Financial data monitoring method and device, electronic equipment and storage medium

Abstract

The invention discloses a financial data monitoring method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining to-be-monitored financial data in a monitoring period, the to-be-monitored financial data comprising an initial category label provided by an associated person of the to-be-monitored financial data; classifying the to-be-monitored financial data based on the initial category label to obtain a target category label of the to-be-monitored financial data, the target category label comprising a parent category label and at least two sub-category labels under the parent category label; screening historical reference data associated with the target category label from a financial database; and performing clustering analysis on the to-be-monitored financial data by adopting the historical reference data to determine whether the to-be-monitored financial data is abnormal or not. According to the method, the to-be-monitored financial data is analyzed by adopting the rule of the historical financial data, and when the historical data is screened, the parent category labels and the multi-level subcategory labels are set, so that the rationality and accuracy of financial data monitoring are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of financial management, and in particular to a financial data monitoring method, device, electronic equipment and storage medium. Background Art

[0002] With the rapid development of informatization and intelligence, a large amount of financial data is generated in the process of business management due to the execution of transactions, operations, reimbursement and other activities, especially for enterprises or companies with more business types and a large number of employees. At present, most enterprises or companies use manual auditing of finance. On the one hand, manual auditing of financial data with a large amount of data is labor-intensive and prone to errors and omissions. On the other hand, financial personnel may only agree to check the company's finances once every half a month or a month. Usually, they only enter data and do simple processing, and cannot detect abnormal data in time. If abnormal or erroneous data is encountered, it may be difficult to recover the loss. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a financial data monitoring method.

[0004] In a first aspect, the present invention provides a financial data monitoring method, comprising:

[0005] Obtaining financial data to be monitored within a monitoring period and initial category labels provided by persons associated with the financial data to be monitored;

[0006] Classifying the financial data to be monitored based on the initial category label to obtain a target category label for the financial data to be monitored, wherein the target category label includes a parent category label and a subcategory label under the parent category label, and the subcategory label has at least two levels;

[0007] Filtering historical reference data associated with the target category label from a financial database;

[0008] The historical reference data is used to perform cluster analysis on the financial data to be monitored to determine whether the financial data to be monitored is abnormal.

[0009] In a second aspect, the present invention provides a financial data monitoring device, comprising:

[0010] A financial data acquisition module, used to acquire the financial data to be monitored within a monitoring period and the initial category labels provided by persons associated with the financial data to be monitored;

[0011] a target category label determination module, configured to classify the financial data to be monitored based on the initial category label to obtain a target category label of the financial data to be monitored, wherein the target category label includes a parent category label and a subcategory label under the parent category label, and the subcategory label has at least two levels;

[0012] A historical reference data acquisition module, used to filter out historical reference data associated with the target category label from a financial database;

[0013] The financial data anomaly judgment module is used to perform cluster analysis on the financial data to be monitored using the historical reference data to determine whether the financial data to be monitored is abnormal.

[0014] In a third aspect, the present invention provides an electronic device, the electronic device comprising:

[0015] at least one processor; and

[0016] a memory communicatively connected to the at least one processor; wherein,

[0017] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the financial data monitoring method described in the first aspect of the present invention.

[0018] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a processor to implement the financial data monitoring method described in the first aspect of the present invention when executed.

[0019] The embodiment of the present invention provides a financial data monitoring method, which obtains the financial data to be monitored within the monitoring period, wherein the financial data to be monitored includes the initial category label provided by the associated personnel of the financial data to be monitored; the financial data to be monitored is classified based on the initial category label to obtain the target category label of the financial data to be monitored, wherein the target category label includes the parent category label and the subcategory label under the parent category label, and the hierarchy of the subcategory label is at least two levels; the historical reference data associated with the target category label is screened out from the financial database; the historical reference data is used to perform cluster analysis on the financial data to be monitored to determine whether the financial data to be monitored is abnormal. In the present invention, the law of historical financial data is used to analyze the financial data to be monitored. When screening the historical data, by setting the parent category label and the multi-level subcategory label, the screened historical reference data can be relatively close to the characteristics of the financial data to be monitored, and then the characteristics of the financial data to be monitored can be judged as abnormal according to the characteristics of the historical reference data, which greatly improves the rationality and accuracy of judging the financial data to be monitored, and can automatically monitor the financial data for a long time without relying on manual processing, which is highly efficient and saves labor costs.

[0020] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 is a flow chart of a financial data monitoring method provided by an embodiment of the present invention;

[0023] Figure 2 is a flow chart of a training process of a classification model provided by an embodiment of the present invention;

[0024] Figure 3 is a structural schematic diagram of a financial data monitoring device provided by an embodiment of the present invention;

[0025] Figure 4 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0027] Figure 1 This is a flowchart of a financial data monitoring method provided by an embodiment of the present invention. This embodiment can be applied to the situation of monitoring the financial data of an enterprise or company to find abnormalities in time. The method can be executed by a financial data monitoring device. The financial data monitoring device can be implemented in the form of hardware and / or software. The financial data monitoring device can be configured in an electronic device. Figure 1 As shown, the financial data monitoring method includes:

[0028] S101, obtaining the financial data to be monitored within a monitoring period and the initial category labels provided by persons associated with the financial data to be monitored.

[0029] The main financial data of an enterprise or company include key data such as assets, liabilities, equity, income, expenses, profits, cash flow, etc. The data generally come from financial statements, directors' reports, management analysis and financial status statements.

[0030] The monitoring cycle can be set according to actual needs. Since the financial data monitoring is automated, a monitoring cycle with a smaller time period can be set. For example, if the monitoring cycle is once every 2 days, the system will automatically obtain the financial data to be monitored from the database within 2 days.

[0031] Financial data can usually appear in the form of Excel reports, invoices, pictures, etc. The associated persons of the financial data to be monitored can specifically be the persons who submit the financial data to be monitored. For example, when the financial data to be monitored is reimbursement data, the person who submits the financial data to be monitored is the reimbursement applicant. The associated persons of the financial data to be monitored can specifically be the staff of the finance department. When submitting / processing the financial data to be monitored, the reimbursement applicant / finance staff first classifies the financial data to be monitored, that is, provides an initial classification label.

[0032] S102. Classify the financial data to be monitored based on the initial category labels to obtain target category labels for the financial data to be monitored. The target category labels include a parent category label and a subcategory label under the parent category label. The subcategory label has at least two levels.

[0033] The initial category label can be provided by the person who submits the financial data to be monitored, or by the personnel of the financial department. Therefore, the initial category label is not necessarily a standard category label. Therefore, the financial data to be monitored is classified based on the initial category label to obtain the target category label of the financial data to be monitored, and the target category label is the standard category label. Specifically, the financial department usually pre-sets a standard category label for each type of financial matter and stores it in the label library.

[0034] In an optional embodiment, the financial data to be monitored is classified based on the initial category label to obtain the target category label of the financial data to be monitored, which specifically includes: determining whether there is a parent category label identical to the initial category label in a preset label library; if so, using the initial category label as the parent category label of the financial data to be monitored; if not, for each parent category label in the label library, calculating the first semantic cosine value of the parent category label and the initial category label; using the parent category label with the largest first semantic cosine value as the parent category label of the financial data to be monitored; and determining the subcategory label of the financial data to be monitored under the parent category label based on a preset classification model.

[0035] In this embodiment, the initial category label is the parent category label to be selected. Based on the semantic cosine value of the parent category label in the label library and the initial category label, the standard parent category label of the financial data to be monitored can be quickly determined, and then the financial data to be monitored can be further refined and classified according to the obtained parent category label.

[0036] The subcategory labels under the parent category labels of the financial data to be monitored are obtained through a classification model, which can be trained based on training data. The classification model can be a PromptT intelligent model, which is a low-code platform that uses visual components to assemble workflows. After receiving the parent category label and the financial data to be monitored, the classification model calls the semantic recognition interface of PromptT, outputs the subcategory label, and locates the specific financial category to which the financial data to be monitored belongs.

[0037] S103: Filter out historical reference data associated with the target category label from the financial database.

[0038] After obtaining the target category label of the financial data to be monitored, the financial category to which the financial data to be monitored specifically belongs is located. Therefore, the associated historical reference data can be obtained from the financial database based on the target category label.

[0039] S104. Perform cluster analysis on the financial data to be monitored using historical reference data to determine whether the financial data to be monitored is abnormal.

[0040] The historical reference data is the historical financial data that is closest or relatively close to the type of the financial data to be monitored. Here, it is assumed that the historical financial data is the monitored and normal financial data. Therefore, the historical reference data can be used to compare and analyze the financial data to be monitored to determine whether there is any abnormality in the financial data to be monitored.

[0041] In an optional embodiment, the financial data to be monitored includes multiple financial types, including balance sheets, income statements, cash flow statements, and financial ratios. The financial data to be monitored can also be clustered according to specific financial types. For each financial type, cluster analysis is performed on the data to be monitored and historical reference data to determine whether there are abnormalities in the financial data to be monitored. Before clustering, the data is standardized to ensure that the dimensions of each feature are consistent, which helps to reduce the cluster center offset caused by different data dimensions, that is, to reduce the error of financial data analysis.

[0042] Balance sheet: reflects the assets, liabilities and owner's equity of an enterprise during the monitoring period.

[0043] Profit and loss statement: reflects the company's operating results during the monitoring period. The profit and loss statement includes operating income, operating costs, taxes, sales expenses, administrative expenses, financial expenses, etc. The profit and loss statement can be used to understand the company's profitability and profit level.

[0044] Cash flow statement: reflects the inflow and outflow of cash of an enterprise during the monitoring period.

[0045] Financial ratios: including current ratio, quick ratio, debt-to-asset ratio, gross profit margin, etc. By analyzing these ratios, we can gain a deeper understanding of the company's financial status and operational efficiency.

[0046] The embodiment of the present invention provides a financial data monitoring method, which obtains the financial data to be monitored within the monitoring period, wherein the financial data to be monitored includes the initial category label provided by the associated personnel of the financial data to be monitored; the financial data to be monitored is classified based on the initial category label to obtain the target category label of the financial data to be monitored, wherein the category label includes the parent category label and the subcategory label under the parent category label, and the hierarchy of the subcategory label is at least two levels; the historical reference data associated with the target category label is screened out from the financial database; the historical reference data is used to perform cluster analysis on the financial data to be monitored to determine whether the financial data to be monitored is abnormal. In the present invention, the law of historical financial data is used to analyze the financial data to be monitored. When screening the historical data, by setting the parent category label and the multi-level subcategory label, the screened historical reference data can be relatively close to the characteristics of the financial data to be monitored, and then the characteristics of the financial data to be monitored can be judged as abnormal according to the characteristics of the historical reference data, which greatly improves the rationality and accuracy of judging the financial data to be monitored, and can automatically monitor the financial data for a long time without relying on manual processing, which is highly efficient and saves labor costs.

[0047] In an optional embodiment, the training process of the classification model is limited. Figure 2 This is a flowchart of the training process of a classification model, such as Figure 2 As shown, the preset classification model is trained through the following steps:

[0048] S1. For each parent category label, obtain financial training data labeled with a preset parent category label and N reference subcategory labels.

[0049] N is a positive integer greater than or equal to 2, and the semantics of the reference subcategory tag at the front position includes the semantics of the reference subcategory tag at the back position. Exemplarily, N=3, that is, the semantics of the first reference subcategory tag includes the semantics of the second reference subcategory tag, and the semantics of the second reference subcategory tag includes the semantics of the third reference subcategory tag.

[0050] S2. Use the preset parent category labels and financial training data to train the classification model to obtain N actual subcategory labels.

[0051] S3. Determine whether the N actual subcategory labels can form a semantic tree hierarchy.

[0052] If yes, execute S5, if no, execute S4. First, obtain the semantic features of each actual subcategory label based on semantic analysis, and then construct a semantic tree hierarchy based on the semantic features. In the semantic tree hierarchy, the number of actual subcategory labels in each layer is 1; in the semantic tree structure, there is a hierarchical relationship between nodes, one node can be the parent node of another node, and the latter becomes the child node of the former.

[0053] S4. Update the parameters of the classification model based on the semantic relationship between the actual subcategory labels, and return to execute S2.

[0054] Through repeated training, the subcategory labels output by the classification model can have a semantic hierarchical relationship, that is, the classification model can learn the ability to output subcategory labels in a hierarchical manner.

[0055] S5. Determine the position ranking of the N actual subcategory labels based on the semantic tree hierarchy.

[0056] The semantics of the actual subcategory label in the front position includes the semantics of the actual subcategory label in the back position. For example, N=3, that is, the semantics of the first actual subcategory label includes the semantics of the second actual subcategory label, and the semantics of the second actual subcategory label includes the semantics of the third actual subcategory label.

[0057] S6. Calculate the second semantic cosine value of the actual subcategory label and the reference subcategory label at the same position.

[0058] S7. Determine whether there is a second semantic cosine value less than 1.

[0059] If yes, execute S8, if no, execute S9.

[0060] The standard value of the second semantic cosine value is set to 1 here because the classification of financial data is usually pre-set professional terms. If some sub-category labels with similar semantics and a large number are expanded, it will not meet the actual needs of financial data monitoring. Therefore, the actual sub-category label of this embodiment needs to be the same as the reference sub-category label to determine that the classification model training is completed.

[0061] S8. Determine that the current classification model training is completed and obtain a trained classification model.

[0062] S9. Update the parameters of the classification model based on the difference between the second cosine value and 1, and return to execute S2.

[0063] In the process of training the classification model, when the second semantic cosine value of the actual subcategory label and the reference subcategory label at the same position is not 1, the parameters of the classification model are adjusted so that the actual subcategory label output by the classification model is close to the reference subcategory label (eventually the same as the reference subcategory label), that is, the classification model learns to output specific subcategory labels based on financial data.

[0064] In an optional embodiment, cluster analysis is performed on the financial data to be monitored using historical reference data to determine whether the financial data to be monitored is abnormal, including: performing a first clustering on the historical reference data to obtain a first cluster and a center point of the first cluster; the number of the first cluster is at least 1; performing a second clustering on the historical reference data and the financial data to be monitored to obtain a second cluster and a center point of the second cluster; the number of the second cluster is at least 1; judging whether the number of the second cluster is the same as the number of the first cluster; if not, determining that the financial data to be monitored is abnormal; if so, for each second cluster, taking the first cluster closest to the second cluster as a neighboring cluster; judging whether the financial data to be monitored is abnormal based on the distance relationship between the second cluster and the neighboring cluster.

[0065] By performing cluster analysis on the historical reference data and the combined data of the historical reference data and the financial data to be monitored, and comparing the two clustering results, it is possible to clearly determine whether the financial data to be monitored is abnormal. On the one hand, if new clusters appear after adding the financial data to be monitored, it means that the financial data to be monitored is abnormal. On the other hand, if no new clusters appear, but the financial data to be monitored has a significant impact on the original clustering results, it also means that the financial data to be monitored is abnormal.

[0066] In an optional example, judging whether the financial data to be monitored is abnormal based on the distance relationship between the second group and the adjacent group includes: calculating the distance from the center point of the second group to the center point of the corresponding adjacent group to obtain a sub-distance; judging whether the sum of all sub-distances is greater than a preset distance value; if so, determining that the financial data to be monitored is abnormal; if not, determining that the financial data to be monitored is normal.

[0067] In this example, after clustering, no new clusters appear in the combined data of historical reference data and the financial data to be monitored. Then, a secondary judgment is made based on the degree of influence of the financial data to be monitored on the original clustering results. The distance (sub-distance) between the center point of the second cluster and the center point of the adjacent cluster is calculated, that is, the degree of influence of the financial data to be monitored on the original clustering results is calculated. All sub-distances are accumulated to obtain the total degree of influence. When the degree of influence is large, it also indicates that there is an abnormality in the financial data to be monitored.

[0068] Through comparative analysis of the above two clustering results, abnormal financial data can be monitored accurately and efficiently, and the entire process is automated.

[0069] In an optional example, after determining that the financial data to be monitored is abnormal, the method further includes: sending the financial data to be monitored to a manual end for review according to the subcategory label of the financial data to be monitored.

[0070] Specifically, the financial data to be monitored is sent to the manual end for review according to the subcategory label of the financial data to be monitored, including: determining the financial review object of the manual end according to the lowest-level subcategory label corresponding to the financial data to be monitored; setting an abnormal label for the financial data to be monitored and sending it to the financial review object for review. The financial data to be monitored can be classified in more detail to accurately locate the classification of the financial data to be monitored, and the abnormal financial data to be monitored can be sent to the corresponding financial review object for review according to the classification, so as to achieve professional and targeted processing, and then the financial data to be monitored can be accurately and efficiently processed, and whether the financial data to be monitored is abnormal can be verified.

[0071] The manual end is the backend, and there are various types of financial data. The manual end is equipped with multiple financial audit objects. The financial audit objects can be personnel who specialize in solving different functional problems. For example, the financial audit objects of Group A specialize in financial data of financial reimbursement, and the work objects of Group B specialize in solving financial data of financial ratios. User feedback information can be fed back to the corresponding group based on the parent category label, and the subcategory label can quickly locate the subcategory to which the financial data to be monitored belongs, which helps the financial audit objects to respond and process quickly, thereby improving work efficiency and accuracy.

[0072] Corresponding to the financial data monitoring method of the present invention, in one embodiment of the present invention, a financial data monitoring device is also provided. Figure 3 The structure diagram of a financial data monitoring device provided by an embodiment of the present invention is shown in FIG. Figure 3 As shown, the financial data monitoring device includes:

[0073] The financial data acquisition module 301 is used to acquire the financial data to be monitored within the monitoring period and the initial category labels provided by the persons associated with the financial data to be monitored;

[0074] A target category label determination module 302 is used to classify the financial data to be monitored based on the initial category label to obtain a target category label of the financial data to be monitored, wherein the target category label includes a parent category label and a subcategory label under the parent category label, and the hierarchy of the subcategory label is at least two levels;

[0075] A historical reference data acquisition module 303 is used to filter out historical reference data associated with the target category tag from a financial database;

[0076] The financial data anomaly judgment module 304 is used to perform cluster analysis on the financial data to be monitored using the historical reference data to determine whether the financial data to be monitored is abnormal.

[0077] Optionally, the target category label determination module 302 includes:

[0078] The label determination submodule is used to determine whether there is a parent category label identical to the initial category label in the preset label library; if so, the content of the first label determination submodule is executed; if not, the content of the first semantic cosine value calculation submodule is executed;

[0079] A first label determination submodule, configured to use the initial category label as a parent category label of the financial data to be monitored;

[0080] A first semantic cosine value calculation submodule, used for calculating, for each parent category label in the label library, a first semantic cosine value of the parent category label and the initial category label;

[0081] A second label determination submodule, configured to use the parent category label with the largest first semantic cosine value as the parent category label of the financial data to be monitored;

[0082] The third label determination submodule is used to determine the subcategory label of the to-be-monitored financial data under the parent category label based on a preset classification model.

[0083] Optionally, a model training module is also included, which includes:

[0084] The training data acquisition module is used to acquire, for each parent category label, financial training data annotated with a preset parent category label and N reference subcategory labels; N is a positive integer greater than or equal to 2, and the semantics of the reference subcategory label in the front position includes the semantics of the reference subcategory label in the back position;

[0085] A training classification module, used for training a classification model using a preset parent category label and the financial training data to obtain N actual subcategory labels;

[0086] A hierarchy judgment module is used to judge whether the N actual sub-category labels can form a semantic tree hierarchy structure; in the semantic tree hierarchy structure, the number of the actual sub-category labels in each layer structure is 1; if so, the content of the position sorting module is executed, if not, the content of the first updating module is executed;

[0087] A first updating module is used to update the parameters of the classification model based on the semantic relationship between the actual sub-category labels; after execution, the content of the training classification module is returned;

[0088] A position sorting module is used to determine the position sorting of N actual subcategory labels based on a semantic tree hierarchy, where the semantics of the actual subcategory label at the front includes the semantics of the actual subcategory label at the back;

[0089] A second semantic cosine value calculation module, used to calculate the second semantic cosine value of the actual subcategory label and the reference subcategory label at the same position;

[0090] A second semantic cosine value judgment module, used to judge whether the second semantic cosine value is less than 1; if so, execute the content of the training completion determination module; if not, execute the content of the second update module;

[0091] The training completion determination module is used to determine that the current classification model training is completed and obtain the trained classification model;

[0092] The second updating module is used to update the parameters of the classification model based on the difference between the second cosine value and 1.

[0093] After execution, return to the content of executing the training classification module.

[0094] Optionally, the financial data abnormality judgment module 304 includes:

[0095] A first clustering submodule, used for performing a first clustering on the historical reference data to obtain a first group and a center point of the first group; the number of the first group is at least 1;

[0096] A second clustering submodule, used for performing a second clustering on the historical reference data and the financial data to be monitored to obtain a second cluster and a center point of the second cluster; the number of the second cluster is at least 1;

[0097] A cluster analysis submodule, used to determine whether the number of the second clusters is the same as the number of the first clusters; if not, execute the content of the abnormality determination submodule; if so, execute the content of the neighboring cluster determination submodule;

[0098] An abnormality determination submodule, used to determine the abnormality of the financial data to be monitored;

[0099] a neighboring group determination submodule, configured to, for each of the second groups, take the first group closest to the second group as a neighboring group;

[0100] The abnormality judgment submodule is used to judge whether the financial data to be monitored has abnormality based on the distance relationship between the second group and the adjacent group.

[0101] Optionally, the abnormality judgment submodule includes:

[0102] a sub-distance calculation unit, configured to calculate the distance from the center point of the second group to the center point of the corresponding adjacent group to obtain a sub-distance;

[0103] The sub-distance judgment unit is used to judge whether the sum of all sub-distances is greater than a preset distance value; if so, the content of the abnormal determination unit is executed; if not, the content of the normal determination unit is executed;

[0104] An abnormality determination unit, used for determining that the financial data to be monitored is abnormal;

[0105] The normal determination unit is used to determine whether the financial data to be monitored is normal.

[0106] Optionally, the financial data abnormality judgment module 304 further includes:

[0107] The review submodule is used to send the financial data to be monitored to a manual end for review according to the subcategory label of the financial data to be monitored after determining that the financial data to be monitored is abnormal.

[0108] Optionally, the audit submodule includes:

[0109] An audit object determination unit, used to determine the financial audit object of the manual end according to the lowest-level subcategory label corresponding to the financial data to be monitored;

[0110] The financial audit unit is used to set an abnormal label for the financial data to be monitored and send it to the financial audit object for audit.

[0111] The financial data monitoring device provided in the embodiment of the present invention can execute the financial data monitoring method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0112] Figure 4A schematic diagram of the structure of an electronic device 40 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0113] like Figure 4 As shown, the electronic device 40 includes at least one processor 41, and a memory connected to the at least one processor 41, such as a read-only memory (ROM) 42, a random access memory (RAM) 43, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 41 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 42 or the computer program loaded from the storage unit 48 to the random access memory (RAM) 43. In the RAM 43, various programs and data required for the operation of the electronic device 40 can also be stored. The processor 41, the ROM 42, and the RAM 43 are connected to each other through a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.

[0114] A number of components in the electronic device 40 are connected to the I / O interface 45, including: an input unit 46, such as a keyboard, a mouse, etc.; an output unit 47, such as various types of displays, speakers, etc.; a storage unit 48, such as a disk, an optical disk, etc.; and a communication unit 49, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 49 allows the electronic device 40 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0115] The processor 41 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 41 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 41 executes the various methods and processes described above, such as a financial data monitoring method.

[0116] In some embodiments, the financial data monitoring method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 48. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 40 via the ROM 42 and / or the communication unit 49. When the computer program is loaded into the RAM 43 and executed by the processor 41, one or more steps of the financial data monitoring method described above may be performed. Alternatively, in other embodiments, the processor 41 may be configured to perform the financial data monitoring method in any other suitable manner (e.g., by means of firmware).

[0117] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0118] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0119] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0120] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0121] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0122] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0123] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0124] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A financial data monitoring method, characterized in that: include: Obtaining financial data to be monitored within a monitoring period and initial category labels provided by persons associated with the financial data to be monitored; Classifying the financial data to be monitored based on the initial category label to obtain a target category label for the financial data to be monitored, wherein the target category label includes a parent category label and a subcategory label under the parent category label, and the subcategory label has at least two levels; Filtering historical reference data associated with the target category label from a financial database; The historical reference data is used to perform cluster analysis on the financial data to be monitored to determine whether the financial data to be monitored is abnormal.

2. The method according to claim 1, characterized in that The classifying the financial data to be monitored based on the initial category label to obtain a target category label of the financial data to be monitored includes: Determine whether there is a parent category label that is the same as the initial category label in the preset label library; If so, using the initial category label as the parent category label of the financial data to be monitored; If not, for each parent category label in the label library, calculate the first semantic cosine value of the parent category label and the initial category label; Using the parent category label with the largest first semantic cosine value as the parent category label of the financial data to be monitored; The subcategory label of the financial data to be monitored under the parent category label is determined based on a preset classification model.

3. The method according to claim 2, characterized in that The preset classification model is trained through the following steps: S1. For each parent category label, obtain financial training data labeled with a preset parent category label and N reference subcategory labels; N is a positive integer greater than or equal to 2, and the semantics of the reference subcategory label at the front position includes the semantics of the reference subcategory label at the back position; S2. Using the preset parent category label and the financial training data to train the classification model to obtain N actual subcategory labels; S3, determining whether the N actual subcategory labels can form a semantic tree hierarchy; if so, executing S5, if not, executing S4; in the semantic tree hierarchy, the number of the actual subcategory labels in each layer structure is 1; S4, updating the parameters of the classification model based on the semantic relationship between the actual subcategory labels, and returning to execute S2; S5. determining the position order of the N actual subcategory labels based on the semantic tree hierarchy, wherein the semantics of the actual subcategory label at the front includes the semantics of the actual subcategory label at the back; S6, calculating the second semantic cosine value of the actual subcategory label and the reference subcategory label at the same position; The second semantic cosine value ranges from 0 to 1, indicating the similarity between the actual subcategory label and the reference subcategory label. When the second semantic cosine value is 1, it indicates that the actual subcategory label is the same as the reference subcategory label. S7, determining whether the second semantic cosine value is less than 1; If yes, execute S8, if no, execute S9; S8, determining that the current classification model training is completed, and obtaining a trained classification model; S9. Update the parameters of the classification model based on the difference between the second cosine value and 1, and return to execute S2.

4. The method according to claim 1, characterized in that The using the historical reference data to perform cluster analysis on the financial data to be monitored to determine whether the financial data to be monitored is abnormal includes: Performing a first clustering on the historical reference data to obtain a first group and a center point of the first group; the number of the first group is at least 1; Performing a second clustering on the historical reference data and the financial data to be monitored to obtain a second cluster and a center point of the second cluster; the number of the second cluster is at least 1; determining whether the number of the second group is the same as the number of the first group; If not, determining that the financial data to be monitored is abnormal; If yes, for each of the second groups, the first group closest to the second group is taken as a neighboring group; Based on the distance relationship between the second group and the adjacent group, it is determined whether the financial data to be monitored has an abnormality.

5. The method according to claim 4, characterized in that The determining whether the to-be-monitored financial data has an abnormality based on the distance relationship between the second group and the adjacent group includes: Calculate the distance between the center point of the second group and the center point of the corresponding adjacent group to obtain a sub-distance; Determine whether the sum of all sub-distances is greater than a preset distance value; If yes, determining that the financial data to be monitored is abnormal; If not, it is determined that the financial data to be monitored is normal.

6. The method according to claim 5, characterized in that After determining that the financial data to be monitored is abnormal, the method further includes: The financial data to be monitored is sent to a manual end for review according to the subcategory label of the financial data to be monitored.

7. The method according to claim 6, characterized in that The step of sending the financial data to be monitored to a manual end for review according to the subcategory label of the financial data to be monitored includes: Determine the financial audit object of the manual end according to the lowest-level subcategory label corresponding to the financial data to be monitored; An abnormal label is set for the financial data to be monitored and sent to the financial audit object for audit.

8. A financial data monitoring device, characterized in that: include: A financial data acquisition module, used to acquire the financial data to be monitored within a monitoring period and the initial category labels provided by persons associated with the financial data to be monitored; a target category label determination module, configured to classify the financial data to be monitored based on the initial category label to obtain a target category label of the financial data to be monitored, wherein the target category label includes a parent category label and a subcategory label under the parent category label, and the subcategory label has at least two levels; A historical reference data acquisition module, used to filter out historical reference data associated with the target category label from a financial database; The financial data anomaly judgment module is used to perform cluster analysis on the financial data to be monitored using the historical reference data to determine whether the financial data to be monitored is abnormal.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the financial data monitoring method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the financial data monitoring method according to any one of claims 1 to 7 when executed.

Images