the last three years;
preprocessing noctilucent image data to eliminate the influence of natural light such as residual sunlight, moonlight, aurora, lightning, cloud and the like;
processing each province monthly noctilucence derivative index based on noctilucence image data, in the embodiment, using monthly average noctilucence total intensity I of each province meeting cityrtAnd representsMoire index M of luminous spatial distribution concentrationrtAs a main characteristic index;
because the noctilucent data has strong seasonal effect, each province city has difference, and the absolute value significance of the noctilucent intensity and the Moran index is not large, the concordance increase of each index is calculated to reflect the development change condition of each province city along with the time, and the concordance increase index calculation logic is as follows:
namely, the characteristic value obtained by subtracting the characteristic value of the same period of the last year from the characteristic value calculated by using the current time period is closed.
In this embodiment, the macroscopic credit risk of each province in the current month is predicted by using the same-ratio increase of luminous intensity and the same-ratio increase of the morland index of each province in the past 6 months, and in consideration of the possible nonlinear relationship between the viewing variables and the system risk of different macros, an extreme gradient boost method (XGBoost) is used for prediction training in this embodiment:
the XGboost model is constructed by using the noctilucence intensity characteristics and the Moran index characteristics of cities in the provinces in the past 6 months, and the total credit risk default rate of each province in the current period is predicted. In addition, other machine learning algorithms, such as linear regression, ridge regression, SVM, random forest, AdaBoost, decision trees, and DNN, may be used to train the model.
Example two
The embodiment of the application provides a default prediction method based on noctilucent remote sensing data, as shown in fig. 2, comprising the following steps:
step S201, determining noctilucence remote sensing data of a target area preset time window, and determining noctilucence derivative characteristic indexes of the target area preset time window based on the noctilucence remote sensing data of the preset time window;
step S202, determining default probability through the default prediction model of any one of the first aspect based on the noctilucent derivative characteristic indexes of the target area in the preset time window, and taking the default probability as a macroscopic credit risk assessment result.
Optionally, the method further comprises:
acquiring boundary vector data of a plurality of target areas and integral noctilucent remote sensing data;
and obtaining the noctilucent remote sensing data of each target area based on the boundary vector data of each target area and the whole noctilucent remote sensing data.
Optionally, the noctilucent derivative characteristic index comprises a noctilucent total intensity index, a Moran index, a week-to-weekend noctilucent intensity contrast index, and a first and second half noctilucent intensity contrast index. Thereby reducing or eliminating the influence caused by the difference of the noctilucence of the first and the second midnight in the middle and at the end of the week.
Optionally, the method further comprises:
determining the proportional increase of the noctilucent derivative characteristic index;
determining the default probability based on the noctilucence derivative characteristic indexes of the predetermined time window of the target area through the default prediction model of any one of the first aspect, including:
and determining the default probability by any default prediction model of the first aspect based on the comparably increasing noctilucent derived characteristic indexes of the target area in the predetermined time window.
Illustratively, the prediction method may comprise the steps of:
obtaining noctilucent data of cities of each province and each meeting in history and carrying out data preprocessing;
calculating and extracting the index of the same-ratio increase of the luminous intensity and the same-ratio increase of the luminous Molan index of each province in the past 6 months of the city based on the boundary vector data of each province;
and calculating the current macro credit risk assessment result of each province based on the pre-trained macro risk assessment model.
The beneficial effects of the embodiments of the present application are the same as those of the first embodiment, and are not described herein again.
EXAMPLE III
Fig. 3 is a default prediction model training device based on noctilucent remote sensing data according to an embodiment of the present application, which includes:
the first determining module 301 is configured to determine historical noctilucent remote sensing data of each time window of the target area, and determine noctilucent derivative characteristic indexes of each time window based on the historical noctilucent remote sensing data of each time window;
a second determining module 302, configured to determine a default rate of each time window of the target area;
the training module 303 is configured to train a default prediction model based on the default rate of each time window of the target area and the noctilucent derivative characteristic indicators of each time window of the target area.
Optionally, the second determining module is further configured to obtain debt item data of a plurality of target objects in a predetermined time period of the target area; and determining the default rate of each time window based on the debt item data of the plurality of target objects in the target area within the predetermined time period.
Optionally, the apparatus further comprises:
the first acquisition module is used for acquiring boundary vector data of a plurality of target areas and integral noctilucent remote sensing data;
and the first obtaining module is used for obtaining the noctilucent remote sensing data of each target area based on the boundary vector data of each target area and the whole noctilucent remote sensing data.
Optionally, the night-light derived characteristic indicators include a night-light total intensity indicator, a morn index indicator.
Optionally, the apparatus further comprises:
the third determining module is used for determining the same-proportion growth of the noctilucent derivative characteristic indexes;
and the training module is also used for training a default prediction model based on the default rate of each time window of the target area and the proportional increase of the noctilucent derivative characteristic indexes of each time window of the target area.
Optionally, the training module is further configured to train the default prediction model using a gradient boosting method, a linear regression method, a ridge regression method, an SVM method, a random forest method, or a deep learning method.
When economic situations in domestic and foreign areas are considered, the traditional risk assessment model often uses official statistical macroscopic data, and has the problems of poor timeliness, inconsistent statistical skills, low accuracy and the like. The invention innovatively uses more objective and more timely noctilucent remote sensing data to evaluate the macroscopic credit risk, and provides the processing logic and more effective characteristic indexes of the noctilucent data in the field of risk control. Based on the economic and financial theory, a more scientific and timely macroscopic credit risk evaluation system is constructed, intelligent response to external environment changes can be rapidly and accurately made, and the overall wind control capability of the financial institution is improved. The embodiment of the application provides a default prediction model training method based on noctilucent remote sensing data, and a default prediction model is trained based on default rates of all time windows of a target area and noctilucent derivative characteristic indexes of all time windows of the target area, so that macroscopic risk prediction through official statistical data is avoided, timeliness of macroscopic credit risk prediction is improved, and losses of financial institutions are avoided.
The beneficial effects of the apparatus of the embodiment of the present application are similar to those of the method shown in the first embodiment, and are not described herein again.
Example four
The embodiment of the application provides a default prediction device based on noctilucent remote sensing data, as shown in fig. 4, including:
a fourth determining module 401, configured to determine noctilucent remote sensing data of a predetermined time window of the target area, and determine a noctilucent derivative characteristic index of the predetermined time window of the target area based on the noctilucent remote sensing data of the predetermined time window;
and the predicting module 402 is used for determining default probability through a default predicting model based on the noctilucence derivative characteristic indexes of the preset time window of the target area, and taking the default probability as a macroscopic credit risk assessment result.
Optionally, the apparatus further comprises:
the second acquisition module is used for acquiring boundary vector data of a plurality of target areas and integral noctilucent remote sensing data;
and the second obtaining module is used for obtaining the noctilucent remote sensing data of each target area based on the boundary vector data of each target area and the whole noctilucent remote sensing data.
Optionally, the night-light derived characteristic indicators include a night-light total intensity indicator, a morn index indicator.
Optionally, the apparatus further comprises:
a fifth determining module, configured to determine a proportional increase of the noctilucent derived feature indicator;
and the prediction module is also used for determining the default probability through a default prediction model based on the isomorphic increase of the noctilucence derivative characteristic indexes of the target region in the preset time window.
When economic situations in domestic and foreign areas are considered, the traditional risk assessment model often uses official statistical macroscopic data, and has the problems of poor timeliness, inconsistent statistical skills, low accuracy and the like. The invention innovatively uses more objective and more timely noctilucent remote sensing data to evaluate the macroscopic credit risk, and provides the processing logic and more effective characteristic indexes of the noctilucent data in the field of risk control. Based on the economic and financial theory, a more scientific and timely macroscopic credit risk evaluation system is constructed, intelligent response to external environment changes can be rapidly and accurately made, and the overall wind control capability of the financial institution is improved. The embodiment of the application provides a default prediction model training method based on noctilucent remote sensing data, and a default prediction model is trained based on default rates of all time windows of a target area and noctilucent derivative characteristic indexes of all time windows of the target area, so that macroscopic risk prediction through official statistical data is avoided, timeliness of macroscopic credit risk prediction is improved, and losses of financial institutions are avoided.
The beneficial effects of the apparatus in the embodiment of the present application are similar to those of the method in the second embodiment, and are not described herein again.
EXAMPLE five
An embodiment of the present application provides an electronic device, as shown in fig. 5, an electronic device 50 shown in fig. 5 includes: a processor 501 and a memory 503. Wherein the processor 501 is coupled to the memory 503, such as via the bus 502. Further, the electronic device 50 may also include a transceiver 503. It should be noted that the transceiver 504 is not limited to one in practical application, and the structure of the electronic device 50 is not limited to the embodiment of the present application. The processor 501 is applied in the embodiment of the present application, and is used to implement the functions of the modules shown in fig. 3 or fig. 4. The transceiver 504 includes a receiver and a transmitter.
The processor 501 may be a CPU, general purpose processor, DSP, ASIC, FPGA or other programmable logic device, transistor logic device, hardware component, or any combination thereof. Which may implement or perform the various illustrative logical blocks, modules, and circuits described in connection with the disclosure. The processor 501 may also be a combination of implementing computing functionality, e.g., comprising one or more microprocessors, a combination of DSPs and microprocessors, and the like.
The memory 503 may be, but is not limited to, a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, an EEPROM, a CD-ROM or other optical disk storage, optical disk storage (including compact disk, laser disk, optical disk, digital versatile disk, blu-ray disk, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer.
The memory 503 is used for storing application program codes for executing the scheme of the application, and the processor 501 controls the execution. The processor 501 is configured to execute application program codes stored in the memory 503 to realize the functions of the apparatus provided by the embodiment shown in fig. 3 or fig. 4.
Compared with the prior art that macroscopic credit risk is carried out through official statistical data, the electronic equipment determines the noctilucent remote sensing data of the target area in the preset time window and determines the noctilucent derivative characteristic index of the target area in the preset time window based on the noctilucent remote sensing data of the preset time window; and determining default probability through a default prediction model based on the noctilucence derivative characteristic indexes of the preset time window of the target area, and taking the default probability as a macroscopic credit risk assessment result. The assessment of the macroscopic credit risk is carried out according to the noctilucent remote sensing data, and the hysteresis quality of the assessment of the macroscopic credit risk based on official statistical data is avoided, so that the timeliness of the assessment of the macroscopic credit risk is improved.
The embodiment of the application provides an electronic device suitable for the embodiment of the device. And will not be described in detail herein.
EXAMPLE VI
The present application provides a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to implement the apparatus shown in the above embodiments.
Compared with the prior art that macroscopic credit risk is carried out through official statistical data, the noctilucent remote sensing data of the target area in the preset time window are determined, and noctilucent derivative characteristic indexes of the target area in the preset time window are determined based on the noctilucent remote sensing data of the preset time window; and determining default probability through a default prediction model based on the noctilucence derivative characteristic indexes of the preset time window of the target area, and taking the default probability as a macroscopic credit risk assessment result. The assessment of the macroscopic credit risk is carried out according to the noctilucent remote sensing data, and the hysteresis quality of the assessment of the macroscopic credit risk based on official statistical data is avoided, so that the timeliness of the assessment of the macroscopic credit risk is improved.
The embodiment of the application provides a computer-readable storage medium which is suitable for the device embodiment. And will not be described in detail herein.
It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and may be performed in other orders unless explicitly stated herein. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
The foregoing is only a partial embodiment of the present application, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present application, and these modifications and decorations should also be regarded as the protection scope of the present application.
Claims (15)
1. A default prediction model training method based on noctilucent remote sensing data is characterized by comprising the following steps:
determining historical noctilucent remote sensing data of each time window of a target area, and determining noctilucent derivative characteristic indexes of each time window based on the historical noctilucent remote sensing data of each time window;
determining default rates of all time windows of a target area;
and training a default prediction model based on the default rate of each time window of the target area and the noctilucence derivative characteristic indexes of each time window of the target area.
2. The method of claim 1, wherein determining the rate of breach of the target area for each time window comprises:
acquiring debt data of a plurality of target objects in a preset time period of a target area;
and determining default rates of the time windows based on the debt item data of the target objects in the target area within the preset time period.
3. The method of claim 1, further comprising:
acquiring boundary vector data of a plurality of target areas and integral noctilucent remote sensing data;
and obtaining the noctilucent remote sensing data of each target area based on the boundary vector data of each target area and the whole noctilucent remote sensing data.
4. The method of claim 1, wherein the night-light derived characteristic indicators comprise a total intensity of night-light indicator, a Morgan index indicator, a weekend night-light intensity comparison indicator, and a first and second night-light intensity comparison indicator.
5. The method of claim 4, further comprising:
determining a geometric growth of the noctilucent derived characteristic indicator;
the training default prediction model based on the default rate of each time window of the target area and the noctilucent derivative characteristic indexes of each time window of the target area comprises the following steps:
and training a default prediction model based on the default rate of each time window of the target area and the comparably increasing of the noctilucent derivative characteristic indexes of each time window of the target area.
6. The method of claim 5, wherein training the default prediction model based on a geometric growth of the default rate for each time window of the target region and the night-light derived feature metrics for each time window of the target region comprises:
training the default prediction model by using a gradient lifting method, a linear regression method, a ridge regression method, an SVM method, a random forest method and a deep learning method.
7. A default prediction method based on noctilucent remote sensing data is characterized by comprising the following steps:
determining noctilucence remote sensing data of a preset time window of a target area, and determining noctilucence derivative characteristic indexes of the preset time window of the target area based on the noctilucence remote sensing data of the preset time window;
determining a default probability based on the night-luminous derivative characteristic indicators of the predetermined time window of the target area by the default prediction model of any one of claims 1 to 6, and using the default probability as a macroscopic credit risk assessment result.
8. The method of claim 7, further comprising:
acquiring boundary vector data of a plurality of target areas and integral noctilucent remote sensing data;
and obtaining the noctilucent remote sensing data of each target area based on the boundary vector data of each target area and the whole noctilucent remote sensing data.
9. The method of claim 8, wherein the night-light derived characteristic indicators comprise a total intensity of night-light indicator, a Morgan index indicator, a weekend night-light intensity comparison indicator, and a first and second night-light intensity comparison indicator.
10. The method of claim 9, further comprising:
determining a geometric growth of the noctilucent derived characteristic indicator;
the determination of the breach probability by the breach prediction model of any of claims 1-6 based on night-light derived feature metrics of a predetermined time window of the target area comprises:
determining a probability of breach by the breach prediction model of any of claims 1-6 based on a concordant increase in the night-light derived feature metrics for a predetermined time window of the target area.
11. The utility model provides a default prediction model training device based on night light remote sensing data which characterized in that includes:
the first determination module is used for determining historical noctilucent remote sensing data of each time window of a target area and determining noctilucent derivative characteristic indexes of each time window based on the historical noctilucent remote sensing data of each time window;
the second determining module is used for determining the default rate of each time window of the target area;
and the training module is used for training the default prediction model based on the default rate of each time window of the target area and the noctilucence derivative characteristic indexes of each time window of the target area.
12. The apparatus of claim 11, further comprising:
a third determining module, configured to determine a proportional increase of the noctilucent derived feature indicator;
the training module is further used for training a default prediction model based on the default rate of each time window of the target area and the proportional increase of the noctilucent derivative characteristic indexes of each time window of the target area.
13. A default prediction device based on noctilucent remote sensing data is characterized by comprising:
the fourth determination module is used for determining noctilucence remote sensing data of a preset time window of the target area and determining noctilucence derivative characteristic indexes of the preset time window of the target area based on the noctilucence remote sensing data of the preset time window;
a prediction module for determining a default probability based on the night-luminous derivative characteristic indicator of the predetermined time window of the target area by the default prediction model of any one of claims 1 to 6, and using the default probability as a macroscopic credit risk assessment result.
14. An electronic device, comprising:
one or more processors;
a memory;
one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more programs configured to: performing the method according to any one of claims 1 to 10.
15. A computer-readable storage medium for storing computer instructions which, when executed on a computer, cause the computer to perform the method of any of claims 1 to 10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011300431.1A CN112488820A (en) | 2020-11-19 | 2020-11-19 | Model training method and default prediction method based on noctilucent remote sensing data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011300431.1A CN112488820A (en) | 2020-11-19 | 2020-11-19 | Model training method and default prediction method based on noctilucent remote sensing data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112488820A true CN112488820A (en) | 2021-03-12 |
Family
ID=74931854
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011300431.1A Pending CN112488820A (en) | 2020-11-19 | 2020-11-19 | Model training method and default prediction method based on noctilucent remote sensing data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112488820A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113378724A (en) * | 2021-06-15 | 2021-09-10 | 中南大学 | Multi-center city hot spot area rapid identification and dynamic monitoring method |
| CN115131685A (en) * | 2022-09-01 | 2022-09-30 | 北京佳格天地科技有限公司 | Analysis method for monitoring enterprise production and operation conditions based on satellite noctilucent data |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106485360A (en) * | 2016-10-14 | 2017-03-08 | 华中师范大学 | Segmental society's prediction of economic indexes method and system based on overall noctilucence remote sensing |
| CN108334838A (en) * | 2018-01-30 | 2018-07-27 | 天津中科智能识别产业技术研究院有限公司 | Based on light remotely-sensed data National Macroscopic economic indicator estimating and measuring method |
| CN108399399A (en) * | 2018-03-23 | 2018-08-14 | 武汉大学 | A kind of city scope extracting method based on noctilucence remote sensing image |
| CN110349012A (en) * | 2019-07-12 | 2019-10-18 | 腾讯科技（深圳）有限公司 | Data predication method and computer readable storage medium |
-
2020
- 2020-11-19 CN CN202011300431.1A patent/CN112488820A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106485360A (en) * | 2016-10-14 | 2017-03-08 | 华中师范大学 | Segmental society's prediction of economic indexes method and system based on overall noctilucence remote sensing |
| CN108334838A (en) * | 2018-01-30 | 2018-07-27 | 天津中科智能识别产业技术研究院有限公司 | Based on light remotely-sensed data National Macroscopic economic indicator estimating and measuring method |
| CN108399399A (en) * | 2018-03-23 | 2018-08-14 | 武汉大学 | A kind of city scope extracting method based on noctilucence remote sensing image |
| CN110349012A (en) * | 2019-07-12 | 2019-10-18 | 腾讯科技（深圳）有限公司 | Data predication method and computer readable storage medium |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113378724A (en) * | 2021-06-15 | 2021-09-10 | 中南大学 | Multi-center city hot spot area rapid identification and dynamic monitoring method |
| CN115131685A (en) * | 2022-09-01 | 2022-09-30 | 北京佳格天地科技有限公司 | Analysis method for monitoring enterprise production and operation conditions based on satellite noctilucent data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109615226B (en) | Operation index abnormity monitoring method | |
| Chen et al. | An intelligent pattern recognition model for supporting investment decisions in stock market | |
| Ma et al. | A method for multiple periodic factor prediction problems using complex fuzzy sets | |
| Fay et al. | On the influence of weather forecast errors in short-term load forecasting models | |
| CN111680841B (en) | Short-term load prediction method, system and terminal equipment based on principal component analysis | |
| CN112488820A (en) | Model training method and default prediction method based on noctilucent remote sensing data | |
| Girimurugan et al. | Application of deep learning to the prediction of solar irradiance through missing data | |
| Tung et al. | Binary classification and data analysis for modeling calendar anomalies in financial markets | |
| Fan | Research on deep learning energy consumption prediction based on generating confrontation network | |
| Priatna et al. | Precipitation prediction using recurrent neural networks and long short-term memory | |
| Karozis et al. | A deep learning approach for spatial error correction of numerical seasonal weather prediction simulation data | |
| Chao et al. | A spatio-temporal neural network learning system for city-scale carbon storage capacity estimating | |
| CN117391466A (en) | Novel early warning method and system for contradictory dispute cases | |
| CN112598181A (en) | Load prediction method, device, equipment and storage medium | |
| CN117277304A (en) | Photovoltaic power generation ultra-short-term power prediction method and system considering sunrise and sunset time | |
| Guo et al. | Research on precipitation prediction based on a complete ensemble empirical mode decomposition with adaptive noise–long short-term memory coupled model | |
| Pan et al. | Modeling optimization method based on Gamma test and NSGA II for forecast of PV power output | |
| Jiang et al. | Evaluation of county-level poverty alleviation progress by deep learning and satellite observations | |
| He et al. | Time Series | |
| Kocharekar et al. | Weather prediction for tourism application using arima | |
| Sahani et al. | Precise single step and multistep short-term photovoltaic parameters forecasting based on reduced deep convolutional stack autoencoder and minimum variance multikernel random vector functional network | |
| Karbasi et al. | Estimation of rural-urban migration in iran (an auto-regressive distributed lag approach) | |
| Zhong et al. | Wind power generation prediction during the COVID-19 epidemic based on novel hybrid deep learning techniques | |
| Jia et al. | A prediction model for the secure issuance scale of Chinese local government bonds | |
| CN116150570B (en) | Storage and superposition analysis calculation method and device for high-precision grid rainfall |
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 |