CN112419046A - Financial data automatic early warning system under artificial intelligence model - Google Patents
Financial data automatic early warning system under artificial intelligence model Download PDFInfo
- Publication number
- CN112419046A CN112419046A CN202011348913.4A CN202011348913A CN112419046A CN 112419046 A CN112419046 A CN 112419046A CN 202011348913 A CN202011348913 A CN 202011348913A CN 112419046 A CN112419046 A CN 112419046A
- Authority
- CN
- China
- Prior art keywords
- data processing
- financial
- financial data
- local data
- artificial intelligence
- 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.)
- Withdrawn
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/02—Banking, e.g. interest calculation or account maintenance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Probability & Statistics with Applications (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- General Business, Economics & Management (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention provides a financial data automatic early warning system under an artificial intelligence model, which is arranged at each local data processing center in a financial data automatic identification system and adopts various artificial intelligence early warning models to carry out risk identification on financial data collected by the local data processing centers; the artificial intelligence early warning model adopts a limited Boltzmann machine to mine the tag data characteristics of the financial data, and establishes a deep confidence network through a classified and partitioned limited Boltzmann mechanism to carry out risk identification on the financial data; the deep belief network comprises an unsupervised learning belief network and a supervised learning belief network with weight adjustment; and sending out an early warning signal to the local data processing center based on the risk identification result. The technical scheme of the invention can collect the financial data in time and give risk feedback in respective broadcast range of the local data center, and meanwhile, the normal use of the financial data terminal is not influenced.
Description
Technical Field
The invention belongs to the technical field of financial data processing, and particularly relates to an automatic financial data early warning system under an artificial intelligence model.
Background
The internet finance (ITFIN) is an organic combination of internet technology and financial functions, depends on a functional financial state and a service system thereof formed on an open internet platform by big data and cloud computing, comprises a financial market system, a financial service system, a financial organization system, a financial product system, an internet financial supervision system and the like based on a network platform, and has financial modes of general finance, platform finance, information finance, fragment finance and the like different from the traditional finance.
The internet financial terminal refers to a platform terminal that sells financial products and provides third-party services for financial product sales using the internet. The internet financial platform terminal is diversified and innovated, forms a third-party financial institution for providing high-end financial investment service and financial products, and an insurance portal website for providing insurance product consultation, price comparison and purchase service, and the like.
However, the information asymmetry problem is aggravated by financial technological innovation, and as financial institutions analyze and process financial data and risks through artificial intelligence and machine learning, very little is known by supervisors, and identification and response to financial risks become sluggish. And if the decision of the financial institution is more sensitive to the data, the economy runs faster, the periodic behavior is strengthened, the financial stability is not facilitated, and the systematic risk is more serious. In addition, most internet financial platforms do not have access to a people bank credit investigation system, do not have a credit information sharing mechanism, do not have wind control, compliance and clearing mechanisms similar to banks, and are easy to cause various risk problems.
Therefore, according to the regulations of the existing laws and regulations, financial institutions need to report their financial data to the monitoring institution regularly, so that the monitoring institution can grasp the attributes of the financial data in time, perform financial risk early warning in advance, and give feedback (monitoring) opinions.
The Chinese invention patent application with the application number of CN202010749615.X provides a wind control analysis early warning system based on financial big data characteristics, the wind control analysis early warning system comprises a temporary database Tep-DB and a service database Bus-DB, the temporary database Tep-DB receives and executes a service request instruction Op-C, the service database Bus-DB synchronously receives the service request instruction Op-C identical to the temporary database Tep-DB, and after the temporary database Tep-DB finishes verification operation, the service database Bus-DB executes the service request instruction Op-C identical to the temporary database Tep-DB so as to finish the execution operation of the service request instruction Op-C. The invention does not influence the efficiency of data wind control analysis early warning under the condition of ensuring data safety; the temporary database and the service database are relatively isolated, the service request instruction is asynchronously executed, data wind control analysis and early warning are carried out through verification operation, the service database is executed after verification, only the sequence exists between the temporary database and the service database, the two sets of computing systems can be used, and computing resources are prevented from being contended for each other.
However, under the existing regulatory system in China, the financial regulatory agencies are managed in different areas. For a large number of existing internet financial innovation bodies, data should be reported to which monitoring organization and how to report specifically, laws only have principle provisions, and the prior art does not provide clear technical means; furthermore, in the environment of encouraging innovation, both to enhance regulation and to avoid impacting the market development of the subject of innovation, and in particular the user experience, the prior art does not present an effective solution in this respect.
Disclosure of Invention
In order to solve the technical problems, the invention provides an automatic financial data early warning system under an artificial intelligence model. The automatic financial data identification system comprises a plurality of local data processing centers; a plurality of mobile terminals exist in the broadcast range of each local data processing center, and the mobile terminals are provided with financial data generators; when the mobile terminal and the local data processing center are in a data exchangeable state, the mobile terminal sends the financial data generated by the financial data generator to the local data processing center and receives a feedback processing result from the local data processing center. The financial data automatic early warning system is arranged in each local data processing center, and risk identification is carried out on financial data collected by the local data processing centers by adopting various artificial intelligence early warning models. The technical scheme of the invention can collect the financial data in time and give risk feedback in respective broadcast range of the local data center, and meanwhile, the normal use of the financial data terminal is not influenced.
As a general illustration, the invention may include N data processing centers C1，C2，…，CNEach data processing center carries out remote data exchange with at least one mobile terminal;
wherein the N data processing centers { C1，C2，…，CNAre arranged in M different geographical locations G1，G2，…，GMAnd collecting financial data generated by the mobile terminal in different geographical ranges in time.
The utility model provides an automatic early warning system of financial data under artificial intelligence model which characterized in that: the financial data automatic early warning system is arranged in each local data processing center in the financial data automatic identification system, and risk identification is carried out on financial data collected by the local data processing centers by adopting various artificial intelligence early warning models; the financial data automatic identification system comprises a plurality of local data processing centers, and each local data processing center is provided with a broadcast range; at least two of the plurality of local data processing centers are contiguous in broadcast range; each local data processing center comprises at least one financial data processing engine, and the financial data processing engine adopts at least one artificial intelligence model to perform attribute identification on the financial data; the financial data processing engines corresponding to at least two local data processing centers which are adjacent on the broadcasting range are different; a plurality of mobile terminals are present within the broadcast range of each of the local data processing centers, the mobile terminals being configured with a financial data generator; when the mobile terminal and the local data processing center are in a data exchangeable state, the mobile terminal sends financial data generated by the financial data generator to the local data processing center and receives a feedback processing result from the local data processing center;
the artificial intelligence early warning model adopts a limited Boltzmann machine to mine the tag data characteristics of the financial data, and establishes a deep confidence network through a classified and partitioned limited Boltzmann mechanism to carry out risk identification on the financial data;
the deep belief network comprises an unsupervised learning belief network and a supervised learning belief network with weight adjustment;
and sending out an early warning signal to the local data processing center based on the risk identification result.
In the scheme, the deep belief network is realized based on a hierarchical Bayes deep belief network algorithm.
In the scheme, each financial data processing engine in the automatic financial data identification system is configured with a plurality of different artificial intelligence models;
the artificial intelligence models of the financial data processing engine configurations corresponding to at least two local data processing centers which are adjacent on the broadcasting range are not identical.
In the scheme, when at least one financial APP on the mobile terminal is in an inactive period, the mobile terminal sends financial data currently generated by the financial APP in the inactive period to the local data processing center, and receives a feedback processing result sent by the local data processing center for the financial data sent last time by the financial APP in the inactive period.
As a first advantage of the present invention, when at least one financial APP on the mobile terminal is in an inactive period, the mobile terminal sends the financial data currently generated by the financial APP in the inactive period to the local data processing center, and receives a feedback processing result sent by the local data processing center for the financial data sent last time by the financial APP in the inactive period.
As a further advantage of the present invention, when at least one financial APP on the mobile terminal is in an inactive period, the mobile terminal determines its own base station location, and determines a base station propagation range based on the base station location;
and based on the base station broadcast range, the mobile terminal sends the financial data to a target data processing center, wherein the target data processing center is a local data processing center of the plurality of local data processing centers, and the broadcast range of the local data processing center is overlapped with the base station propagation range.
As a more critical technical means associated with the above advantages, each of the financial data processing engines is configured with a plurality of different artificial intelligence models;
the artificial intelligence models of the financial data processing engine configurations corresponding to at least two local data processing centers which are adjacent on the broadcasting range are not identical.
In the above technical solution of the present invention, the Deep Belief Network (DBN) is a multi-level neural network that integrates deep learning and feature learning. The traditional neural network also tries to learn more profound features through a multi-level network structure, but the multi-level neural network is difficult to achieve a good effect through a simple gradient descent method training. The deep belief network better solves the problem by adopting a layer-by-layer unsupervised pre-training mechanism. The deep neural network model is composed of a plurality of layers of unsupervised limited Boltzmann machines and a layer of supervised BP neural network.
In the above technical solution of the present invention, the financial data is from a plurality of different types of financial APPs installed on the mobile terminal, and the active periods of the plurality of different types of financial APPs are different.
According to the technical scheme, the mobile terminal generating the financial data can automatically select the optimal local data processing center to report the financial data; moreover, the data transmission is carried out when the mobile terminal is in a data exchangeable state, so that the interference to the normal use time period of a user is avoided; moreover, different artificial intelligence models are adopted by different local data processing centers, and especially different artificial intelligence models are adopted by the local data processing centers with overlapped broadcasting ranges (basically equal to adjacent broadcasting ranges), so that the singleness of data feedback results can be avoided; finally, the invention adopts a restricted Boltzmann mechanism of classification and partition to build a deep confidence network for risk prediction, thereby avoiding the defects of the traditional method.
Further advantages of the invention will be apparent in the detailed description section in conjunction with the drawings attached hereto.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a diagram of the subject architecture of an automated financial data recognition system under an artificial intelligence model, in accordance with an embodiment of the present invention
FIG. 2 is a schematic diagram of the layout of the automatic identification system for financial data of FIG. 1 in a wide space
FIG. 3 is a schematic diagram of data communication between a mobile terminal and a local data processing center in the system of FIG. 1 or 2
FIG. 4 is a diagram of the subject architecture of an automated early warning system for financial data under an artificial intelligence model according to an embodiment of the invention
FIG. 5 is a flow chart of a learning algorithm based on a class-partition restricted Boltzmann machine used by the system of FIG. 4
FIG. 6 is a schematic diagram of a mobile terminal used in the embodiment described in FIGS. 1-4
Detailed Description
The invention is further described with reference to the following drawings and detailed description.
First, as a general description of each embodiment of the present invention, the local data processing center mentioned in each embodiment of the present invention may be a financial monitoring center or an institution legally set, such as a local data monitoring processing center legally set in each jurisdiction city (geographical location) of a monitoring institution such as a people bank, a bank insurance policy, or a certificate authority; each geographical location represents the jurisdiction city of the financial administration center or institution or its affiliates, and the administrative region or other jurisdiction within a predetermined range centered on the geographical location may be referred to as the broadcast range of the geographical location in the present invention.
As a general example, the N local data processing centers { C }1,C2,…,CNAre arranged in M different geographical locations G1,G2,…,GM}; for example, multiple local data processing centers are located in multiple different city centers.
In the general example described above, at least one data processing center is provided for each geographic location. I.e., N > M, two or more data processing centers may be located in a single geographic location.
Obviously, N and M are both positive integers.
The embodiments shown in each of the figures will be described in detail below with reference to the figures.
Fig. 1 is a diagram of a subject architecture of an automatic financial data recognition system under an artificial intelligence model according to an embodiment of the present invention.
In fig. 1, the automatic financial data recognition system includes a plurality of local data processing centers,
each local data processing center comprises at least one financial data processing engine, and the financial data processing engine adopts at least one artificial intelligence model to perform attribute identification on the financial data.
In fig. 1, the local data processing center includes two financial data processing engines 01 and 02, each of which performs attribute recognition using 2-person intelligence models.
In fig. 1, the artificial intelligence models used by the financial data processing engines 01 and 02, respectively, are not identical, and the financial data processing engine 01 uses an artificial intelligence model a and an artificial intelligence model B; the financial data processing engine 02 uses the artificial intelligence model B and the artificial intelligence model C, and by means of the arrangement, at least two processing results can be obtained for the same financial data set by the same local data center, and meanwhile, the two processing results have relevance and are convenient to compare and analyze.
In the embodiment of FIG. 1, each local data processing center is provided with a broadcast range;
due to the contiguous presence of partial regulatory regions, wherein at least two of the plurality of local data processing centers are contiguous in broadcast range; the financial data processing engines corresponding to at least two local data processing centers adjacent on the broadcasting range are different.
The configuration also enables that when two target data processing centers may exist in the same mobile terminal, the two target data processing centers may be adjacent to each other in the broadcast range, in this case, two local data centers adjacent to each other in the broadcast range will obtain two processing results for the same financial data set, and in order to ensure that the analysis and supervision are not repeated and to facilitate the comparative analysis, the financial data processing engines corresponding to at least two local data processing centers adjacent to each other in the broadcast range are different.
Further, fig. 2 shows a schematic diagram of a larger layout.
In general, there must be tens of thousands of active mobile terminals within the broadcast range of each geographic location (local data processing center), and only one representative symbol is shown in fig. 2, not to indicate only this mobile terminal, as is the case in other figures or embodiments.
There are a plurality of mobile terminals within the broadcast range of each of the local data processing centers, the mobile terminals being configured with a financial data generator.
When the mobile terminal and the local data processing center are in a data exchangeable state, the mobile terminal sends the financial data generated by the financial data generator to the local data processing center and receives a feedback processing result from the local data processing center.
As a more specific example, when the mobile terminal itself is in a data exchangeable state, the mobile terminal transmits financial data generated by the financial data generator in a previous processing cycle to a target data processing center, and acquires a feedback result of the previously transmitted previous financial data from the target data processing center.
More specifically, referring to fig. 6, the financial data generator includes a plurality of different types of financial APPs installed on the mobile terminal, and the different types of financial APPs have different active periods.
Regarding the determination of the target local data center, under the schematic diagram of fig. 2, when at least one financial APP on the mobile terminal is in an inactive period, the mobile terminal determines its own base station location, and determines a base station propagation range based on the base station location;
and based on the base station broadcast range, the mobile terminal sends the financial data to a target data processing center, wherein the target data processing center is a local data processing center of the plurality of local data processing centers, and the broadcast range of the local data processing center is overlapped with the base station propagation range.
For each mobile terminal, the mobile terminal itself is in a data exchangeable state, including: at least one of said financial APPs (financial applications) is not currently in a period of high frequency use.
When a certain financial APP is in a non-high-frequency use period, the data exchange is carried out, so that the timeliness of data uploading (for example, the data is uploaded once every week or every day because the non-high-frequency use period is inevitably existed) is ensured, the data transmission blockage can be avoided, meanwhile, the data exchange is carried out under the non-inductive condition of a user, and the actual use of the user is not influenced.
Particularly for the case of fig. 3, the plurality of mobile terminals may transmit financial data generated by different financial applications to the same target data center at different time periods, or may transmit financial data generated by the same financial application to the same target data at the same time period.
Obviously, the local data processing center, especially the processing mode of the latter, can collect data of the same financial application in a centralized manner, which facilitates subsequent rapid centralized processing, because the non-high frequency use periods of all different mobile terminals for the same financial application are basically the same.
In fig. 3, each of the local data processing centers receives financial data transmitted from a plurality of mobile terminals within its broadcast range, and transmits a feedback processing result of the previously received financial data to the mobile terminals, and more particularly, to a financial APP that generates corresponding financial data.
In combination with the foregoing description, the specific operations include: when at least one financial APP on the mobile terminal is in an inactive period, the mobile terminal sends the financial data currently generated by the financial APP in the inactive period to the local data processing center, and receives a feedback processing result sent by the local data processing center for the financial data sent last time by the financial APP in the inactive period.
It should be noted that the feedback process for the financial data includes a plurality of processes, and the risk identification and the early warning are one of the most important feedback processes.
Reference is next made to fig. 4.
Fig. 4 shows an automatic financial data early warning system under an artificial intelligence model, where the automatic financial data early warning system is arranged in each local data processing center in the automatic financial data identification system in fig. 2, and multiple artificial intelligence early warning models are used to perform risk identification on financial data collected by the local data processing centers.
More specifically, in the embodiment shown in fig. 4, the artificial intelligence early warning model adopts a limited boltzmann machine to mine the tag data features of the financial data, and establishes a deep confidence network through a classified and partitioned limited boltzmann mechanism to perform risk identification on the financial data;
the deep belief network comprises an unsupervised learning belief network and a supervised learning belief network with weight adjustment;
sending an early warning signal to the local data processing center based on the risk identification result;
and sending the early warning signal to an operator of a financial APP generating the financial data.
In the embodiment of fig. 4, the deep belief network is composed of multiple layers of unsupervised constrained boltzmann machines and one layer of supervised neural networks.
The deep belief network is developed on the basis of the traditional neural network, and solves the problem that the depth of the hidden layer of the traditional neural network is limited and cannot be expanded. And training the limited Boltzmann machine layer by layer through a gradient descent greedy algorithm to obtain an interlayer weight W, a visible layer bias c and an implicit layer bias b. In the training process, unsupervised learning is used, and feature representation of input data is mainly extracted. And (3) carrying out fine adjustment on system parameters by using a BP algorithm on the basis of training a laminated limited Boltzmann machine.
As an improvement of the present invention, when solving the classification problem based on the constrained Boltzmann machine, the constrained Boltzmann machine functions as a feature extractor. Firstly, training a limited Boltzmann machine by using unlabeled training data, and then performing supervised learning on the training data by using other algorithms (such as a BP algorithm in a DBN).
In this example, in two learning steps of the deep belief network, the first stage information is generated from bottom to top, and the second stage information is generated from top to bottom.
More specifically, the flow of the learning algorithm based on the classification partition limited boltzmann machine in the embodiment illustrated in fig. 4 is shown in fig. 5.
For the restricted boltzmann machine contrast divergence learning model used in this embodiment, the model has an input layer (visible layer), two hidden layers and an output layer, where X is input sample data and Y is a sample label.
Fig. 5 generates a classification partition penalty term Q according to the label data y, wherein Q simultaneously affects the connection weight w between the visible layer and the hidden layer unit and the bias c of the visible layer, without changing the b-setting function of the hidden layer. The meaning of other parameters not labeled in fig. 5 follows a general understanding of the art.
It should be noted that in fig. 5, the initial purpose of the classification partition is to increase the uncertainty in the training process, and to give different weight penalties to different data sets, so that the weights have different effects. The training process of the classification partition limited Boltzmann machine must use small batch data (Mini-batch) to train a single sample, and aims to balance the effects of different punishment items so as not to cause system non-convergence. The classification partition vector is determined during system initialization, so a constant penalty term is relatively applied to the weight in the learning error, and although the penalty vector obeys Gaussian distribution, the independence among hidden layer units is not influenced, and more uncertainty is not added to system convergence. The penalty vector is set to be a constant, mainly because the limited Boltzmann machine is better than the Boltzmann machine algorithm, namely because the limiting interlayer units are not connected, the intra-layer correlation is reduced, and if the penalty vector is defined as a variable, the complexity of the system algorithm is greatly increased, and even the performance is reduced.
In addition, the fact li of the invention aims at the multi-task prediction problem that the input samples are the same and the output is different continuous variables, and the supervised learning prediction capability can be improved by adopting the hierarchical Bayesian depth confidence network algorithm.
Referring to fig. 6, the mobile terminal for sending the financial data includes an android or IOS or blackberry system, and a plurality of financial applications 01-03 that can be installed under different systems.
Simulation proves that the mobile terminal generating the financial data can automatically select the optimal local data processing center to report the financial data; moreover, the data transmission is carried out when the mobile terminal is in a data exchangeable state, so that the interference to the normal use time period of a user is avoided; moreover, different artificial intelligence models are adopted by different local data processing centers, and especially different artificial intelligence models are adopted by the local data processing centers with overlapped broadcasting ranges (basically equal to adjacent broadcasting ranges), so that the singleness of data feedback results can be avoided; finally, the invention adopts a restricted Boltzmann mechanism of classification and partition to build a deep confidence network for risk prediction, thereby avoiding the defects of the traditional method.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (4)
1. The utility model provides an automatic early warning system of financial data under artificial intelligence model which characterized in that: the financial data automatic early warning system is arranged in each local data processing center in the financial data automatic identification system, and risk identification is carried out on financial data collected by the local data processing centers by adopting various artificial intelligence early warning models; the financial data automatic identification system comprises a plurality of local data processing centers, and each local data processing center is provided with a broadcast range; at least two of the plurality of local data processing centers are contiguous in broadcast range; each local data processing center comprises at least one financial data processing engine, and the financial data processing engine adopts at least one artificial intelligence model to perform attribute identification on the financial data; the financial data processing engines corresponding to at least two local data processing centers which are adjacent on the broadcasting range are different; a plurality of mobile terminals are present within the broadcast range of each of the local data processing centers, the mobile terminals being configured with a financial data generator; when the mobile terminal and the local data processing center are in a data exchangeable state, the mobile terminal sends financial data generated by the financial data generator to the local data processing center and receives a feedback processing result from the local data processing center;
the artificial intelligence early warning model adopts a limited Boltzmann machine to mine the tag data characteristics of the financial data, and establishes a deep confidence network through a classified and partitioned limited Boltzmann mechanism to carry out risk identification on the financial data;
the deep belief network comprises an unsupervised learning belief network and a supervised learning belief network with weight adjustment;
and sending out an early warning signal to the local data processing center based on the risk identification result.
2. The system of claim 1, wherein the system comprises:
the deep confidence network is realized based on a hierarchical Bayesian deep confidence network algorithm.
3. The system of claim 1, wherein the system comprises:
each financial data processing engine in the automatic financial data identification system is configured with a plurality of different artificial intelligence models;
the artificial intelligence models of the financial data processing engine configurations corresponding to at least two local data processing centers which are adjacent on the broadcasting range are not identical.
4. The system of claim 1, wherein the system comprises:
when at least one financial APP on the mobile terminal is in an inactive period, the mobile terminal sends the financial data currently generated by the financial APP in the inactive period to the local data processing center, and receives a feedback processing result sent by the local data processing center for the financial data sent last time by the financial APP in the inactive period.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011348913.4A CN112419046A (en) | 2020-11-26 | 2020-11-26 | Financial data automatic early warning system under artificial intelligence model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011348913.4A CN112419046A (en) | 2020-11-26 | 2020-11-26 | Financial data automatic early warning system under artificial intelligence model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112419046A true CN112419046A (en) | 2021-02-26 |
Family
ID=74842562
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011348913.4A Withdrawn CN112419046A (en) | 2020-11-26 | 2020-11-26 | Financial data automatic early warning system under artificial intelligence model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112419046A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113934760A (en) * | 2021-10-15 | 2022-01-14 | 珠海百丰网络科技有限公司 | Financial data identification and transmission system and method based on artificial intelligence model |
-
2020
- 2020-11-26 CN CN202011348913.4A patent/CN112419046A/en not_active Withdrawn
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113934760A (en) * | 2021-10-15 | 2022-01-14 | 珠海百丰网络科技有限公司 | Financial data identification and transmission system and method based on artificial intelligence model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220076165A1 (en) | Systems and methods for automating data science machine learning analytical workflows | |
| US11710033B2 (en) | Unsupervised machine learning system to automate functions on a graph structure | |
| Yu et al. | Prediction of highway tunnel pavement performance based on digital twin and multiple time series stacking | |
| Tekouabou | Intelligent management of bike sharing in smart cities using machine learning and Internet of Things | |
| US20190378049A1 (en) | Ensemble of machine learning engines coupled to a graph structure that spreads heat | |
| CN102567391B (en) | Method and device for building classification forecasting mixed model | |
| Hashemi et al. | End‐to‐end deep learning methodology for real‐time traffic network management | |
| CN107169598A (en) | A kind of weather prediction method and system based on deep learning | |
| Feng et al. | [Retracted] Design and Simulation of Human Resource Allocation Model Based on Double‐Cycle Neural Network | |
| CN112529261A (en) | Financial data automatic identification and early warning system under artificial intelligence model | |
| CN108549907A (en) | A kind of data verification method based on multi-source transfer learning | |
| Liu et al. | A systematic procurement supply chain optimization technique based on industrial internet of things and application | |
| CN108734567A (en) | A kind of asset management system and its appraisal procedure based on big data artificial intelligence air control | |
| Fang et al. | Cdgnet: A cross-time dynamic graph-based deep learning model for traffic forecasting | |
| Jin et al. | A visual analytics system for improving attention-based traffic forecasting models | |
| CN112419046A (en) | Financial data automatic early warning system under artificial intelligence model | |
| Liu et al. | Manufacturing Process Optimization in the Process Industry | |
| Liu et al. | Fuzzy neural networks for machine maintenance in mass transit railway system | |
| Reddy et al. | A Beginner's Guide to Federated Learning | |
| Chen et al. | Research on Life‐Cycle Project Cost Management Based on Random Matrix Weight Algorithm | |
| Yang | Intelligent informatization early warning analysis of agricultural economy based on support vector sequential regression model | |
| Kang et al. | Performance Evaluation Model of Agricultural Enterprise Technology Innovation Based on GA‐BP Neural Network | |
| Ma et al. | MetaSTNet: Multimodal Meta-learning for Cellular Traffic Conformal Prediction | |
| US11734157B2 (en) | Systems with software engines configured for detection of high impact scenarios with machine learning-based simulation and methods of use thereof | |
| Hu | Intelligent Financial Decision Support System Using Artificial Intelligence |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20210226 |
|
| WW01 | Invention patent application withdrawn after publication |