A method and device for classifying financial text data based on a pre-trained model

Abstract

The application discloses a kind of based on pre-training model's financial text data classification method and device, applied to data classification technical field.The method is first by pre-processing to financial text data, constructs the data containing classification mark, wherein the number of the classification mark is not less than one.Then, construct the data classification model based on pre-training model BERT, and the model is trained by using pre-processing training data by subnetwork fine-tuning method.Finally, the financial text data to be classified is classified using the trained data classification model.The application optimizes fine-tuning algorithm, not only improves the efficiency of financial text data classification, but also enhances the accuracy of training model, ensures the accuracy of classification result, can meet the accuracy requirement in practical application.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to a financial text data classification method and device based on a pre-trained model. BACKGROUND

[0002] With the development of information technology, numerous financial basic businesses, core processes, and inter-industry transactions and activities have been running on information-based support carriers. The information generated in the production and operation process of financial institutions is gradually transformed into digital assets in different forms and circulated between different information networks and systems. With the in-depth application of new technologies such as big data, artificial intelligence, and cloud computing in the financial industry, data has gradually transformed from information assets to production factors, and its importance has become increasingly prominent. The influence of data security threats in the financial industry has gradually expanded from within the institution to between industries, and even affected national security, social order, public interests, and financial market stability.

[0003] Financial data is complex and diverse. Implementing hierarchical management of data can further clarify the data protection objects, help financial institutions allocate data protection resources and costs reasonably, and is the basis for financial institutions to establish a complete financial data lifecycle protection framework and the prerequisite for targeted implementation of data security management.

[0004] Under such circumstances, financial institutions have an increasing demand for technologies that can accurately classify and process financial text data. The BERT (Bidirectional Encoder Representations from Transformers) model, as an advanced natural language processing (NLP) technology, provides a new solution for the classification and processing of financial text data through its powerful language understanding and text representation capabilities. The BERT model can deeply understand the semantics and context of text data, thereby improving the accuracy and efficiency of financial text data classification. However, in the field of financial text classification, the application of the BERT model still faces some challenges and limitations. Due to the large model parameter capacity and limited financial domain annotation data, it may not be able to directly capture professional terms and specific contexts in the financial field, resulting in a decline in test data results due to overfitting, which limits the precision in the financial text classification task. SUMMARY

[0005] The purpose of the present application is to provide a financial text data classification method and device based on a pre-trained model to address the above-mentioned deficiencies in the prior art.

[0006] The purpose of the present application is achieved by the following technical solution: a financial text data classification method based on a pre-trained model, comprising the following steps:

[0007] (1) Preprocess the financial text data to construct data containing classification labels as training data, wherein the number of classification labels is not less than one;

[0008] (2) Train a data classification model based on the pre-trained model BERT using the training data, and in the back propagation process, train the model by a subnetwork fine-tuning method, including:

[0009] (2.1) Initialize parameters: set learning rate η, probability p, initial pre-trained model weight w0, and hyperparameters β1 and β2 of AdamW optimizer, wherein β1 and β2 are exponential decay rates for calculating first and second moment estimates of gradients;

[0010] (2.2) Calculate the gradient: in the tth iteration, calculate the gradient of the loss function with respect to the parameter corresponding to the current model parameter where w t is the model weight after the tth iteration, is the loss function with w t as the parameter;

[0011] (2.3) Generate a subnetwork gradient mask: in the tth iteration, generate a 0-1 mask matrix M t , the elements of the mask matrix are randomly selected from a Bernoulli distribution with probability p, which is used to selectively update the model parameters;

[0012] (2.4) Apply gradient mask: perform Hadamard product (element-wise multiplication) between the gradient g t and the mask matrix M t generated in (3.3) to obtain the masked gradient g ′ t = g t ⊙M t , so that the gradient matrix only retains the elements corresponding to 1 in the mask matrix M t ;

[0013] (2.5) Update the first moment estimate: m t = β1·m t-1 +(1-β1)·g ′ t ;

[0014] (2.6) Update the second moment estimate: v t = β2·v t-1 +(1-β2)·g ′ t 2 ;

[0015] (2.7) Bias correction: bias correction is performed on the first moment estimate and the second moment estimate to obtain a corrected first moment estimate and a corrected second moment estimate

[0016] (2.8) Update model weights: update the model weights according to the corrected first moment estimate and the corrected second moment estimate, and the learning rate

[0017] (2.9) Iterative training: repeat steps (2.2) to (2.8) until the model converges or reaches a preset number of iterations.

[0018] (2.10) Fine-tuning training is completed: during fine-tuning, the model weights are updated through selective updating of the mask matrix M t .

[0019] (3) The trained data classification model is used to classify the financial text data to be classified.

[0020] Further, it also includes:

[0021] The trained data classification model is deployed in a data classification system to classify the financial text data to be classified.

[0022] Further, the trained data classification model is deployed in a data classification system to classify the financial text data to be classified, including:

[0023] Data upload: through the file upload module in the user interface, the user uploads the financial text data file to be classified to the data classification system;

[0024] Data preprocessing: the system automatically preprocesses the uploaded financial text data;

[0025] Model application: use the trained data classification model to automatically classify and label the text data preprocessed in step (4.2);

[0026] Result storage: store the classified and labeled data set in the system database and the specified local storage location for subsequent query, analysis and application;

[0027] Report generation: provide a user interface for users to view the classification results, and the system can automatically generate data analysis charts according to the classification results for users to conduct in-depth analysis and decision support.

[0028] The present application also provides a financial text data classification device based on a pre-trained model, comprising:

[0029] The preprocessing module is configured to preprocess the financial text data to construct data containing classification labels as training data, wherein the number of the classification labels is not less than one.

[0030] The model training module is configured to train a data classification model based on a pre-trained model BERT using the training data, and train the model by a subnetwork fine-tuning method in a back propagation process, including:

[0031] Initialization parameters: setting a learning rate η, a probability p, an initial pre-trained model weight w0, and hyperparameters β1 and β2 of an AdamW optimizer, wherein β1 and β2 are exponential decay rates for calculating first-order and second-order moment estimates of gradients;

[0032] Gradient calculation: in the tth iteration, the gradient of the loss function with respect to the parameter corresponding to the current model parameter is calculated where w t is the model weight after the tth iteration, is the loss function with w t as the parameter;

[0033] Generation of subnetwork gradient mask: in the tth iteration, a 0-1 mask matrix M t is generated, the elements of the mask matrix are randomly selected from a Bernoulli distribution with a probability p, and are used to selectively update the model parameters;

[0034] Application of gradient mask: the gradient g t is subjected to Hadamard product (element-by-element multiplication) with the mask matrix M t generated in (3.3) to obtain the masked gradient g ′ t =g t ⊙M t , so that the gradient matrix only retains the elements corresponding to 1 in the mask matrix M t ;

[0035] Update of first-order moment estimate: m t =β1·m t-1 +(1-β1)·g ′ t ;

[0036] Update of second-order moment estimate: v t =β2·v t-1 +(1-β2)·g ′ t 2 ;

[0037] Bias correction: bias correction is performed on the first-order moment estimate and the second-order moment estimate to obtain the corrected first-order moment estimate and the second-order moment estimate

[0038] updating model weights: updating model weights according to the corrected first moment estimate and second moment estimate, and learning rate

[0039]

[0040] iterative training: repeating the above process until the model converges or reaches a preset number of iterations;

[0041] fine-tuning training is completed: in the fine-tuning process, the update of the model weights is realized through the selective update of the mask matrix M t

[0042] data classification module, including the trained data classification model, used for classifying the financial text data to be classified.

[0043] Further comprising:

[0044] deployment module, for deploying the trained data classification model to the data classification system to classify the financial text data to be classified.

[0045] Further, the trained data classification model is deployed to the data classification system to classify the financial text data to be classified, comprising:

[0046] data uploading: through the file uploading module in the user interface, the user uploads the financial text data file to be classified to the data classification system;

[0047] data preprocessing: the system automatically preprocesses the uploaded financial text data;

[0048] model application: using the trained data classification model, the text data preprocessed in step (4.2) is automatically classified and labeled;

[0049] result storage: storing the data set after classification and labeling in the system database and the specified local storage location, so as to facilitate subsequent query, analysis and application;

[0050] report generation: providing a user interface for the user to view the classification results, and the system can automatically generate data analysis charts according to the classification results, so as to provide in-depth analysis and decision support for the user.

[0051] The application also provides an electronic device comprising a memory and a processor, wherein the memory is coupled to the processor; wherein the memory is used to store program data, and the processor is used to execute the program data to realize the above-mentioned financial text data classification method based on a pre-trained model.

[0052] ​The application further provides a computer readable storage medium, which stores a computer program, and the program is executed by a processor to realize the financial text data classification method based on a pre-trained model.

[0053] The application further provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to realize the financial text data classification method based on a pre-trained model.

[0054] The application has the beneficial effect of enhancing the generalization ability of the model for financial field texts, enabling the model to better understand and process various financial text data, reducing the overfitting phenomenon, and the implementation method is simple, and the precision of financial text data classification can be effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0056] Figure 1 is a flowchart of the financial text data classification method based on a pre-trained model provided by the embodiments of the application.

[0057] Figure 2 is a flowchart of training the model by the sub-network fine-tuning method provided by the embodiments of the application.

[0058] Figure 3 is a schematic diagram of a data classification system module provided by the embodiments of the application.

[0059] Figure 4 is a schematic diagram of an electronic device structure provided by the embodiments of the application. DETAILED DESCRIPTION

[0060] The application will be described in detail below with reference to the drawings. The features in the following embodiments and implementation manners can be combined with each other without conflict.

[0061] The application provides a financial text data classification method based on a pre-trained model and a data classification system, Figure 1 is a flowchart of the data classification method and the data classification system of the application, and specifically comprises the following steps:

[0062] Step 1: constructing data containing classification identifiers by preprocessing the financial text data, wherein the number of the classification identifiers is not less than one;

[0063] Step two: training a data classification model based on a pre-trained model BERT using training data, in the back propagation process, the model is trained by a subnetwork fine-tuning method, including:

[0064] Figure 2 is a flowchart of training a model by a subnetwork fine-tuning method provided by an embodiment of the application, which is the core of the application and is divided into the following sub-steps:

[0065] (2.1) initialize parameters: set learning rate η, probability p, initial pre-trained model weight w0, and AdamW optimizer hyperparameters β1 and β2, wherein β1 and β2 are exponential decay rates for calculating first and second moment estimates of gradients;

[0066] (2.2) calculate gradient: in the tth iteration, calculate the gradient of the loss function with respect to the parameter corresponding to the current model parameter where w t is the model weight after the tth iteration, t is the loss function with w as the parameter;

[0067] (2.3) generate subnetwork gradient mask: in the tth iteration, generate a 0-1 mask matrix M t , the elements of the mask matrix are randomly selected from a Bernoulli distribution with probability p, which is used to selectively update the model parameters;

[0068] (2.4) apply gradient mask: perform Hadamard product (element-wise multiplication) between the gradient g t and the mask matrix M t generated in (3.3) to obtain the masked gradient g ′ t = g t ⊙M t , so that the gradient matrix only retains the elements corresponding to 1 in the mask matrix M t ;

[0069] (2.5) update first moment estimate: m t = β1·m t-1 +(1-β1)·g ′ t ;

[0070] (2.6) update second moment estimate: v t = β2·v t-1 +(1-β2)·g ′ t 2 ;

[0071] (2.7) Bias correction: bias correction is performed on the first moment estimate and the second moment estimate to obtain the corrected first moment estimate and the second moment estimate

[0072] (2.8) Update model weights: update the model weights according to the corrected first moment estimate and the second moment estimate, and the learning rate

[0073] (2.9) Iterative training: repeat the above steps until the model converges or reaches a preset number of iterations;

[0074] (2.10) Fine-tuning training is completed: in the fine-tuning process, the update of the model weights is realized through the selective update of the mask matrix M t ;

[0075] Step three: the trained data classification model is used to classify the financial text data to be classified.

[0076] In an embodiment, further comprising: deploying the trained data classification model to a data classification system to classify the financial text data to be classified; as shown in Figure 3 , the data classification system includes a file upload module, a data preprocessing module, a model classification module, a file storage module and a report generation module. Specifically, the following steps are included:

[0077] Data upload: through the file upload module in the user interface, the user can upload the financial text data file to be classified to the classification system;

[0078] Data preprocessing: the system automatically preprocesses the uploaded financial text data through the data preprocessing module;

[0079] Model application: the model classification module uses the trained financial text data classification model to automatically classify and label the preprocessed text data;

[0080] Result file storage: store the classified and labeled data set in the system database and the specified local storage location to facilitate subsequent query, analysis and application;

[0081] Report generation: provide a user interface for users to view the classification results, and the system can automatically generate data analysis charts and exportable tables according to the classification results to support in-depth analysis and decision-making for users;

[0082] The performance comparison of the model trained by the sub-network fine-tuning method of the present application and the bert-base-chinese model on the financeSimple data set is shown in Table 1 as follows:

[0083] Table 1: Performance comparison

[0084] macro metrics precision recal‖ f1 bert-base-chinese 90.892 89.424 89.424 childNetwork-bert 92.29 91.584 91.468 improvement 1.398 2.16 2.044

[0085] From the above table 1, the evaluation index of the model trained by the child network fine-tuning method (childNetwork-bert) is compared with the bert-base-chinese model as the baseline method. The data set contains the information of the same flush 2000-2022 listed companies and various research reports and company announcements collected from public channels, which contains 50,000 samples, each sample has 6 features. Compared with the baseline method, the precision, recall and f1 evaluation indexes of the macro average are improved, which has practical value.

[0086] The application further provides a financial text data classification device based on a pre-training model, comprising:

[0087] A preprocessing module is configured to preprocess the financial text data and construct data containing classification identifiers as training data, wherein the number of classification identifiers is not less than one.

[0088] A model training module is configured to train a data classification model based on a pre-training model BERT using the training data, and train the model by a child network fine-tuning method in a back propagation process, comprising:

[0089] Initialize parameters: set learning rate η, probability p, initial pre-training model weight w0 and hyperparameters β1 and β2 of AdamW optimizer, wherein β1 and β2 are exponential decay rates, used to calculate the first and second moment estimates of the gradient;

[0090] Calculate the gradient: in the tth iteration, calculate the gradient of the loss function with respect to the parameter corresponding to the current model parameter where w t is the model weight after the tth iteration, is the loss function with w t as the parameter;

[0091] Generate a child network gradient mask: in the tth iteration, generate a 0-1 mask matrix M t , the elements of the mask matrix are randomly selected from a Bernoulli distribution with probability p, used to selectively update the model parameters;

[0092] Apply the gradient mask: perform Hadamard product (element-wise multiplication) between the gradient g t and the mask matrix M t generated in (3.3) to obtain the masked gradient g ′ t =gt ⊙M t , so that the gradient matrix only retains the elements in the mask matrix M t corresponding to 1;

[0093] Update the first moment estimate: m t = β1·m t-1 +(1-β1)·g ′ t ;

[0094] Update the second moment estimate: v t = β2·v t-1 +(1-β2)·g ′ t 2 ;

[0095] Bias correction: bias correction is performed on the first moment estimate and the second moment estimate to obtain the corrected first moment estimate and the second moment estimate

[0096] Update model weights: update the model weights according to the corrected first moment estimate and the second moment estimate, and the learning rate

[0097]

[0098] Iterative training: repeat the above process until the model converges or reaches a preset number of iterations;

[0099] Fine-tuning training is completed: in the fine-tuning process, the update of the model weights is realized through the selective update of the mask matrix M t ;

[0100] The data classification module includes a trained data classification model, which is used to classify the financial text data to be classified.

[0101] In an embodiment, it further includes a deployment module for deploying the trained data classification model to a data classification system to classify the financial text data to be classified.

[0102] Specifically, the following processes are included:

[0103] Data upload: through the file upload module in the user interface, the user uploads the financial text data file to be classified to the data classification system;

[0104] Data preprocessing: the system automatically preprocesses the uploaded financial text data;

[0105] Model application: use the trained data classification model to automatically classify and label the text data preprocessed in step (4.2);

[0106] Result storage: store the data set after classification and labeling in the system database and the designated local storage location, so as to facilitate subsequent query, analysis and application;

[0107] Report generation: provide a user interface for users to view the classification results, and the system can automatically generate data analysis charts according to the classification results for users to conduct in-depth analysis and decision support.

[0108] The application also provides an electronic device, which refers to Figure 4 , comprising a memory and a processor, the memory is coupled with the processor; wherein the memory is used for storing program data, and the processor is used for executing the program data to realize the above-mentioned financial text data classification method based on a pre-trained model. It should be noted that in addition to the memory and the processor shown in Figure 4 , the electronic device can also include other hardware according to its actual function, which will not be described here.

[0109] The application also provides a computer readable storage medium, which stores a computer program, and the program is executed by a processor to realize the above-mentioned financial text data classification method based on a pre-trained model.

[0110] The application also provides a computer program product, which includes a computer program, and the computer program is executed by a processor to realize the above-mentioned financial text data classification method based on a pre-trained model.

[0111] The computer readable storage medium can be an internal storage unit of any device with data processing capability, such as a hard disk or a memory. The computer readable storage medium can also be any device with data processing capability, such as a plug-in hard disk, a smart media card (SMC), an SD card, a flash card, etc. Further, the computer readable storage medium can include both the internal storage unit of any device with data processing capability and the external storage device. The computer readable storage medium is used to store the computer program and other programs and data required by the device with data processing capability, and can also be used to temporarily store data that has been output or will be output.

[0112] Those skilled in the art will appreciate that embodiments of the application can be readily used as software, hardware, or a combination of software and hardware. In a software embodiment, various elements are implemented in hardware, software, or a combination of both hardware and software. In a software embodiment, the software implementation can include but is not limited to a bit stream, program, routine, applet, object, component, data, instruction, instruction set, or

[0113] The present application is described in reference to the flowchart illustrations and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart Figure 1 one or more functions specified in the flowchart block or blocks. Figure 1 one or more functions specified in the flowchart block or blocks.

[0114] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more functions specified in the flowchart block or blocks. Figure 1 one or more functions specified in the flowchart block or blocks.

[0115] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more functions specified in the flowchart block or blocks. Figure 1 one or more functions specified in the flowchart block or blocks.

[0116] The above embodiments are only used to illustrate the design ideas and features of the present application, and the purpose is to make those skilled in the art understand the present application and implement it, and the protection scope of the present application is not limited to the above embodiments. Therefore, any equivalent changes or modifications made according to the principles and design ideas disclosed by the present application are within the protection scope of the present application.

Claims

1. A pre-training model-based financial text data classification method, characterized in that, Comprise the following steps: (1) by preprocessing the financial text data, build data containing classification identification as training data, wherein the number of classification identification is not less than one; (2) using training data to train data classification model based on pre-training model BERT, in the process of back propagation, the model is trained by subnetwork fine-tuning method, including: (2.1) Initialize parameters: set learning rate , probability p, initial pre-trained model weights and hyperparameters of AdamW optimizer and where and are the exponential decay rates for the first and second moment estimates of the gradient; (2.2) Compute the gradient: in the t-th iteration, compute the gradient of the loss function with respect to the parameters corresponding to the current model parameters where is the model weight after the t-th iteration, is the loss function with as the parameters; (2.3) Generating sub-network gradient mask: at the t-th iteration, a 0-1 mask matrix is generated , the elements of which are randomly drawn from a Bernoulli distribution with probability p, for selectively updating the model parameters; (2.4) Apply gradient mask: apply the gradient generated in (3.3) to the gradient to obtain the masked gradient such that the gradient matrix only retains the elements in the mask that are 1. (2.5) updating the first moment estimate: ; (2.6) update the second moment estimate: ; (2.7) Bias correction: bias correction is performed on the first moment estimate and the second moment estimate to obtain a corrected first moment estimate and a corrected second moment estimate ; (2.8) updating the model weights: updating the model weights according to the corrected first moment estimate and the second moment estimate, and the learning rate ; (2.9) iterative training: repeat step (2.2)~step (2.8), until the model converges or reaches the preset iteration number; (2.10) Fine-tuning training is completed: in the fine-tuning process, the update of the model weight is realized through the selective update of the mask matrix ​ (3) the trained data classification model is used for classifying financial text data to be classified.

2. The method of claim 1, wherein, Also include: The trained data classification model is deployed in the data classification system to classify the financial text data to be classified.

3. The method of claim 2, wherein, The trained data classification model is deployed in the data classification system to classify the financial text data to be classified, including: Data upload: through the file upload module in the user interface, the user uploads the financial text data file to be classified to the data classification system; Data preprocessing: the system automatically preprocesses the uploaded financial text data; Model application: using the trained data classification model, the text data preprocessed in step (4.2) is automatically classified and labeled; Result storage: the data set after classification and labeling is stored in the system database and the specified local storage location, so as to facilitate subsequent query, analysis and application; Report generation: provide user interface for users to view classification results, and the system can automatically generate data analysis charts according to the classification results to provide users with in-depth analysis and decision support.

4. A pre-trained model-based financial text data classification apparatus, characterized by, Include: Preprocessing module, for preprocessing the financial text data, building data containing classification identification as training data, wherein the number of classification identification is not less than one; Model training module, for training data classification model based on pre-training model BERT using training data, in the process of back propagation, the model is trained by subnetwork fine-tuning method, including: Initialization parameters: set learning rate , probability p, initial pre-trained model weights and hyperparameters for AdamW optimizer and where and are the exponential decay rates for the first and second moment estimates of the gradient Compute the gradient: in the t-th iteration, compute the gradient of the loss function with respect to the parameters corresponding to the current model parameters where is the model weight after the t-th iteration, is the loss function with as the parameters; Generating subnetwork gradient mask: at the t-th iteration, a 0-1 mask matrix is generated , the elements of which are randomly drawn from a Bernoulli distribution with probability p, for selectively updating the model parameters; Apply gradient mask: apply the gradient with the mask matrix generated in (3.3) Perform Hadamard product to get the masked gradient such that the gradient matrix only retains the elements in the mask matrix that have a corresponding 1 in updating the first moment estimate: ; Updating the second moment estimate: ; Bias correction: bias correction is performed on the first moment estimate and the second moment estimate to obtain a corrected first moment estimate and a corrected second moment estimate ; updating the model weights according to the corrected first moment estimate and the second moment estimate, and a learning rate ; Iterative training: repeat the above process until the model converges or reaches the preset iteration number; Fine-tuning training is completed: in the fine-tuning process, the update of the model weight is realized through the selective update of the mask matrix ​ Data classification module, including the trained data classification model, for classifying financial text data to be classified.

5. The financial text data classification apparatus according to claim 4, characterized by Also include: Deployment module, for deploying the trained data classification model in the data classification system to classify the financial text data to be classified.

6. The financial text data classifying apparatus according to claim 5, wherein The trained data classification model is deployed in the data classification system to classify the financial text data to be classified, including: Data upload: through the file upload module in the user interface, the user uploads the financial text data file to be classified to the data classification system; Data preprocessing: the system automatically preprocesses the uploaded financial text data; Model application: using the trained data classification model, the text data preprocessed in step (4.2) is automatically classified and labeled; Result storage: the data set after classification and labeling is stored in the system database and the specified local storage location, so as to facilitate subsequent query, analysis and application; Report generation: provide user interface for users to view classification results, and the system can automatically generate data analysis charts according to the classification results to provide users with in-depth analysis and decision support. Report generation: a user interface is provided for users to view the classification results, and the system can automatically generate data analysis charts based on the classification results for in-depth analysis and decision support of users.

7. An electronic device comprising a memory and a processor, characterized in that The memory is coupled with the processor; wherein the memory is configured to store program data, and the processor is configured to execute the program data to implement the financial text data classification method based on the pre-trained model according to any one of claims 1-3.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the financial text data classification method based on the pre-trained model according to any one of claims 1-3.

9. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the financial text data classification method based on the pre-trained model according to any one of claims 1-3.

Images