CN117290505A - Method and device for determining flow condition of virtual resource and electronic device - Google Patents

Method and device for determining flow condition of virtual resource and electronic device Download PDF

Info

Publication number
CN117290505A
CN117290505A CN202311267730.3A CN202311267730A CN117290505A CN 117290505 A CN117290505 A CN 117290505A CN 202311267730 A CN202311267730 A CN 202311267730A CN 117290505 A CN117290505 A CN 117290505A
Authority
CN
China
Prior art keywords
data
transaction data
transaction
sample data
model
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
CN202311267730.3A
Other languages
Chinese (zh)
Inventor
乔杰
霍晓梅
常征
张航
初莹莹
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.)
China Construction Bank Corp
Original Assignee
China Construction Bank Corp
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 China Construction Bank Corp filed Critical China Construction Bank Corp
Priority to CN202311267730.3A priority Critical patent/CN117290505A/en
Publication of CN117290505A publication Critical patent/CN117290505A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/22—Indexing; Data structures therefor; Storage structures
    • G06F16/2282—Tablespace storage structures; Management thereof
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/38—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/205—Parsing
    • G06F40/216—Parsing using statistical methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
    • 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
    • G06N3/0455—Auto-encoder networks; Encoder-decoder networks
    • 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
    • G06Q40/12—Accounting
    • G06Q40/125—Finance or payroll
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/27—Regression, e.g. linear or logistic regression
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/01—Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Business, Economics & Management (AREA)
  • Databases & Information Systems (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • Development Economics (AREA)
  • Marketing (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Biophysics (AREA)
  • Library & Information Science (AREA)
  • Biomedical Technology (AREA)
  • Economics (AREA)
  • Molecular Biology (AREA)
  • Strategic Management (AREA)
  • Technology Law (AREA)
  • General Business, Economics & Management (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a method and a device for determining the flow condition of virtual resources and an electronic device, wherein the method comprises the following steps: in the case that the usage authorization of the target object is determined to have been acquired, acquiring transaction data of the target object, wherein the usage authorization is used for indicating that the target object allows the transaction data to be acquired; establishing a purpose tag library of the transaction data based on the purpose of the transaction data; establishing a purpose classification model of the transaction data according to the purpose tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library; and classifying the use of the transaction data according to the use classification model, and inputting the transaction data subjected to use classification into the fund flow model to determine the flowing condition of the virtual resource corresponding to the target object. The problems that the acquisition delay exists and the acquired fund flow report is inaccurate when the mechanism acquires the fund flow report of the public client are solved.

Description

Method and device for determining flow condition of virtual resource and electronic device
Technical Field
The invention relates to the field of finance, in particular to a method and a device for determining the flow condition of virtual resources and an electronic device.
Background
In the big data age, the financial industry relies on various wind control or marketing models to help it reduce loan (legal) violation rates or to enable accurate marketing. In public business, the business condition of enterprises is an important consideration factor, but the financial statement acquisition for measuring the business condition of enterprises is not easy: on the one hand, financial reports of marketing companies, large enterprises or administrative institutions have delay, are usually released in quarters or year, and meanwhile, the financial reports released by the enterprises have false phenomena, and the marketing companies who explode mines in recent years have fictitious financial reporting behaviors. On the other hand, over 90% of national enterprises are small and medium-sized micro-enterprises, and the small and medium-sized micro-enterprises hardly issue financial reports to the outside. These all greatly increase the difficulty of the financial institution in grasping the business operation.
Specifically, there are two traditional ways to obtain enterprise financial reports:
first, enterprises such as marketing companies actively disclose financial reports of the last quarter or year. Under the condition, the financial statement acquired by the organization has delay of a few months or even a year on one hand, and on the other hand, the accuracy of the financial statement data disclosed by the enterprise cannot be completely guaranteed.
Secondly, enterprises actively provide financial statements to institutions due to business needs. In this way, the organization can only obtain the financial statement of the enterprise at that time when the enterprise applies for the related business, and when the business is completed (for example, after loan is issued), the organization cannot obtain the latest financial statement of the enterprise to continuously follow up the business condition of the enterprise, so that the organization cannot control the subsequent risk. In addition, the mechanism can only acquire financial reports of part of enterprises in the mode, and cannot cover the full quantity of clients of the mechanism, so that the mechanism is blocked for digital wind control and marketing.
Aiming at the problems that in the related art, when an organization acquires a fund flow report of a public client, the acquisition delay exists, the acquired fund flow report is inaccurate and the like, no effective solution is proposed yet.
Disclosure of Invention
The embodiment of the invention provides a method and a device for determining the flowing condition of virtual resources and an electronic device, which at least solve the problems that in the related art, when an organization acquires a fund flow report of a public customer, the acquisition delay exists, the acquired fund flow report is inaccurate and the like.
According to an aspect of the embodiment of the present invention, there is provided a method for determining a flow condition of a virtual resource, including: obtaining transaction data of a target object in the case that the use authorization of the target object is determined to be obtained, wherein the use authorization is used for indicating that the target object allows the transaction data to be obtained; establishing a purpose tag library of the transaction data based on the purpose of the transaction data; establishing a purpose classification model of the transaction data according to the purpose tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library; and classifying the use of the transaction data according to the use classification model, and inputting the transaction data subjected to use classification into the fund flow model to determine the flow condition of the virtual resource corresponding to the target object.
In one exemplary embodiment, building a usage classification model of the transaction data from the usage tag library and the transaction data comprises: determining sample data corresponding to the transaction data based on the application tag library, wherein the sample data is transaction data with application tags; performing feature construction operation on the sample data to obtain feature data corresponding to the sample data, wherein the feature data comprises at least one of the following: text feature vector representation, target object attribute features, and transaction business type features; training a classification model by the sample data and the feature data to obtain the usage classification model.
In an exemplary embodiment, determining sample data corresponding to the transaction data based on the usage tag library includes: extracting text feature data in the transaction data, wherein the text feature data at least comprises: summary data, remark data, target object industry, target object name; constructing a tag rule model according to the text feature data and the application tag library; and inputting the transaction data into the label rule model to determine sample data corresponding to the transaction data.
In an exemplary embodiment, performing a feature construction operation on the sample data to obtain feature data corresponding to the sample data, including: determining text feature data in the sample data and application labels corresponding to the text feature data; training to obtain a first sub-model based on the text feature data and the application label; the sample data is input into the first sub-model to obtain a text feature vector representation of the sample data.
In an exemplary embodiment, performing a feature construction operation on the sample data to obtain feature data corresponding to the sample data, including: determining all categories of text feature data in the sample data; determining first sample data and second sample data in the sample data based on the all categories, the first sample data being sample data having text feature data of all categories, the second sample data being sample data not having text feature data of all categories; training based on the first sample data to obtain a second sub-model; filling the second sample data through a second sub-model to obtain filling data corresponding to the second sample data; and extracting keywords of the sample data based on the types of the target objects, and determining the keywords and the filling data as target object attribute characteristics of the sample data, wherein the types are included in text characteristic data in the first sample data and the filling data.
In an exemplary embodiment, performing a feature construction operation on the sample data to obtain feature data corresponding to the sample data, including: determining a business classification label corresponding to the transaction data, and determining a labeling rule of the transaction data according to an initiating component of the transaction data; and classifying the sample data according to the labeling rules and the service classification labels to obtain transaction service type characteristics corresponding to the sample data.
In an exemplary embodiment, establishing a fund flow model corresponding to transaction data according to the usage tag library includes: determining a system category to which the target object belongs, and determining a balance type label indicated by a preset rule corresponding to the system category; constructing a fund flow label system based on the system category and the expense type label; dividing the application tag library into the fund flow tag system according to the preset rule so as to establish a fund flow model corresponding to the transaction data.
According to another aspect of the embodiment of the present invention, there is also provided a device for determining a flow condition of a virtual resource, including: an acquisition module, configured to acquire transaction data of a target object if it is determined that a usage authorization of the target object has been acquired, where the usage authorization is used to instruct the target object to allow the transaction data to be acquired; the first establishing module is used for establishing a purpose tag library of the transaction data based on the purpose of the transaction data; the second building module is used for building a purpose classification model of the transaction data according to the purpose tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library; and the classification module is used for classifying the use of the transaction data according to the use classification model, and inputting the transaction data subjected to use classification into the fund flow model so as to determine the flow condition of the virtual resource corresponding to the target object.
According to another aspect of the embodiment of the present invention, there is further provided a computer readable storage medium, where the computer readable storage medium includes a stored program, where the program runs the method for determining a flow condition of a virtual resource.
According to still another aspect of the embodiments of the present invention, there is further provided an electronic device, including a memory and a processor, where the memory stores a computer program, and the processor runs the method for determining the flow condition of the virtual resource by using the computer program.
In the embodiment of the invention, under the condition that the usage authorization of the target object is determined to be acquired, acquiring the transaction data of the target object, wherein the usage authorization is used for indicating that the target object allows the transaction data to be acquired; establishing a purpose tag library of the transaction data based on the purpose of the transaction data; establishing a purpose classification model of the transaction data according to the purpose tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library; and classifying the use of the transaction data according to the use classification model, and inputting the transaction data subjected to use classification into the fund flow model to determine the flow condition of the virtual resource corresponding to the target object. That is, when the usage authorization of the transaction data by the target object is acquired, the usage classification model and the fund flow model are constructed based on the transaction data, and the flow condition of the virtual resource corresponding to the transaction data is determined based on the usage classification model and the fund flow model. The method solves the problems that in the related art, when an institution acquires a fund flow report of a public client, the acquired fund flow report is delayed and inaccurate, and further accurately determines the flowing condition of virtual resources corresponding to a target object based on transaction data under the condition that the use authorization of the target object is obtained.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention and do not constitute a limitation on the invention. In the drawings:
FIG. 1 is a block diagram of a hardware architecture of a computer terminal of a method for determining a flow condition of a virtual resource according to an embodiment of the present invention;
FIG. 2 is a flow chart of an alternative method of determining the flow condition of virtual resources according to an embodiment of the invention;
FIG. 3 is a flow diagram illustration (one) of an alternative method of determining the flow condition of virtual resources according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of an architecture for creating a usage classification model for a flow scenario of an alternative virtual resource according to an embodiment of the present invention;
FIG. 5 is a flow diagram view (II) of an alternative method of determining the flow condition of virtual resources according to an embodiment of the present invention;
fig. 6 is a block diagram of a determining apparatus for a flow condition of a virtual resource according to an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
The method embodiment provided by the embodiment of the invention can be operated in a computer terminal. Taking the operation on a computer terminal as an example, fig. 1 is a block diagram of a hardware structure of a computer terminal according to a method for determining a flow condition of a virtual resource according to an embodiment of the present invention. As shown in fig. 1, the computer terminal may include one or more (only one is shown in fig. 1) processors 102 (the processor 102 may include, but is not limited to, a processing system such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, and in one exemplary embodiment, may also include a transmission device 106 for communication functions and an input-output device 108. It will be appreciated by those skilled in the art that the configuration shown in fig. 1 is merely illustrative and is not intended to limit the configuration of the computer terminal described above. For example, the image pickup apparatus may further include more or less components than those shown in fig. 1, or have a different configuration equivalent to the functions shown in fig. 1 or more than the functions shown in fig. 1.
The memory 104 may be used to store a computer program, for example, a software program of application software and a module, such as a computer program corresponding to a method for determining a flow condition of a virtual resource in an embodiment of the present invention, and the processor 102 executes various functional applications and data processing by executing the computer program stored in the memory 104, that is, implements the method described above. Memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage systems, flash memory, or other non-volatile solid-state memory. In some examples, memory 104 may further include memory remotely located with respect to processor 102, which may be connected to secure text via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The transmission device 106 is used to receive or transmit data via a network. The specific example of the network described above may include a wireless network provided by a communication provider of the image capturing apparatus. In one example, the transmission system 106 includes a network adapter (Network Interface Controller, simply referred to as NIC) that can connect to other network devices through a base station to communicate with the internet.
Technical terms adopted by the embodiments of the present invention are explained below:
1) The institution pair male clients (equivalent to the target object in the embodiment of the present invention): refers to non-individual clients with business in institutions, including marketing companies, large, medium and small enterprises, individual business merchants, administrative institutions, organizations and the like.
2) The institutions are responsible for public service: refers to a business that an organization provides to a public customer.
3) Financial statement (corresponding to the flow of virtual resources in embodiments of the invention): accounting statements reflecting the fund and profit conditions of an enterprise or budget unit for a period of time, including balance tables, damage tables, fund flow tables or financial condition change tables, balance tables and notes.
In this embodiment, a method for determining a flow condition of a virtual resource is provided, including but not limited to being applied to the computer terminal described above, and fig. 2 is a flowchart of an alternative method for determining a flow condition of a virtual resource according to an embodiment of the present invention, where the flowchart includes the following steps:
step S202, in the case that the usage authorization of the target object is determined to be acquired, acquiring transaction data of the target object, wherein the usage authorization is used for indicating that the target object allows the transaction data to be acquired;
It should be noted that, in the embodiment of the present invention, all the data used, including the transaction data, are already authorized to be used by the target object.
It will be appreciated that the transaction data obtained is data that includes text feature information including, but not limited to: information such as transaction party names (including the names of target objects and the names of transaction objects generating transaction data with the target objects), transaction directions, transaction remarks, summaries, transaction channel types, clearing channel types, source components and the like. The transaction data existence range can be transaction data of all target objects of the service provided by the receiving mechanism in one day, one week and one month. The actual time range of acquiring the transaction data, or the data amount of the acquired transaction data, should be determined according to the actual situation, which is not necessarily limited by the embodiment of the present invention.
Step S204, a purpose tag library of the transaction data is established based on the purpose of the transaction data;
and determining the system category corresponding to the target object, such as: enterprises, administrative institutions, financial institutions and the like, and based on the system types, the general fund uses of transaction behaviors are carried out through the institutions, and a fund use tag library is built, wherein the use tag library is marked as T= { T 1 ,t 2 ,…,t p }, t is i The name of the ith funds use label is represented, and the labels are mutually exclusive.
Step S206, establishing a purpose classification model of the transaction data according to the purpose tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library;
and step S208, carrying out purpose classification on the transaction data according to the purpose classification model, and inputting the transaction data subjected to purpose classification into the fund flow model so as to determine the flow condition of the virtual resource corresponding to the target object.
Through the steps, under the condition that the usage authorization of the target object is determined to be acquired, acquiring the transaction data of the target object, wherein the usage authorization is used for indicating that the target object allows the transaction data to be acquired; establishing a purpose tag library of the transaction data based on the purpose of the transaction data; establishing a purpose classification model of the transaction data according to the purpose tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library; and classifying the use of the transaction data according to the use classification model, and inputting the transaction data subjected to use classification into the fund flow model to determine the flow condition of the virtual resource corresponding to the target object. That is, when the usage authorization of the transaction data by the target object is acquired, the usage classification model and the fund flow model are constructed based on the transaction data, and the flow condition of the virtual resource corresponding to the transaction data is determined based on the usage classification model and the fund flow model. The method solves the problems that in the related art, when an institution acquires a fund flow report of a public client, the acquired fund flow report is delayed and inaccurate, and further accurately determines the flowing condition of virtual resources corresponding to a target object based on transaction data under the condition that the use authorization of the target object is obtained.
In one exemplary embodiment, building a usage classification model of the transaction data from the usage tag library and the transaction data comprises: determining sample data corresponding to the transaction data based on the application tag library, wherein the sample data is transaction data with application tags; performing feature construction operation on the sample data to obtain feature data corresponding to the sample data, wherein the feature data comprises at least one of the following: text feature vector representation, target object attribute features, and transaction business type features; training a classification model by the sample data and the feature data to obtain the usage classification model.
Optionally, determining sample data corresponding to the transaction data based on the usage tag library includes: extracting text feature data in the transaction data, wherein the text feature data at least comprises: summary data, remark data, target object industry, target object name; constructing a tag rule model according to the text feature data and the application tag library; and inputting the transaction data into the label rule model to determine sample data corresponding to the transaction data.
For the usage tag library t= { T 1 ,t 2 ,…,t p A rule model (corresponding to the label rule model in the embodiment) is formulated according to abstract types, transaction remark keywords, customers (corresponding to target objects in the embodiment of the invention), customer names and the like in text feature data, and is recorded as R= { t 1 :rule 1 ,t 2 :rule 2 ,…,t n :rule n For meeting the condition rule i The transaction of which the fund uses is a label t i . For example, rule 1 : customer name A, customer industry 2, trade remark keyword XX, corresponding usage label is t 1 。
Classifying transaction data through a label rule model to meet rule 1 、rule 2 Etc. under the corresponding usage label. It is also understood that to satisfy rule 1 、rule 2 Etc. are labeled with corresponding usage labels.
In an exemplary embodiment, performing a feature construction operation on the sample data to obtain feature data corresponding to the sample data, including: determining text feature data in the sample data and application labels corresponding to the text feature data; training to obtain a first sub-model based on the text feature data and the application label; the sample data is input into the first sub-model to obtain a text feature vector representation of the sample data.
It should be noted that, the sample data may be understood as transaction data labeled with a usage tag, and further the sample data still has text feature data in the transaction data. Further, determining text feature data in the sample data and a usage label corresponding to the text feature data, including: determining each piece of text feature data and each usage label in each piece of sample data in the sample data combines all of each piece of text feature data into first combined data (corresponding to the text feature data in the above embodiment), and combines all of each piece of usage label into second combined data (corresponding to the usage label corresponding to the text feature data in the above embodiment).
Text pre-training models employed in the present application include, but are not limited to BERT, roBERTa, XLNet and the like. Training the text pre-training model through the text characteristic data and the application label to obtain a first sub-model.
The text feature vector output after the sample data is processed based on the first sub-model can be understood as: in the case where n pieces of sample data are included in the sample data, an n×m-dimensional fund use tag probability matrix (i.e., text feature vector) is output, where n is the number of pieces of sample data input and m is the number of fund use tag categories. I.e. the text features of each sample data are represented by vectors of dimension m.
In an exemplary embodiment, performing a feature construction operation on the sample data to obtain feature data corresponding to the sample data, including: determining all categories of text feature data in the sample data; determining first sample data and second sample data in the sample data based on the all categories, the first sample data being sample data having text feature data of all categories, the second sample data being sample data not having text feature data of all categories; training based on the first sample data to obtain a second sub-model; filling the second sample data through a second sub-model to obtain filling data corresponding to the second sample data; and extracting keywords of the sample data based on the types of the target objects, and determining the keywords and the filling data as target object attribute characteristics of the sample data, wherein the types are included in text characteristic data in the first sample data and the filling data.
It can be understood that when the target object performs transaction, only the name of the target object needs to be filled in, and other text information in the text feature data is selectively filled in, so that in the transaction data generated by the transaction acquired by the institution, the text information in the text feature data may be missing. Thus, the first sample data with complete text information and the second sample data with incomplete text information in the template data can be screened out based on all the categories of the text characteristic data. Specifically, all categories of text feature data in the sample data are first determined, including but not limited to: customer name, trade opponent customer name, customer industry, trade opponent customer industry, customer type, trade opponent customer type.
Furthermore, the second sample data of the missing text information is filled, and the filling mode can be filled by adopting methods of external data purchase, manual filling, machine learning classification and the like.
Taking machine learning for filling as an example, the first sample data is used as data for training a filling model to obtain a second sub-model, and the filling model can include, but is not limited to, a support vector machine, logistic regression, decision tree, and the like. The processing operation on the first sample data prior to training the filling model includes: for sample data deduplication processing, customer names (certain existence) in the sample data subjected to the deduplication processing are used as training features, the rest of text information in the text feature data, such as a customer type and the like, are used as labels, and the training features are mapped into vectors which can be used by a model through algorithms including but not limited to TF-IDF, word2Vec, gloVe, fastText and the like. It should be noted that, the second sample data may also need to be input into the second sub-model after the processing operation for the filling process.
The names of both transaction parties (namely the names of target objects or clients and the names of transaction opponents of target objects) contain important information related to the use of transaction funds, and the use of the transaction funds can be determined only by means of the names of both transaction parties, the type of transaction business and other information parties under the condition that the information such as remarks and abstracts of the transaction is lost.
And extracting keywords of the sample data based on the type of the target object, including:
determining the position of a name keyword according to the type of the target object; acquiring names of both transaction parties in the text feature data; preprocessing operation is carried out on names of both parties of the transaction to obtain preprocessed data, wherein the preprocessing operation sequentially comprises the following steps: removing special characters, splitting words, wherein the pretreatment data comprises a plurality of words of names of both transaction sides obtained after splitting words, and selecting words corresponding to the positions in the pretreatment data based on the positions; and combining the word segmentation into the keywords.
In an exemplary embodiment, performing a feature construction operation on the sample data to obtain feature data corresponding to the sample data, including: determining a business classification label corresponding to the transaction data, and determining a labeling rule of the transaction data according to an initiating component of the transaction data; and classifying the sample data according to the labeling rules and the service classification labels to obtain transaction service type characteristics corresponding to the sample data.
Traffic class labels include, but are not limited to, the following several broad categories: deposit and credit business, fund supervision business, peer business, proxy financial business, etc.; more specific divisions based on the several broad categories are also possible, and embodiments of the present invention are not limited in this regard.
Further, transactions of different services are typically completed by different transaction service components or data platforms through processes of initiation, invocation, recording, etc., so labeling rules can be constructed by, but not limited to, constructing from information of transaction service components, caller components, product types, etc.
And the sample data can be classified under a specific service classification label based on the labeling rule, and the service classification label corresponding to the sample data fragment can be understood.
In an exemplary embodiment, establishing a fund flow model corresponding to transaction data according to the usage tag library includes: determining a system category to which the target object belongs, and determining a balance type label indicated by a preset rule corresponding to the system category; constructing a fund flow label system based on the system category and the expense type label; dividing the application tag library into the fund flow tag system according to the preset rule so as to establish a fund flow model corresponding to the transaction data.
It will be appreciated that the system categories include: enterprises, administrative institutions and financial institutions, wherein the preset rule is an accounting rule used by the institutions, and then the established fund flow label system is as follows:
U i ＝{Fn:Re j }；
i e (business, administrative, financial institution),
fn e (business fund inflow, business fund outflow, financing fund inflow,
and (3) sending out the fund of the financing activity, sending in the fund of the investment activity, and sending out the fund of the investment activity).
Wherein U is i The system is suitable for i-type clients to obtain a fund flow label system, fn is the fund flow label of the fund flow label system, re j And a corresponding expense type label for the fund flow label.
And dividing the fund use label library under the balance type labels of the corresponding fund flow label system according to accounting rules and combining different customer industries and transaction directions, so as to obtain the fund flow model.
In order to better understand the above-mentioned determination scheme of the flow condition of the virtual resource, in an alternative embodiment, a scheme is also provided for explaining the above-mentioned scheme.
The invention aims at obtaining a fund flow report which is updated by the whole amount of an organization to public clients in time and has correct data and can reflect the operation condition of the public clients, and comprises 2 modules: the system comprises a public client transaction purpose classification module (a first module) and a fund flow report construction module (a second module). The module one builds a classification model for the transaction purpose of the public clients based on the real-time transaction data of the public clients by an organization and through text analysis and deep learning algorithm, and marks the real fund purpose of each transaction of the public clients. The second module gives different fund flow classification systems (namely, different fund uses belong to different fund flow items for different types of enterprises) based on industry accounting rules for different types of enterprises, and the result of the first module is summarized and calculated according to the fund flow classification systems to generate a fund flow report of institutions to public clients on a monthly, quarterly and annual basis.
Further, the overall flow of the present invention is shown in fig. 3, and specifically, the present invention may be implemented in two modules. The first module is a transaction purpose classification module, which aims to obtain the fund purpose of each transaction of a public customer. After the fund use is obtained, different fund flow reports can be obtained according to different enterprise types through a module two fund flow construction module.
The steps shown in fig. 3 are as follows:
step S301, data preparation;
step S302, constructing a label;
step S303, feature construction;
step S304, fund usage classification model construction (corresponding to the establishment of the usage classification model in the above embodiment);
step S305, cash flow model construction (corresponding to the establishment of the cash flow model in the above embodiment);
and step S306, outputting a cash flow report.
Specifically, the steps of acquiring the fund usage classification module for public client transaction are as follows, the construction structure diagram is shown in fig. 4, and the implementation process can be divided into four sub-processes of data preparation, label definition, feature construction and model training, specifically, the following steps 1-4 are as follows:
step 1: data preparation. The method is mainly used for trading data of public clients, including but not limited to trade party names, trade directions, trade remarks, abstracts, trade channel types, clearing channel types, source components and the like. The data range can take the total quantity of the mechanism or the sampling of the public transportation easy data in a certain day or a certain period of time, and the actual data time range or the data quantity is determined according to the actual situation without mandatory limitation.
Step 2: and (5) constructing a label. The process comprises label library construction, label rule model and sample label generation. The method comprises the following steps of:
step 21: and (5) constructing a tag library. Based on common fund uses of enterprises, administrative institutions, financial institutions and the like for carrying out transaction behaviors on public clients through institutions, a fund use tag library is built and is marked as T= { T 1 ,t 2 ,…,t p }, t is i The name of the ith funds use label is represented, and the labels are mutually exclusive.
Step 22: a label rule model. For T= { T 1 ,t 2 ,…,t p A rule model is formulated according to abstract types, transaction remark keywords, customer industries, customer names and the like and is recorded as R= { t }, wherein the rule model is formed according to abstract types, transaction remark keywords, customer industries, customer names and the like 1 :rule 1 ,t 2 :rule 2 ,…,t n :rule n For meeting the condition rule i The transaction of which the fund uses is a label t i 。
Step 23: and generating a sample label. The label rule model established in the step 22 is applied to the prepared data, the fund use label of the data hit by the label rule model is output, and the part of the data is used as sample data for subsequent feature construction and model training.
Step 3: and (5) feature construction. The part uses the sample data generated in the step 2, and is specifically divided into the following steps:
step 31: text feature vector representation. A pre-trained model is employed herein to vector the textual features of trade notes, customer names, trade adversary names, etc. The method comprises the following steps:
Step 311: data preparation. Based on the sample data generated in step 2, two types of data are prepared: the text feature data of the punctuation mark is deleted and is not empty in the sample data, and the text feature data can be a text spliced by a customer name, a transaction remark, a transaction opponent name and the like, or only the transaction remark can be used as a text feature. The other is the funds use label data in these sample data.
Step 312: and (5) model training. The text pre-training model used here includes, but is not limited to BERT, roBERTa, XLNet, etc., and the data prepared in step 311 is input into the pre-training model to train, thereby obtaining the fund usage classification model based on the transaction remarks.
Step 313: and (5) vector representation. And inputting n pieces of sample data needing fund use prediction into the fund use classification model obtained in the step 312 after processing, and outputting an n multiplied by m-dimensional fund use label probability matrix, wherein n is the number of input sample data pieces, and m is the number of fund use label categories. I.e. the text features of each sample data are represented by vectors of dimension m.
Step 32: and constructing the customer attribute characteristics. The client attribute features of the invention include, but are not limited to, client name, trade opponent client name, client industry, trade opponent client industry, client type, trade opponent client type, etc., and the construction method can be divided into deletion filling and keyword extraction according to different data characteristics. The method comprises the following steps:
Step 321: missing fills. Other customer attribute features besides customer names are not necessarily filled in institution transaction data, so transaction data lacking relevant information needs to be filled in. The data loss can be filled by adopting methods of external data purchase, manual filling, machine learning classification and the like. Taking the client type as an example, a filling method adopting machine learning classification is provided, and the filling method specifically comprises the following steps:
step 3211: data preparation. For data with the client type not missing, the client name is used as a text feature, and the client type is used as a label. And de-duplication processing is performed on the data, and algorithms including but not limited to TF-IDF, word2Vec, gloVe, fastText and the like are used for mapping text features into vectors for a model.
Step 3212: and (5) model training. The classification algorithms mainly used herein include, but are not limited to, support vector machines, logistic regression, decision trees, etc., and the data prepared in step 3211 is input into the classification algorithm for training, thereby obtaining a client type classification model.
Step 3213: model prediction. And processing the data with the missing client types by the step 3211, and inputting the client names into the model trained by the step 3212 to obtain the client types predicted by the model, namely finishing filling the missing client types.
Step 322: and extracting keywords. The trade name of both sides contains important information related to trade fund usage, and under the condition of missing information such as trade remarks, abstracts and the like, the trade fund usage can be determined only by means of the trade name of both sides, trade business type and other information sides, so that keyword extraction is considered to be carried out on the trade name of both sides according to the client type, and the classification effect of a final model is enhanced. The method comprises the following steps:
step 3221: acquiring name keyword positions, and storing different client types and the name keyword positions corresponding to the client types into a dictionary form: k= { t 1 :key 1 ,t 2 :key 2 ,…,t n :key n }. Where n is the number of categories of client type, t i For the ith customer type, key i And the name keyword position corresponding to the ith client type. Wherein the key is i May be a location interval, e.g. key i ＝[3,5]Then the 3 rd, 4 th and 5 th words are keywords, and the key i Also referred to as single word position, e.g. key i =2 means that the 2 nd word is a keyword.
Step 3222: removing special characters in names of both transaction sides, and then performing word segmentation, wherein each name is expressed as follows: c= { C 1 ,c 2 ,…,c m }, wherein c i And (3) referring to the ith word after name segmentation, wherein m represents the number of words obtained after name segmentation.
Step 3223: word c= { C for each name 1 ,c 2 ,…,c m First combine its client type t i Obtaining the corresponding keyword position key i Will fall to the key i And splicing the segmented words of the intervals to obtain the keywords corresponding to the names.
Step 33: and constructing transaction service type characteristics. The method comprises the following steps:
step 331: and determining the public traffic easy service classification labels according to the public traffic range of the institutions. Generally, public services can be classified into deposit and credit services, funds management services, peer services, proxy financial services, etc., and transaction service classification labels can be determined according to actual conditions of the institutions. The large number of complex institutional services may also be further subdivided into transaction service classification subclasses such as, without limitation, agent social security, agent financial tax, and the like.
Step 332: after determining the transaction service classification labels, constructing corresponding labeling rules for each label. Transactions for different services are typically completed by different transaction service components or data platforms through processes such as initiation, invocation, logging, etc., and thus labeling rules may be constructed from, but not limited to, information such as transaction service components, caller components, product types, etc.
Step 333: based on the labeling rules of step 332, a transaction traffic classification for each transaction of the male customer is obtained.
Step 4: and (5) model training. A training set is first constructed. And (3) dividing the characteristic data in the step (3) and the tag data in the step (2) into a plurality of training sets according to different transaction service types. These training sets are then separately input into a classification model for training, where the classification model includes, but is not limited to, a support vector machine, logistic regression, decision tree, etc. And finally outputting the trained model as a fund use classification model.
Further, alternative embodiments of the present invention may require a funding flow model construction,
optionally, the fund flow model building process is shown in the following steps 5-6, specifically:
step 5: and (5) constructing a fund flow label system. Dividing clients into enterprises, administrative enterprises and financial institutions, constructing a fund flow label system by referring to corresponding accounting rules, and marking as follows:
U i ＝{Fn:Re j },
i e (business, administrative, financial institution),
fn e (business fund inflow, business fund outflow, financing fund inflow,
and (3) sending out the fund of the financing activity, sending in the fund of the investment activity, and sending out the fund of the investment activity).
Wherein U is i Applicable funds for type i customers Stream label system, fn is its fund stream label, re j And a corresponding expense type label for the fund flow label.
Step 6: and (5) constructing a fund flow model. Dividing the fund use label library into the balance type labels of the corresponding fund flow label system according to accounting rules and combining the different and transaction directions of the client industry for the fund flow label system of enterprises, administrative enterprises and financial institutions constructed in the step 5, and marking as U i ＝{Fn:Re j :rule jm }, wherein rule jm Is Re (Re) j M decision rules of the balance type, namely, the fund flow model is constructed.
Based on the above description of the optional embodiment of the present invention, the fund usage classification module and the fund flow model have been obtained, so that a fund flow report for a public client can be obtained through the fund usage classification module and the fund flow model, and the flowchart is shown in fig. 5, and includes the following steps:
step S701, data preparation;
the method is mainly used for transaction data of public clients, including but not limited to information such as names of transaction parties, transaction directions, transaction remarks of transaction amounts, abstracts, transaction channel types, clearing channel types, source components and the like. The data range is transaction data of a public client for one month, one quarter, one year and the like, and the actual data time range or the data volume is determined according to the actual situation, so that the mechanism for calculating the fund flow report is not restricted.
Step S702, acquiring a fund use label through a fund use classification module;
step S703, acquiring a fund flow label and a balance type label through a fund flow model;
step S704, generating a funds flow report of the enterprise, the administrative institution and the financial institution.
Specifically, the prepared data is firstly input into a fund usage classification module for public client transaction to obtain a fund usage label of each transaction data. And inputting the data into a fund flow model to obtain fund flow labels and expense type labels corresponding to the fund use labels of transactions of different clients. And finally, summarizing the data according to the customer, the fund flow label and the balance type label to obtain the fund flow report of the public customer with different customer types.
The optional embodiment of the invention constructs a fund use classification module for public client transactions and a fund flow report module based on the real transaction data of the institutions, and finally obtains a fund flow report for the public clients
The financial statement can reflect the actual operation condition of the public clients based on the actual transaction data of the institutions, has the advantages of being real and reliable, wide in coverage and high in timeliness, solves the problems of delaying, low in coverage and fictional statement in the traditional method of acquiring the public client financial statement by the institutions, and has the advantages of being high in timeliness and real and reliable in public clients and data by the full-scale coverage mechanism.
From the description of the above embodiments, it will be clear to a person skilled in the art that the method according to the above embodiments may be implemented by means of software plus the necessary general hardware platform, but of course also by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) comprising instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to run the method of the embodiments of the present invention.
The embodiment also provides a device for determining the flow condition of the virtual resource, which is used for implementing the foregoing embodiment and the preferred implementation manner, and is not described in detail. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. While the devices described in the following embodiments are preferably implemented in software, implementations in hardware, or a combination of software and hardware, are also possible and contemplated.
FIG. 6 is a block diagram of a determination apparatus for flow conditions of virtual resources according to an embodiment of the present invention; as shown in fig. 6, includes:
an acquisition module 62 for acquiring transaction data of a target object in a case where it is determined that a usage authorization of the target object has been acquired, wherein the usage authorization is used to instruct the target object to allow the transaction data to be acquired;
a first establishing module 64 for establishing a usage tag library of the transaction data based on the usage of the transaction data;
a second building module 66 for building a usage classification model of the transaction data based on the usage tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library;
and the classification module 68 is configured to classify the transaction data according to the usage classification model, and input the transaction data after the usage classification to the fund flow model to determine the flow condition of the virtual resource corresponding to the target object.
By the device, under the condition that the usage authorization of the target object is acquired, transaction data of the target object is acquired, wherein the usage authorization is used for indicating the target object to allow the transaction data to be acquired; establishing a purpose tag library of the transaction data based on the purpose of the transaction data; establishing a purpose classification model of the transaction data according to the purpose tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library; and classifying the use of the transaction data according to the use classification model, and inputting the transaction data subjected to use classification into the fund flow model to determine the flow condition of the virtual resource corresponding to the target object. That is, when the usage authorization of the transaction data by the target object is acquired, the usage classification model and the fund flow model are constructed based on the transaction data, and the flow condition of the virtual resource corresponding to the transaction data is determined based on the usage classification model and the fund flow model. The method solves the problems that in the related art, when an institution acquires a fund flow report of a public client, the acquired fund flow report is delayed and inaccurate, and further accurately determines the flowing condition of virtual resources corresponding to a target object based on transaction data under the condition that the use authorization of the target object is obtained.
In an exemplary embodiment, the second establishing module 66 is further configured to determine sample data corresponding to the transaction data based on the usage tag library, where the sample data is transaction data having a usage tag; performing feature construction operation on the sample data to obtain feature data corresponding to the sample data, wherein the feature data comprises at least one of the following: text feature vector representation, target object attribute features, and transaction business type features; training a classification model by the sample data and the feature data to obtain the usage classification model.
In an exemplary embodiment, the second establishing module 66 is further configured to extract text feature data in the transaction data, where the text feature data includes at least: summary data, remark data, target object industry, target object name; constructing a tag rule model according to the text feature data and the application tag library; and inputting the transaction data into the label rule model to determine sample data corresponding to the transaction data.
In an exemplary embodiment, the second establishing module 66 is further configured to determine text feature data in the sample data and a usage label corresponding to the text feature data; training to obtain a first sub-model based on the text feature data and the application label; the sample data is input into the first sub-model to obtain a text feature vector representation of the sample data.
In one exemplary embodiment, the second building block 66 is further configured to determine all categories of text feature data in the sample data; determining first sample data and second sample data in the sample data based on the all categories, the first sample data being sample data having text feature data of all categories, the second sample data being sample data not having text feature data of all categories; training based on the first sample data to obtain a second sub-model; filling the second sample data through a second sub-model to obtain filling data corresponding to the second sample data; and extracting keywords of the sample data based on the types of the target objects, and determining the keywords and the filling data as target object attribute characteristics of the sample data, wherein the types are included in text characteristic data in the first sample data and the filling data.
In an exemplary embodiment, the second establishing module 66 is further configured to determine a service classification label corresponding to the transaction data, and determine a labeling rule of the transaction data according to an initiating component of the transaction data; and classifying the sample data according to the labeling rules and the service classification labels to obtain transaction service type characteristics corresponding to the sample data.
In an exemplary embodiment, the second establishing module 66 is further configured to determine a class of the system to which the target object belongs, and determine a balance type tag indicated by a preset rule corresponding to the class of the system; constructing a fund flow label system based on the system category and the expense type label; dividing the application tag library into the fund flow tag system according to the preset rule so as to establish a fund flow model corresponding to the transaction data.
An embodiment of the present invention also provides a storage medium including a stored program, wherein the program runs the method of any one of the above.
Alternatively, in the present embodiment, the above-described storage medium may be configured to store program code for executing the steps of:
s1, acquiring transaction data of a target object under the condition that the use authorization of the target object is acquired, wherein the use authorization is used for indicating the target object to allow the transaction data to be acquired;
s2, establishing a purpose tag library of the transaction data based on the purpose of the transaction data;
s3, establishing a purpose classification model of the transaction data according to the purpose tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library;
And S4, carrying out purpose classification on the transaction data according to the purpose classification model, and inputting the transaction data subjected to purpose classification into the fund flow model so as to determine the flow condition of the virtual resource corresponding to the target object.
In one exemplary embodiment, the computer readable storage medium may include, but is not limited to: a usb disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a removable hard disk, a magnetic disk, or an optical disk, or other various media capable of storing a computer program.
Specific examples in this embodiment may refer to the examples described in the foregoing embodiments and the exemplary implementation, and this embodiment is not described herein.
An embodiment of the invention also provides an electronic device comprising a memory having stored therein a computer program and a processor arranged to run the computer program to perform the steps of any of the method embodiments described above.
Alternatively, in this embodiment, the above-mentioned processor may be configured to execute the following steps by a computer program:
s1, acquiring transaction data of a target object under the condition that the use authorization of the target object is acquired, wherein the use authorization is used for indicating the target object to allow the transaction data to be acquired;
S2, establishing a purpose tag library of the transaction data based on the purpose of the transaction data;
s3, establishing a purpose classification model of the transaction data according to the purpose tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library;
and S4, carrying out purpose classification on the transaction data according to the purpose classification model, and inputting the transaction data subjected to purpose classification into the fund flow model so as to determine the flow condition of the virtual resource corresponding to the target object.
Optionally, the electronic apparatus may further include a transmission device and an input/output device, where the transmission device is connected to the processor, and the input/output device is connected to the processor.
Alternatively, specific examples in this embodiment may refer to examples described in the foregoing embodiments and optional implementations, and this embodiment is not described herein.
It will be appreciated by those skilled in the art that the modules or steps of the invention described above may be implemented in a general purpose computing system, they may be centralized in a single computing system, or distributed across a network of computing systems, and they may alternatively be implemented in program code that is executable by the computing system, such that they are stored in a memory system and, in some cases, executed in a different order than that shown or described, or they may be implemented as individual integrated circuit modules, or as individual integrated circuit modules. Thus, the present invention is not limited to any specific combination of hardware and software.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A method for determining a flow condition of a virtual resource, comprising:
obtaining transaction data of a target object in the case that the use authorization of the target object is determined to be obtained, wherein the use authorization is used for indicating that the target object allows the transaction data to be obtained;
establishing a purpose tag library of the transaction data based on the purpose of the transaction data;
establishing a purpose classification model of the transaction data according to the purpose tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library;
and classifying the use of the transaction data according to the use classification model, and inputting the transaction data subjected to use classification into the fund flow model to determine the flow condition of the virtual resource corresponding to the target object.
2. The method of determining a streaming situation of a virtual resource according to claim 1, wherein establishing a usage classification model of the transaction data according to the usage tag library and the transaction data comprises:
determining sample data corresponding to the transaction data based on the application tag library, wherein the sample data is transaction data with application tags;
performing feature construction operation on the sample data to obtain feature data corresponding to the sample data, wherein the feature data comprises at least one of the following: text feature vector representation, target object attribute features, and transaction business type features;
training a classification model by the sample data and the feature data to obtain the usage classification model.
3. The method for determining the flow condition of the virtual resource according to claim 2, wherein determining sample data corresponding to the transaction data based on the usage tag library comprises:
extracting text feature data in the transaction data, wherein the text feature data at least comprises: summary data, remark data, target object industry, target object name;
Constructing a tag rule model according to the text feature data and the application tag library;
and inputting the transaction data into the label rule model to determine sample data corresponding to the transaction data.
4. The method for determining the flow condition of the virtual resource according to claim 2, wherein performing a feature construction operation on the sample data to obtain feature data corresponding to the sample data includes:
determining text feature data in the sample data and application labels corresponding to the text feature data;
training to obtain a first sub-model based on the text feature data and the application label;
the sample data is input into the first sub-model to obtain a text feature vector representation of the sample data.
5. The method for determining the flow condition of the virtual resource according to claim 2, wherein performing a feature construction operation on the sample data to obtain feature data corresponding to the sample data includes:
determining all categories of text feature data in the sample data;
determining first sample data and second sample data in the sample data based on the all categories, the first sample data being sample data having text feature data of all categories, the second sample data being sample data not having text feature data of all categories;
Training based on the first sample data to obtain a second sub-model; filling the second sample data through a second sub-model to obtain filling data corresponding to the second sample data;
and extracting keywords of the sample data based on the types of the target objects, and determining the keywords and the filling data as target object attribute characteristics of the sample data, wherein the types are included in text characteristic data in the first sample data and the filling data.
6. The method for determining the flow condition of the virtual resource according to claim 2, wherein performing a feature construction operation on the sample data to obtain feature data corresponding to the sample data includes:
determining a business classification label corresponding to the transaction data, and determining a labeling rule of the transaction data according to an initiating component of the transaction data;
and classifying the sample data according to the labeling rules and the service classification labels to obtain transaction service type characteristics corresponding to the sample data.
7. The method for determining the flow condition of the virtual resource according to claim 1, wherein establishing a fund flow model corresponding to the transaction data according to the usage tag library comprises:
Determining a system category to which the target object belongs, and determining a balance type label indicated by a preset rule corresponding to the system category;
constructing a fund flow label system based on the system category and the expense type label;
dividing the application tag library into the fund flow tag system according to the preset rule so as to establish a fund flow model corresponding to the transaction data.
8. A device for determining a flow condition of a virtual resource, comprising:
an acquisition module, configured to acquire transaction data of a target object if it is determined that a usage authorization of the target object has been acquired, where the usage authorization is used to instruct the target object to allow the transaction data to be acquired;
the first establishing module is used for establishing a purpose tag library of the transaction data based on the purpose of the transaction data;
the second building module is used for building a purpose classification model of the transaction data according to the purpose tag library and the transaction data; establishing a fund flow model corresponding to the transaction data according to the application tag library;
and the classification module is used for classifying the use of the transaction data according to the use classification model, and inputting the transaction data subjected to use classification into the fund flow model so as to determine the flow condition of the virtual resource corresponding to the target object.
9. A computer readable storage medium, characterized in that the computer readable storage medium comprises a stored program, wherein the program runs the method according to any one of the preceding claims 1 to 7.
10. An electronic device comprising a memory and a processor, characterized in that the memory has stored therein a computer program, the processor being arranged to run the method according to any of the claims 1 to 7 by means of the computer program.
CN202311267730.3A 2023-09-27 2023-09-27 Method and device for determining flow condition of virtual resource and electronic device Pending CN117290505A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311267730.3A CN117290505A (en) 2023-09-27 2023-09-27 Method and device for determining flow condition of virtual resource and electronic device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311267730.3A CN117290505A (en) 2023-09-27 2023-09-27 Method and device for determining flow condition of virtual resource and electronic device

Publications (1)

Publication Number Publication Date
CN117290505A true CN117290505A (en) 2023-12-26

Family

ID=89247668

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311267730.3A Pending CN117290505A (en) 2023-09-27 2023-09-27 Method and device for determining flow condition of virtual resource and electronic device

Country Status (1)

Country Link
CN (1) CN117290505A (en)

Similar Documents

Publication Publication Date Title
TWI788529B (en) Credit risk prediction method and device based on LSTM model
CN110020660B (en) Integrity assessment of unstructured processes using Artificial Intelligence (AI) techniques
US20210173711A1 (en) Integrated value chain risk-based profiling and optimization
US20120221486A1 (en) Methods and systems for risk mining and for generating entity risk profiles and for predicting behavior of security
CN103154991A (en) Credit risk mining
CN109102394A (en) Methods of risk assessment, device and computer readable storage medium
CN111861716B (en) Method for generating monitoring early warning level in credit based on software system
Strauss et al. Tax risk assessment and assurance reform in response to the digitalised economy
US10922633B2 (en) Utilizing econometric and machine learning models to maximize total returns for an entity
CN113283806A (en) Enterprise information evaluation method and device, computer equipment and storage medium
CN116151840B (en) User service data intelligent management system and method based on big data
CN111639916A (en) Online auditing method, system and readable storage medium based on block chain technology and deep learning
WO2023114637A1 (en) Computer-implemented system and method of facilitating artificial intelligence based lending strategies and business revenue management
US11880394B2 (en) System and method for machine learning architecture for interdependence detection
CN115880077A (en) Recommendation method and device based on client label, electronic device and storage medium
CN117290505A (en) Method and device for determining flow condition of virtual resource and electronic device
Bineid et al. CADM: big data to limit creative accounting in Saudi-listed companies
US20160239914A1 (en) Analyzing Financial Market Transactions
WO2019136628A1 (en) Method, apparatus and system for assisting in enterprise listing
Price et al. Making monitoring manageable: a framework to guide learning
Kang Fraud Detection in Mobile Money Transactions Using Machine Learning
Sandhya Competitive Digital Strategic Model of Amazon-A Case Study Analysis
US20230306279A1 (en) Guided feedback loop for automated information categorization
US20210049553A1 (en) Rate Ingestion Tool
Chesney Using smart contracts to attract conservation investment

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