CN111242452A - Financial risk data analysis control system and method - Google Patents

Financial risk data analysis control system and method Download PDF

Info

Publication number
CN111242452A
CN111242452A CN202010012637.8A CN202010012637A CN111242452A CN 111242452 A CN111242452 A CN 111242452A CN 202010012637 A CN202010012637 A CN 202010012637A CN 111242452 A CN111242452 A CN 111242452A
Authority
CN
China
Prior art keywords
index
particle
sequence
indexes
time sequence
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
Application number
CN202010012637.8A
Other languages
Chinese (zh)
Inventor
章英海
周胜
李鸿达
潘翔鹰
吴镇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Wangxinsi Software Co Ltd
Original Assignee
Shenzhen Wangxinsi Software Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenzhen Wangxinsi Software Co Ltd filed Critical Shenzhen Wangxinsi Software Co Ltd
Priority to CN202010012637.8A priority Critical patent/CN111242452A/en
Publication of CN111242452A publication Critical patent/CN111242452A/en
Pending legal-status Critical Current

Links

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
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • 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
    • 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
    • G06Q40/06—Asset management; Financial planning or analysis

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • Development Economics (AREA)
  • Theoretical Computer Science (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Game Theory and Decision Science (AREA)
  • Technology Law (AREA)
  • Educational Administration (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Quality & Reliability (AREA)
  • Data Mining & Analysis (AREA)
  • Tourism & Hospitality (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

A regional financial risk data analysis control system and method are provided. The method simultaneously adopts a time series algorithm and a particle swarm algorithm to perform fusion analysis on regional financial risk data; not only the case of positive correlation with the prediction index sequence is concerned, but also the case of negative correlation with the prediction index sequence is concerned; and external indexes which are ranked at the top are selected during index optimization, an initial time sequence is formed together with the prediction index time sequence, and the accuracy and the stability of the algorithm are greatly improved by combining the optimization of the particle swarm algorithm on parameters.

Description

Financial risk data analysis control system and method
Technical Field
The invention belongs to the technical field of green intelligent finance; to a data analysis control system and method, and more particularly, to a regional financial risk data analysis control system and method.
Background
Finance is the first core competitiveness in countries and regions, and similarly, financial security is an important component of national and regional security. In recent years, with the innovative development of the international financial market and the rapid rise of internet financial science and technology, the method increases the national and regional financial risk and instability while increasing the vitality of the continuously developed financial market in China. In addition, with the continuous advance of supply-side innovation, partial areas and industries are in a liquidity crisis, financial risks caused by the local debt and the folk loan risk of each area are more and more prominent, and each economic department in the area is continuously accumulated and spread, so that the stability of the area finance and economy is influenced through various channels.
This situation has attracted much attention from the center, and formally put preventive and financial risks into the first three major combat in the report. With the increasing perfection of financial supervision means, the accumulated financial risk is gradually relieved, and a safe and controllable state is presented generally.
However, due to different resource endowments and economic and social development levels in different regions, the development stages and levels of regional financial markets have great individual differences, and the financial risk bearing capacity is different.
Therefore, reasonable analysis of regional financial data and effective control of regional financial risk have essential and important values for maintaining the stability of regional finance, and the importance and urgency of regional finance are increasingly significant.
People mainly quantitatively judge the degree of regional financial risk through different index systems. For example, the national statistical institute has proposed an index system with 33 indexes in a total of 5 major categories. But some index systems are not specific and have poor operability; in addition, the selection standards of part of indexes are not clear enough, and the statistical compliance existing among the indexes is not fully considered, so that the calculation workload is huge, and even the precision of an index system is influenced.
On the other hand, a plurality of quantization indexes may be arranged in a chronological order to form a dynamic sequence, also referred to as a time sequence. By establishing a mathematical model for financial data within a certain length range, the dynamic dependence relationship contained in each index of the time sequence can be accurately analyzed and fitted, and the future numerical value or behavior can be predicted by utilizing the dynamic dependence relationship. However, because the data adopted in the actual prediction is complex, the adaptability of the data is not high only by adopting a single time sequence, and the technical defects of low precision and poor stability exist; when a plurality of time sequences are adopted, the distance between different sequences after magnitude difference is eliminated is mainly concerned, namely the situation of positive correlation with the prediction index sequence is mainly considered, and the situation of negative correlation with the prediction index sequence is ignored. All of the above results in poor precision and stability of the above method.
Therefore, in view of the technical defects, it is desirable to provide a regional financial risk data analysis control system and method with higher accuracy and better stability.
Disclosure of Invention
The invention aims to provide a regional financial risk data analysis control system and a regional financial risk data analysis control method.
In order to achieve the above object, in one aspect, the present invention provides a regional financial risk data analysis control method, including:
step S1, selecting technical indexes, importing data and establishing a time sequence of each index;
step S2, taking a certain technical index as a prediction index and the other indexes as external indexes, and optimizing the external indexes by calculating the statistical relevance to obtain an initial time sequence;
step S3, accumulating the initial time sequence to obtain a new sequence and a whitening differential equation thereof; obtaining parameters to be estimated according to a least square formula, and simultaneously obtaining a predicted value model of an initial sequence;
step S4, performing parameter optimization by using a particle swarm algorithm, and obtaining a new predicted value model based on the optimized parameters to be estimated;
and step S5, comparing the predicted value of the prediction index with a preset risk threshold value, and determining the risk grade of the prediction index.
The analysis control method according to the present invention is characterized in that step S1 specifically includes:
and step S11, selecting 15 technical indexes of the balance rate, the sales profit rate, the relative labor production rate, the product sales rate, the M1/M2 ratio, the deposit-loan ratio, the bad loan rate, the financial deficit rate, the capital abundance rate, the housing sales price index, the commodity housing sales area/completion area, the stock market profit rate, the foreign exchange reserve amount, the PMI and the foreign net asset growth rate which reflect the regional finance average level, importing data and establishing a time sequence of each index.
The analysis control method according to the present invention is characterized in that the prediction index of step S2 is a rate of liability.
The analysis control method according to the present invention is characterized in that step S2 specifically includes:
step S21, normalizing the time series of each index;
step S22, calculating the statistical relevance of the external index time sequence by taking the rate of capital and debt as the prediction index time sequence and the other indexes as the external index time sequence;
s23, sorting the external indexes according to the statistical relevance, further selecting the n-1 external indexes before ranking, and forming an initial time sequence together with the time sequence of the prediction indexes
Figure BDA0002357704810000041
Figure BDA0002357704810000042
The analysis control method according to the present invention is characterized in that n is an integer of 4 to 10. Preferably, n is an integer from 4 to 9; more preferably, n is an integer from 4 to 8; and, most preferably, n is an integer from 4 to 6.
In a specific embodiment, n is 5.
The analysis control method according to the present invention is characterized in that the normalization processing formula of step S21 is:
Figure BDA0002357704810000043
the analysis control method according to the present invention is characterized in that the statistical association degree formula of step S22 is:
Figure BDA0002357704810000044
wherein the content of the first and second substances,
Figure BDA0002357704810000045
Figure BDA0002357704810000051
Figure BDA0002357704810000052
Figure BDA0002357704810000053
rho is a decimal between 0 and 1;
Figure BDA0002357704810000054
Figure BDA0002357704810000055
preferably, ρ is a fraction between 0.2 and 0.8; more preferably, ρ is between 0.3 and 0.7; and, most preferably, ρ is a fraction between 0.4 and 0.6.
In a specific embodiment, ρ is 0.5.
The analysis control method according to the present invention is characterized in that step S3 specifically includes:
step S31, accumulating the initial time sequence to obtain a new sequence
Figure BDA0002357704810000056
Figure BDA0002357704810000057
Wherein the content of the first and second substances,
Figure BDA0002357704810000061
step S32, obtaining a new sequence
Figure BDA0002357704810000062
The whitening differential equation of (a) is as follows:
Figure BDA0002357704810000063
wherein, b1，b2，…，bnAnd u is the parameter to be estimated, t is 1,2, …, f; f is the number of predicted terms;
step S33, integral transformation is carried out on the whitening differential equation in the [ k-1, k ] interval, and the following steps are obtained:
Figure BDA0002357704810000064
(ii) a Wherein the content of the first and second substances,
Figure BDA0002357704810000065
step S34, calculating the parameters to be estimated according to the following least square formula;
Figure BDA0002357704810000066
wherein the content of the first and second substances,
Figure BDA0002357704810000067
Figure BDA0002357704810000071
Figure BDA0002357704810000072
step S35, obtaining a predicted value of the initial sequence,
Figure BDA0002357704810000073
the analysis control method according to the present invention is characterized in that step S4 specifically includes:
step S41, initializing a particle swarm, setting a swarm size N, a particle dimension D and a maximum iteration number TmaxIn the definition space RDIn which j particles X are randomly generated1，X2，…，XjAnd randomly generating an initial velocity V of each particle1，V2，…，Vj；
Step S42, selecting a fitness function, and calculating the fitness value of each particle in the population according to the fitness function;
step S43, the initialization positions of the respective particles are determined as the individually initialized individual optimum values pbestjDetermining an initial population optimal value gbest according to the fitness value of each particle;
step S44, updating the speed and position of each particle according to the speed and position
Figure BDA0002357704810000074
The particle velocity is constrained, and the particle position outside the search space range is reset, wherein,
Figure BDA0002357704810000075
denotes the jth particle P at the time k +1jA position vector in a d-dimensional search space;
step S45, calculating the fitness value of each particle after updating, and comparing the new fitness value with pbestjIf more preferred, the particles are usedNew location of replacement pbestiOtherwise, the replacement is not carried out; compare each pbestjDetermining a new gbest according to the fitness value of the gbest;
step S46, checking whether the iteration number reaches the maximum value TmaxIf the stop condition of (3) is satisfied, the search is stopped and the search result gbest is output, otherwise, let t be t +1, and return to step S44 to continue the search.
The analysis control method according to the present invention is characterized in that the population size N of step S41 is 40; particle dimension D10; the maximum iteration time Tmax is 500.
The analysis control method according to the present invention is characterized in that the fitness function of step S42 is an average relative error function,
Figure BDA0002357704810000081
wherein the content of the first and second substances,
Figure BDA0002357704810000082
represents the predicted value of the prediction index sequence at the time k,
Figure BDA0002357704810000083
for its true value, m is the number of sequences.
In another aspect, the present invention further provides a regional financial risk data analysis control system, including:
the data import module selects technical indexes, imports data and establishes a time sequence of each index;
the external index optimization module takes a certain technical index as a prediction index and the other indexes as external indexes, and optimizes the external indexes by calculating the statistical relevance to obtain an initial time sequence;
a time sequence analysis module, which accumulates the initial time sequence to obtain a new sequence and a whitening differential equation thereof; obtaining parameters to be estimated according to a least square formula, and simultaneously obtaining a predicted value model of an initial sequence;
the particle swarm optimization module is used for optimizing parameters by using a particle swarm algorithm and obtaining a new predicted value model based on the optimized parameters to be estimated;
and the risk rating module compares the predicted value of the prediction index with a preset risk threshold value to determine the risk level of the prediction index.
Compared with the prior art, the method has the advantages that the method adopts the time sequence algorithm and the particle swarm algorithm to perform fusion analysis on the regional financial risk data, and particularly pays attention to the situation of positive correlation with the prediction index sequence (namely α) when a plurality of time sequences are adoptedi) Attention is also paid to the case of negative correlation (β)i) (ii) a And when the indexes are optimized, external indexes n-1 before the ranking are selected, the external indexes and the prediction index time sequence form an initial time sequence together, and the accuracy and the stability of the algorithm are greatly improved by combining the optimization of the particle swarm algorithm on the parameters.
Detailed Description
Hereinafter, specific embodiments of the present invention will be described in detail, and those skilled in the art can clearly understand the present invention and can implement the present invention according to the detailed description. Features from different embodiments may be combined to yield new embodiments, or some features may be substituted for others to yield yet further preferred embodiments, without departing from the principles of the present invention.
In a specific embodiment, there is provided a regional financial risk data analysis control method, including:
step S1, selecting technical indexes, importing data and establishing a time sequence of each index;
step S2, taking a certain technical index as a prediction index and the other indexes as external indexes, and optimizing the external indexes by calculating the statistical relevance to obtain an initial time sequence;
step S3, accumulating the initial time sequence to obtain a new sequence and a whitening differential equation thereof; obtaining parameters to be estimated according to a least square formula, and simultaneously obtaining a predicted value model of an initial sequence;
step S4, performing parameter optimization by using a particle swarm algorithm, and obtaining a new predicted value model based on the optimized parameters to be estimated;
and step S5, comparing the predicted value of the prediction index with a preset risk threshold value, and determining the risk grade of the prediction index.
In a specific embodiment, step S1 is specifically:
and step S11, selecting 15 technical indexes of the balance rate, the sales profit rate, the relative labor production rate, the product sales rate, the M1/M2 ratio, the deposit-loan ratio, the bad loan rate, the financial deficit rate, the capital abundance rate, the housing sales price index, the commodity housing sales area/completion area, the stock market profit rate, the foreign exchange reserve amount, the PMI and the foreign net asset growth rate which reflect the regional finance average level, importing data and establishing a time sequence of each index.
In one specific embodiment, the prediction index of step S2 is a rate of assets liability.
Generally speaking, poor loan rate, financial deficit rate, housing sales price index, stock market profitability, commodity housing sales area/completion area, PMI, and foreign net asset growth rate have a positive correlation with the rate of assets liability; and the sales profit rate, the relative labor production rate, the product sales rate, the M1/M2 ratio, the inventory rate, the capital abundance rate, the housing sales price index, the stock market abundance rate, the foreign exchange reserve amount and the asset liability rate present a positive correlation.
In a specific embodiment, step S2 is specifically:
step S21, normalizing the time series of each index;
step S22, calculating the statistical relevance of the external index time sequence by taking the rate of capital and debt as the prediction index time sequence and the other indexes as the external index time sequence;
s23, sorting the external indexes according to the statistical relevance, further selecting the n-1 external indexes before ranking, and forming an initial time together with the time sequence of the prediction indexesIntermediate sequence
Figure BDA0002357704810000111
Figure BDA0002357704810000112
In a specific embodiment, the normalization processing formula of step S21 is:
Figure BDA0002357704810000113
in a specific embodiment, the statistical association formula of step S22 is:
Figure BDA0002357704810000114
wherein the content of the first and second substances,
Figure BDA0002357704810000115
Figure BDA0002357704810000121
Figure BDA0002357704810000122
Figure BDA0002357704810000123
rho is a decimal between 0 and 1;
Figure BDA0002357704810000124
Figure BDA0002357704810000125
in a specific embodiment, ρ is 0.5.
In a specific embodiment, n is 5. The statistical association degrees of the 4 external indexes obtained by the statistical association degree calculation method are 0.8724, 0.8327, 0.7968 and 0.7841 respectively; these 4 indices are then combined with the prediction index time series to form an initial time series.
In a specific embodiment, step S3 is specifically:
step S31, accumulating the initial time sequence to obtain a new sequence
Figure BDA0002357704810000126
Figure BDA0002357704810000127
Wherein the content of the first and second substances,
Figure BDA0002357704810000131
step S32, obtaining a new sequence
Figure BDA0002357704810000132
The whitening differential equation of (a) is as follows:
Figure BDA0002357704810000133
wherein, b1，b2，…，bnAnd u is the parameter to be estimated, t is 1,2, …, f; f is the number of predicted terms;
step S33, integral transformation is carried out on the whitening differential equation in the [ k-1, k ] interval, and the following steps are obtained:
Figure BDA0002357704810000134
(ii) a Wherein the content of the first and second substances,
Figure BDA0002357704810000135
step S34, calculating the parameters to be estimated according to the following least square formula;
Figure BDA0002357704810000136
wherein the content of the first and second substances,
Figure BDA0002357704810000137
Figure BDA0002357704810000141
Figure BDA0002357704810000142
step S35, obtaining a predicted value of the initial sequence,
Figure BDA0002357704810000143
in a specific embodiment, the parameter b is obtained according to the above calculation method of the present invention1＝0.7463；b2＝-0.4982；b3＝0.3146；b4＝0.9127；b5＝0.0584；u＝2842.5。
In a specific embodiment, step S4 is specifically:
step S41, initializing a particle swarm, setting a swarm size N, a particle dimension D and a maximum iteration number TmaxIn the definition space RDIn which j particles X are randomly generated1，X2，…，XjAnd randomly generating an initial velocity V of each particle1，V2，…，Vj；
Step S42, selecting a fitness function, and calculating the fitness value of each particle in the population according to the fitness function;
step S43, the initialization positions of the respective particles are determined as the individually initialized individual optimum values pbestjDetermining an initial population optimal value gbest according to the fitness value of each particle;
step S44, updating the speed and position of each particle according to the speed and position
Figure BDA0002357704810000144
The particle velocity is constrained, and the particle position outside the search space range is reset, wherein,
Figure BDA0002357704810000151
denotes the jth particle P at the time k +1jA position vector in a d-dimensional search space;
step S45, calculating the fitness value of each particle after updating, and comparing the new fitness value with pbestjIf better, the new position of the particle is used to replace pbestiOtherwise, the replacement is not carried out; compare each pbestjDetermining a new gbest according to the fitness value of the gbest;
step S46, checking whether the iteration number reaches the maximum value TmaxIf the stop condition of (3) is satisfied, the search is stopped and the search result gbest is output, otherwise, let t be t +1, and return to step S44 to continue the search.
In a specific embodiment, the population size N of step S41 is 40; particle dimension D10; maximum number of iterations Tmax＝500。
In one specific embodiment, the fitness function of step S42 is the following average relative error function,
Figure BDA0002357704810000152
wherein the content of the first and second substances,
Figure BDA0002357704810000153
represents the predicted value of the prediction index sequence at the time k,
Figure BDA0002357704810000154
for its true value, m is the number of sequences.
In a specific embodiment, the preset risk threshold of step S5 is 60-100%.
In a more specific embodiment, the preset risk threshold of step S5 is 80%.
In another specific embodiment, there is also provided a regional financial risk data analysis control system, including:
the data import module selects technical indexes, imports data and establishes a time sequence of each index;
the external index optimization module takes a certain technical index as a prediction index and the other indexes as external indexes, and optimizes the external indexes by calculating the statistical relevance to obtain an initial time sequence;
a time sequence analysis module, which accumulates the initial time sequence to obtain a new sequence and a whitening differential equation thereof; obtaining parameters to be estimated according to a least square formula, and simultaneously obtaining a predicted value model of an initial sequence;
the particle swarm optimization module is used for optimizing parameters by using a particle swarm algorithm and obtaining a new predicted value model based on the optimized parameters to be estimated;
and the risk rating module compares the predicted value of the prediction index with a preset risk threshold value to determine the risk level of the prediction index.
The relative error Q, the variance ratio C and the small probability error p of the aforementioned analysis control system and method of the present invention are 0.0094, 0.285 and 1.0, respectively, the system and method obtained using all 15 indices without using step S2 are 0.062, 0.179 and 1.0, respectively, and only α is usediThe relative error Q, the variance ratio C and the small probability error p of the system and the method obtained by the analysis control method are respectively 0.017, 0.261 and 1.0; the relative error Q, the variance ratio C and the small probability error p of the obtained system and method without optimization by adopting a particle swarm optimization are respectively 0.028, 0.182 and 1.0.
According to the method, the time sequence algorithm and the particle swarm optimization are adopted to perform fusion analysis on the regional golden fusion risk data; especially, when a plurality of time sequences are adopted, not only the situation of positive correlation with the prediction index sequence is concerned, but also the situation of negative correlation with the prediction index sequence is concerned; and in the index optimization, external indexes n-1 before the ranking are selected, an initial time sequence is formed together with the prediction index time sequence, and the relative error is greatly improved by combining the optimization of the particle swarm algorithm on the parameters, so that the accuracy and the stability of the algorithm are improved.
Although the present invention has been described above with reference to specific embodiments, it will be appreciated by those skilled in the art that many modifications are possible in the arrangement and details of the invention disclosed within the principle and scope of the invention. The scope of the invention is to be determined by the appended claims, and all changes that come within the meaning and range of equivalency of the technical features are intended to be embraced therein.

Claims (10)

1. A regional financial risk data analysis control method, the method comprising:
step S1, selecting technical indexes, importing data and establishing a time sequence of each index;
step S2, taking a certain technical index as a prediction index and the other indexes as external indexes, and optimizing the external indexes by calculating the statistical relevance to obtain an initial time sequence;
step S3, accumulating the initial time sequence to obtain a new sequence and a whitening differential equation thereof; obtaining parameters to be estimated according to a least square formula, and simultaneously obtaining a predicted value model of an initial sequence;
step S4, performing parameter optimization by using a particle swarm algorithm, and obtaining a new predicted value model based on the optimized parameters to be estimated;
and step S5, comparing the predicted value of the prediction index with a preset risk threshold value, and determining the risk grade of the prediction index.
2. The analysis control method according to claim 1, wherein step S1 is specifically:
and step S11, selecting 15 technical indexes of the balance rate, the sales profit rate, the relative labor production rate, the product sales rate, the M1/M2 ratio, the deposit-loan ratio, the bad loan rate, the financial deficit rate, the capital abundance rate, the housing sales price index, the commodity housing sales area/completion area, the stock market profit rate, the foreign exchange reserve amount, the PMI and the foreign net asset growth rate which reflect the regional finance average level, importing data and establishing a time sequence of each index.
3. The analysis control method according to claim 1, wherein step S2 is specifically:
step S21, normalizing the time series of each index;
step S22, calculating the statistical relevance of the external index time sequence by taking the rate of capital and debt as the prediction index time sequence and the other indexes as the external index time sequence;
s23, sorting the external indexes according to the statistical relevance, further selecting the n-1 external indexes before ranking, and forming an initial time sequence together with the time sequence of the prediction indexes
Figure FDA0002357704800000021
Figure FDA0002357704800000022
4. The analysis control method according to claim 3, wherein the normalization processing formula of step S21 is:
Figure FDA0002357704800000023
5. the analysis control method according to claim 3, wherein the statistical association degree formula of step S22 is:
Figure FDA0002357704800000024
wherein the content of the first and second substances,
Figure FDA0002357704800000025
Figure FDA0002357704800000026
Figure FDA0002357704800000031
Figure FDA0002357704800000032
rho is a decimal between 0 and 1;
Figure FDA0002357704800000033
Figure FDA0002357704800000034
6. the analysis control method according to claim 1, wherein step S3 is specifically:
step S31, accumulating the initial time sequence to obtain a new sequence
Figure FDA0002357704800000035
Figure FDA0002357704800000036
Wherein the content of the first and second substances,
Figure FDA0002357704800000037
step S32, obtaining a new sequence
Figure FDA0002357704800000038
The whitening differential equation of (a) is as follows:
Figure FDA0002357704800000039
wherein, b1，b2，…，bnAnd u is the parameter to be estimated, t is 1,2, …, f; f is the number of predicted terms;
step S33, integral transformation is carried out on the whitening differential equation in the [ k-1, k ] interval, and the following steps are obtained:
Figure FDA0002357704800000041
；
wherein the content of the first and second substances,
Figure FDA0002357704800000042
step S34, calculating the parameters to be estimated according to the following least square formula;
Figure FDA0002357704800000043
wherein the content of the first and second substances,
Figure FDA0002357704800000044
Figure FDA0002357704800000045
Figure FDA0002357704800000046
step S35, obtaining a predicted value of the initial sequence,
Figure FDA0002357704800000051
7. the analysis control method according to claim 1, wherein step S4 is specifically:
step S41, initializing a particle swarm, setting the size N of the swarm, the dimension D of the particle and the maximum iteration time Tmax, and defining a space RDIn which j particles X are randomly generated1，X2，…，XjAnd randomly generating an initial velocity V of each particle1，V2，…，Vj；
Step S42, selecting a fitness function, and calculating the fitness value of each particle in the population according to the fitness function;
step S43, the initialization positions of the respective particles are determined as the individually initialized individual optimum values pbestjDetermining an initial population optimal value gbest according to the fitness value of each particle;
step S44, updating the speed and position of each particle according to the speed and position
Figure FDA0002357704800000052
The particle velocity is constrained, the particle position outside the search space is reset, wherein,
Figure FDA0002357704800000053
denotes the jth particle P at the time k +1jA position vector in a d-dimensional search space;
step S45, calculating the fitness value of each particle after updating, and comparing the new fitness value with pbestjIf better, the new position of the particle is used to replace pbestjOtherwise, not replacing; compare each pbestjDetermining a new gbest according to the fitness value of the gbest;
step S46, checking whether the iteration number is satisfiedMaximum value TmaxIf the stop condition of (3) is satisfied, the search is stopped and the search result gbest is output, otherwise, let t be t +1, and return to step S44 to continue the search.
8. The analysis control method according to claim 7, wherein the population size N of step S41 is 40; particle dimension D10; the maximum iteration time Tmax is 500.
9. The analysis control method according to claim 7, wherein the fitness function of step S42 is the following average relative error function,
Figure FDA0002357704800000061
wherein the content of the first and second substances,
Figure FDA0002357704800000062
represents the predicted value of the prediction index sequence at the time k,
Figure FDA0002357704800000063
for its true value, m is the number of sequences.
10. A regional financial risk data analysis control system for performing the method of any of claims 1 to 9, the system comprising:
the data import module selects technical indexes, imports data and establishes a time sequence of each index;
the external index optimization module takes a certain technical index as a prediction index and the other indexes as external indexes, and optimizes the external indexes by calculating the statistical relevance to obtain an initial time sequence;
a time sequence analysis module, which accumulates the initial time sequence to obtain a new sequence and a whitening differential equation thereof; obtaining parameters to be estimated according to a least square formula, and simultaneously obtaining a predicted value model of an initial sequence;
the particle swarm optimization module is used for optimizing parameters by using a particle swarm algorithm and obtaining a new predicted value model based on the optimized parameters to be estimated;
and the risk rating module compares the predicted value of the prediction index with a preset risk threshold value to determine the risk level of the prediction index.
CN202010012637.8A 2020-01-07 2020-01-07 Financial risk data analysis control system and method Pending CN111242452A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010012637.8A CN111242452A (en) 2020-01-07 2020-01-07 Financial risk data analysis control system and method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010012637.8A CN111242452A (en) 2020-01-07 2020-01-07 Financial risk data analysis control system and method

Publications (1)

Publication Number Publication Date
CN111242452A true CN111242452A (en) 2020-06-05

Family

ID=70864843

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010012637.8A Pending CN111242452A (en) 2020-01-07 2020-01-07 Financial risk data analysis control system and method

Country Status (1)

Country Link
CN (1) CN111242452A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113947248A (en) * 2021-10-21 2022-01-18 广东电网有限责任公司广州供电局 Method and related device for predicting risk of cable moisture aging tripping
CN116361860A (en) * 2022-12-27 2023-06-30 深圳市网新新思软件有限公司 Information storage and verification method, device, equipment and storage medium

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113947248A (en) * 2021-10-21 2022-01-18 广东电网有限责任公司广州供电局 Method and related device for predicting risk of cable moisture aging tripping
CN113947248B (en) * 2021-10-21 2024-05-31 广东电网有限责任公司广州供电局 Risk prediction method and related device for cable damp aging tripping
CN116361860A (en) * 2022-12-27 2023-06-30 深圳市网新新思软件有限公司 Information storage and verification method, device, equipment and storage medium
CN116361860B (en) * 2022-12-27 2024-02-09 深圳市网新新思软件有限公司 Information storage and verification method, device, equipment and storage medium

Similar Documents

Publication Publication Date Title
Lin et al. Novel feature selection methods to financial distress prediction
CN110009479A (en) Credit assessment method and device, storage medium, computer equipment
CN108711107A (en) Intelligent financing services recommend method and its system
CN110956273A (en) Credit scoring method and system integrating multiple machine learning models
Zowid et al. Multi-criteria inventory ABC classification using Gaussian Mixture Model
CN107563645A (en) A kind of Financial Risk Analysis method based on big data
WO2016183391A1 (en) System, method and computer-accessible medium for making a prediction from market data
CN109102386A (en) Risk control method and device based on big data
Chen et al. Research on credit card default prediction based on k-means SMOTE and BP neural network
CN111242452A (en) Financial risk data analysis control system and method
CN113111924A (en) Electric power customer classification method and device
US8463678B2 (en) Generating method for transaction models with indicators for option
Haga et al. Initial stage clustering when estimating accounting quality measures with self-organizing maps
Mukherjee et al. Detection of defaulters in P2P lending platforms using unsupervised learning
CN107093005A (en) The method that tax handling service hall's automatic classification is realized based on big data mining algorithm
CN111292188A (en) Intelligent integrated return test system for pairing transaction
CN109978300B (en) Customer risk tolerance quantification method and system, and asset configuration method and system
CN115115447A (en) Hyper-parameter optimization method and device of credit wind control model
Li Design of Financial Crisis Early Warning Model Based on PSO‐SVM Algorithm
Chen [Retracted] BP Neural Network Based on Simulated Annealing Algorithm Optimization for Financial Crisis Dynamic Early Warning Model
Huang et al. An enhanced approach to optimizing the stock portfolio selection based on Modified Markowitz MV Method
CN113077189A (en) Method and device for evaluating life cycle of small and micro enterprise
Kunasekaran Research on E-commerce Customer Loyalty under Big Data
CN110610420A (en) Stock price trend prediction method and system
Bílková Assessment of Living Standard Indicators in the OECD Member Countries

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20200605