Green item classification and model training method and device, equipment, medium and product

Abstract

The invention provides a green item classification method which can be applied to the field of artificial intelligence, and the method comprises the steps: obtaining target data which at least comprises at least one of the name of a target item, a loan item in the target item, and a loan purpose in the target item; based on the classification model, predicting a classification identifier for the target data, the classification identifier corresponding to one of a plurality of statistical identifier fields related to the green industry; wherein the classification model is obtained based on target sample training, and the target sample at least comprises a first sample and a second sample; wherein the first sample at least comprises financial data related to the green industry; wherein the second sample at least comprises classification basis information related to the green industry and / or description information related to the green industry. The invention further provides a classification model training method, and a device, equipment, a medium and a product for executing the method.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and more specifically, to a green project classification and model training method, device, equipment, medium and product. Background Art

[0002] In green finance lending, banks typically need to identify and classify corporate loan projects as green industries based on standards such as the Green Industry Guidance Catalogue. Currently, financial institutions rely primarily on traditional text matching technology to identify these classification labels. This technology uses word embedding models (Word2Vec) or the Term Frequency-Inverse Document Frequency (TF-IDF) algorithm to convert project description text into static word vectors. These vectors are then matched against a predefined library of green labels using cosine similarity, ultimately outputting a single classification label. This solution suffers from fundamental flaws: At the semantic understanding level, the static word vectors generated by Word2Vec fail to capture dynamic contextual semantics (for example, the ambiguous semantics of "photovoltaic" in energy or building contexts), while TF-IDF completely ignores word order and struggles to distinguish between synonyms and polysemy. At the data processing level, it is unable to generate effective vectors for specialized terms outside the training corpus (such as the names of new environmental protection technologies). Furthermore, it relies solely on historical project form data and fails to incorporate authoritative classification criteria from green labeling standard documents, resulting in low accuracy in identifying long-tail categories. At the output mechanism level, the single-label output lacks fault tolerance, resulting in a high rate of missed edge cases for complex loan projects (such as integrated energy retrofits), and an inability to support multiple candidate recommendation results. These flaws hinder the accuracy of classification identification, significantly increase the time required for manual pre-loan review, and potentially lead to compliance risks in which highly polluting projects are mistakenly labeled as green projects. Summary of the Invention

[0003] In view of this, the present application provides a green project classification and model training method, device, equipment, medium and product.

[0004] One aspect of the present application provides a green project classification method, comprising: obtaining target data, the target data including at least one of the name of the target project, the loan project in the target project, and the purpose of the loan in the target project; based on a classification model, predicting a classification identifier for the target data, the classification identifier corresponding to one of a plurality of statistical identification fields related to the green industry; wherein the classification model is obtained based on target sample training, the target sample including at least a first sample and a second sample; wherein the first sample includes at least financial data related to the green industry; wherein the second sample includes at least classification basis information related to the green industry, and / or descriptive information related to the green industry.

[0005] According to an embodiment of the present application, before predicting the classification identification for the target data, the method also includes: generating text information for describing the target project based on the name, and / or the loan project name, and / or the purpose of the loan; predicting the classification identification for the target data includes: inputting the text information into the classification model to generate a classification identification for the target data.

[0006] According to an embodiment of the present application, the classification model includes an embedding layer, an encoding layer, and a classification layer, and predicts a classification identifier for the target text, including: based on the embedding layer, converting the text information into a word embedding vector; based on the encoding layer, context encoding the word embedding vector to generate a dynamic semantic representation vector; based on the classification layer, generating at least one classification identifier according to the dynamic semantic representation vector.

[0007] Another aspect of the present application provides a classification model training method, including: obtaining a first sample, where the source of the first sample is a financial data table related to financial transactions, and the financial data table includes at least financial data related to the green industry; obtaining a second sample, where the source of the second sample is a standard document of the green industry, including classification basis information related to the green industry, and / or descriptive information related to the green industry; training a target model based on the first sample and / or the second sample to obtain a classification model.

[0008] According to an embodiment of the present application, obtaining a first sample includes: obtaining a financial data table related to financial transactions; generating a first sample based on data in the financial data table, the first sample including at least one of a customer name, a project name, a loan purpose, and a classification identifier, the classification identifier being used to mark the green industry classification to which the first sample belongs.

[0009] According to an embodiment of the present application, obtaining a second sample includes: obtaining a standard document for the green industry; generating at least one second sample based on the names of the statistical identification fields recorded in the standard document as training sample labels, the classification basis of the statistical identification fields, and / or the descriptive text as input in the training sample.

[0010] Another aspect of the present application provides a green project classification device, including: a first acquisition module, used to acquire target data, the target data at least including at least one of the name of the target project, the loan project in the target project, and the purpose of the loan in the target project; and, a first prediction module, used to predict a classification identifier for the target data, the classification identifier corresponding to one of a plurality of statistical identification fields related to the green industry; wherein the classification model is obtained based on target sample training, the target sample includes at least a first sample and a second sample; wherein the first sample includes at least financial data related to the green industry; wherein the second sample includes at least classification basis information related to the green industry, and / or descriptive information related to the green industry.

[0011] Another aspect of the present application provides a classification model training device, including: a second acquisition module, used to obtain a first sample, the source of the first sample is a financial data table related to financial transactions, and the financial data table includes at least financial data related to the green industry; a third acquisition module, used to obtain a second sample, the source of the second sample is a standard document of the green industry, including classification basis information related to the green industry, and / or descriptive information related to the green industry; and a training module, used to train a target model based on the first sample and / or the second sample to obtain a classification model.

[0012] Another aspect of the present application provides an electronic device comprising: at least one processor; and a memory connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the green project classification method and / or classification model training method of any one of the aforementioned embodiments.

[0013] Another aspect of the present application provides a computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute the green project classification method and / or classification model training method according to any one of the aforementioned embodiments.

[0014] Another aspect of the present application provides a computer program product, including a computer program / instruction, characterized in that when the computer program / instruction is executed by a processor, it implements the operations of the green project classification method and / or classification model training method of any of the aforementioned embodiments.

[0015] According to the embodiments of the present application, the green project classification method or classification model training method provided by the present application has at least one of the following beneficial effects: on the one hand, by integrating the first sample in the financial data table, the model can capture the financial characteristics related to the green project, thereby improving the recognition ability of the project's financial background and supporting more accurate green industry classification. On the other hand, the second sample is derived from the green industry standard document, which supplements the classification identification category information that may be missing in the first sample, thereby enhancing the model's ability to understand environmental attributes, ensuring that the classification model can fully cover the green project category, and improving recognition accuracy and efficiency. In addition, the introduction of the second sample enables the model to adapt to different green project standards and classification bases, so when dealing with complex and diverse projects, the model exhibits greater adaptability and robustness, supporting the intelligent transformation of business processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The above and other objects, features and advantages of the present application will become more apparent through the following description of the embodiments of the present application with reference to the accompanying drawings, in which:

[0017] Figure 1 The following schematically illustrates an application scenario of the green project classification method according to an embodiment of the present application;

[0018] Figure 2 A flowchart schematically illustrates a green project classification method according to an embodiment of the present application;

[0019] Figure 3 Another flowchart of the green project classification method according to an embodiment of the present application is schematically shown;

[0020] Figure 4 The following schematically illustrates a flow chart for generating classification identifiers in a green project classification method according to an embodiment of the present application;

[0021] Figure 5 A flowchart of a classification model training method according to an embodiment of the present application is schematically shown;

[0022] Figure 6 A flowchart of obtaining a first sample in a classification model training method according to an embodiment of the present application is schematically shown;

[0023] Figure 7 A flowchart of obtaining a second sample in a classification model training method according to an embodiment of the present application is schematically shown;

[0024] Figure 8 Schematically shows a block diagram of a green project classification device according to an embodiment of the present application;

[0025] Figure 9 A block diagram schematically illustrates a classification model training device according to an embodiment of the present application; and

[0026] Figure 10 A block diagram of an electronic device suitable for implementing the method described above according to an embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0027] Hereinafter, embodiments of the present application will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present application. In the detailed description below, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present application. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessarily confusing the concepts of the present application.

[0028] The terms used herein are only for describing specific embodiments and are not intended to limit this application. The terms "comprise," "include," etc. used herein indicate the presence of the features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0029] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0030] When expressions such as "at least one of A, B, and C, etc." are used, they should generally be interpreted in accordance with the meaning commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0031] In the technical solution of this application, the user information involved (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data comply with relevant laws, regulations and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0032] An embodiment of the present application provides a green project classification method, including: obtaining target data, the target data including at least one of the name of the target project, the loan project in the target project, and the purpose of the loan in the target project; based on a classification model, predicting a classification identifier for the target data, the classification identifier corresponding to one of multiple statistical identification fields related to the green industry; wherein the classification model is obtained based on target sample training, and the target sample includes at least a first sample and a second sample; wherein the first sample includes at least financial data related to the green industry; wherein the second sample includes at least classification basis information related to the green industry, and / or descriptive information related to the green industry.

[0033] Figure 1 The application scenario diagram of the green project classification method according to the embodiment of the present application is schematically shown.

[0034] like Figure 1As shown, application scenario 100 includes network interactions between multiple terminal devices and a server. Network 104, serving as the medium for the communication link, connects a first terminal device 101, a second terminal device 102, a third terminal device 103, and a server 105. Network 104 can utilize various connection methods, such as wired connections, wireless communication links, or fiber optic cables, to ensure stable and fast data transmission.

[0035] Users can interact with the server 105 through the first terminal device 101, the second terminal device 102, and the third terminal device 103 to receive or send messages related to the green project classification. These terminal devices can be installed with various communication client applications, such as applications for green project management, web browsers, instant messaging tools, email clients, and social platform software (for example only).

[0036] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be different types of electronic devices that have display screens and support web browsing functions, including but not limited to smart phones, tablet computers, laptop computers, and desktop computers.

[0037] Server 105 is responsible for providing various services, such as a backend management server that supports users browsing green project-related websites through their terminal devices. The backend management server can analyze and process received user requests and feed back the processing results (e.g., green project information or data obtained or generated based on the user request) to the terminal device.

[0038] It should be noted that the green project classification method provided in the embodiments of the present application is typically executed by the server 105. Accordingly, the green project classification device is generally located in the server 105. Furthermore, the green project classification method may also be executed by a server or server cluster other than the server 105, which servers are capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or the server 105. Accordingly, the green project classification device may also be located in a server or server cluster other than the server 105.

[0039] It should be understood that Figure 1 The number of terminal devices, networks, and servers in the figure is for illustration only. Depending on the specific implementation requirements, any number of terminal devices, networks, and servers may be used to meet different business requirements and application scenarios.

[0040] Figure 2 The flowchart of the green project classification method according to an embodiment of the present application is schematically shown.

[0041] like Figure 2As shown, the green project classification method may at least include operations S210 to S220.

[0042] In operation S210, target data is acquired. The target data includes at least one of the target project name, the loan project within the target project, and the loan purpose within the target project. The target data may be structured fields representing loan project attributes in a credit business. For example, the customer name, loan project name, and loan purpose fields may be captured in real time through a financial institution's internal data interface.

[0043] For example, the target data combination can be Scenario 1: only the purpose of the loan (such as "biomass power generation equipment procurement"); or Scenario 2: project name + loan purpose (such as "offshore wind farm construction, wind turbine purchase"); or Scenario 3: customer name + project name + loan purpose (such as "an environmental protection company, a waste battery recycling plant, recycling and processing equipment procurement").

[0044] In operation S220, a classification identifier is predicted for the target data based on the classification model. The classification identifier corresponds to one of multiple statistical identifier fields related to green industries. A classification model is a machine learning model or deep learning model used to predict the green project category to which the target data belongs. It utilizes natural language processing techniques, such as the BERT model, to process complex semantic information. The classification model analyzes the semantic features of the project description text to generate statistical identifier fields related to green industries, helping to identify the environmental attributes of the project. The model is optimized using training, validation, and test sets to improve prediction accuracy.

[0045] The classification model is obtained based on target sample training, and the target sample includes at least a first sample and a second sample;

[0046] The first sample includes at least financial data related to green industries. Financial data related to green industries refers to economic information related to green project financing, such as the name of the target project, the loan items within the target project, the purpose of the loan within the target project, and may also include the loan amount, interest rate, and repayment period. For example, the loan amount and interest rate of a specific project can be extracted from a bank's green credit records and used to train a classification model to ensure that the model can identify the relationship between financial attributes and green projects based on the loan amount and interest rate.

[0047] The second sample includes at least green industry-related classification basis information and / or green industry-related descriptive information. The green industry-related classification basis information and descriptive information refer to textual information used to define green project standards, including policies, regulations, industry standards, etc. The green industry-related classification basis information and / or green industry-related descriptive information are used as training sample input to supplement categories that may be missing in conventional samples. For example, definitions and standards for green buildings are extracted from environmental policy documents and used to train a classification model to ensure that the model can identify and classify the environmental attributes of a project.

[0048] According to the embodiments of this application, by integrating regular samples with supplementary samples (second samples), the classification model can more comprehensively cover green project categories, improve identification accuracy and efficiency, reduce the burden of manual judgment, and support the digital transformation of business processes. This integration strategy makes the model more adaptable and accurate when handling complex projects.

[0049] Figure 3 Another flowchart of the green project classification method according to an embodiment of the present application is schematically shown.

[0050] like Figure 3 As shown, based on the above embodiment, the green project classification method may further include operation S310.

[0051] In operation S310, text information describing the target project is generated based on the name, and / or the name of the loan project, and / or the purpose of the loan. The text information can be a descriptive text formed by merging relevant information of the target project, which is used to input into the classification model for analysis. Specifically, the text information can be formed by connecting the customer name, project name and loan purpose with commas to form a complete project description, so that the classification model can perform semantic analysis. For example, "Customer Name: Zhang San", "Project Name: Green Building", and "Loan Purpose: Environmentally Friendly Material Procurement" are merged into "Zhang San, Green Building, Environmentally Friendly Material Procurement". It is also possible to use connecting words such as auxiliary words, verbs, and adverbs to connect the various contents into a complete sentence, for example, "Zhang San executed a green building-related project, the content of which is environmentally friendly material procurement."

[0052] Operation S220 may include operation S320.

[0053] In operation S320, the text information is input into the classification model to generate a classification identifier for the target data. For example, after tokenization, the text information is input into the Bidirectional Encoder Representations from Transformers (BERT) model to generate a semantic embedding vector. This is then processed through a linear fully connected layer (FC) and a sigmoid activation function to generate a classification identifier. For example, the text information "Zhang San, green building, environmentally friendly material procurement" is processed by the BERT model to generate a statistical identifier for the green building category.

[0054] According to the embodiments of this application, by integrating project-related information into text and inputting it into a classification model for processing, the accuracy and efficiency of green project classification can be significantly improved. This method not only reduces the complexity of manual participation but also enhances the model's ability to understand multi-dimensional information, enabling it to more accurately identify the environmental attributes of projects and support the intelligent and digital transformation of business processes.

[0055] Figure 4 The flowchart of generating classification identification in the green project classification method according to an embodiment of the present application is schematically shown.

[0056] like Figure 4 As shown, based on the above embodiment, the classification model includes an embedding layer, an encoding layer, and a classification layer, and operation S320 may include operations S410 to S430.

[0057] In operation S410, the text information is converted into word embedding vectors based on the embedding layer. Specifically, the embedding layer converts the input text information into high-dimensional word vectors by using pre-trained embeddings of the Bidirectional Encoder Representations from Transformers (BERT) model.

[0058] In operation S420, context encoding is performed on the word embedding vector based on the encoding layer to generate a dynamic semantic representation vector. The dynamic semantic representation vector refers to performing context analysis on the word embedding vector through the encoding layer to generate a semantic representation that can capture the relationship between words.

[0059] In operation S430, at least one classification identifier is generated based on the dynamic semantic representation vector using the classification layer. The classification identifier is generated by analyzing the dynamic semantic representation vector through the classification layer to identify the green category to which the project belongs. The classification layer may include, for example, a linear fully connected layer and an activation function. The dynamic semantic representation vector is processed using the linear fully connected layer and the Sigmoid activation function to generate a classification identifier (also known as a classification identifier) ​​for the project, indicating the green industry category to which it belongs.

[0060] According to the embodiments of this application, by processing text information layer by layer into word embedding vectors, dynamic semantic representation vectors, and finally classification identifiers, the classification model can more accurately capture the semantic information in the text and achieve efficient green project classification. This multi-layer processing approach improves the accuracy and robustness of classification, supports the intelligentization of business processes, and reduces the complexity of manual judgment.

[0061] Figure 5 The flowchart of the classification model training method according to an embodiment of the present application is schematically shown.

[0062] like Figure 5 As shown, based on the above embodiments, the classification model training method may include operations S510 to S530.

[0063] In operation S510, a first sample is obtained. The source of the first sample is a financial data table related to financial transactions, and the financial data table includes at least financial data related to green industries. The financial data table may contain database records of financial information related to green industries, and the database records are used to identify the financial background and green attributes of the project. Specifically, the first sample is formed by extracting relevant fields such as customer name, project name, loan purpose, and classification identification from the financial data table to form a project description text as the input sample in the first sample, and its corresponding green project identification is used as the training label. For example, a loan record with the project name "Green Building" is extracted from the bank's credit management system, including the loan amount and purpose, as the input sample of the first sample, and its corresponding green project identification is used as the training label.

[0064] In operation S520, a second sample is obtained. The second sample is sourced from a standard document on the green industry, including classification basis information and / or descriptive information related to the green industry. Standard documents for the green industry refer to documents that contain green project classification standards and descriptive information, providing the basis and detailed description of project classification. The second sample is formed by extracting the classification basis and descriptive text for each classification identifier from the standard document to form a training sample input, thereby supplementing the category information that may be missing in the first sample.

[0065] For example, definitions and standards about green buildings are extracted from environmental policy documents as the second sample for training the model to ensure that the model can identify the environmental attributes of the project.

[0066] In operation S530, a target model is trained based on the first sample and / or the second sample to obtain a classification model. The target model may be a model for green project classification obtained through training using a machine learning algorithm or a deep learning algorithm. The target model is trained by combining the first sample and the second sample to improve the model's ability to identify green projects and ensure that it can accurately classify multiple classification identification categories.

[0067] According to the embodiments of this application, by combining financial data tables with samples of green industry standard documents for model training, the classification model can more comprehensively cover green project categories, improving recognition accuracy and efficiency. This training method not only supplements the green category information that may be missing in conventional financial data, but also enhances the model's understanding of environmental attributes, supporting the intelligent and digital transformation of business processes.

[0068] Figure 6 The flowchart of obtaining the first sample in the classification model training method according to an embodiment of the present application is schematically shown.

[0069] like Figure 6 As shown, based on the above embodiment, operation S510 may include operations S610 to S620.

[0070] In operation S610, a financial transaction-related financial data table is obtained. The financial transaction-related financial data table refers to a database that records detailed information related to loan projects in a bank or financial institution. The table contains multiple fields that describe the financial and business attributes of each project.

[0071] In operation S620, a first sample is generated based on the data in the financial data table. The first sample includes at least one of the following: the client name, the project name, the loan purpose, and a classification identifier. The classification identifier is used to identify the green industry category to which the first sample belongs. Generating the first sample involves extracting relevant information from the financial data table and organizing it into standardized input samples for training the classification model. The classification identifier is a label used to identify the green industry category to which the project belongs.

[0072] Figure 7 The flowchart of obtaining the second sample in the classification model training method according to an embodiment of the present application is schematically shown.

[0073] like Figure 7 As shown, based on the above embodiment, S520 may include operations S710 to S720.

[0074] In operation S710, a standard document for the green industry is obtained. This document contains green project classification standards and detailed descriptions, typically published by a specific organization to guide and regulate green industry project classification. For example, relevant information can be extracted from a green building standards manual published by an environmental protection department as the source of the standard document.

[0075] In operation S720, at least one second sample is generated based on the names of the various statistical identification fields recorded in the standard document as training sample labels, the classification basis of the statistical identification fields, and / or the descriptive text as input in the training sample. The statistical identification field refers to the field defined in the standard document for identifying the category of green projects, including the name, classification basis, and descriptive text. The second sample forms the input for training the classification model by extracting information from these fields. For example, the definition and classification standards of "green building" are extracted from the green building standard manual as the second sample for training the model to ensure that the model can recognize the environmental attribute of the project "green building".

[0076] According to the embodiments of this application, by combining information from green industry standard documents to generate a second sample, the classification model can more comprehensively cover green project categories, improving the accuracy and efficiency of identification. This approach not only supplements the green category information that may be missing in conventional financial data, but also enhances the model's understanding of environmental attributes, supporting the intelligent and digital transformation of business processes.

[0077] According to an embodiment of the present application, multiple classification identifiers can be generated for the target data. For example, by outputting k classification identifier prediction values, the model can capture more potential classification information, thereby improving the overall recognition accuracy. For example, when k=3, that is, when 3 prediction values ​​are output, the business personnel can select one from the 3 classification identifiers output by the model and click to fill in the system. In addition, the business personnel can also make independent judgments based on their own professional knowledge and experience and in combination with the three classification identifiers output, and select the green identifier that best suits the actual situation to fill in.

[0078] According to the embodiment of the present application, the method of outputting multiple prediction values ​​not only enhances the recognition ability and accuracy of the model, but also provides business personnel with a more flexible and efficient operation method, thereby promoting the efficiency improvement of green project identification.

[0079] Figure 8 The block diagram of the green project classification device according to an embodiment of the present application is schematically shown.

[0080] like Figure 8 As shown, the green project classification device 800 may include a first acquisition module 810 and a first prediction module 820 .

[0081] First acquisition module 810 is used to acquire target data, which includes at least one of the name of the target project, the loan project within the target project, and the purpose of the loan within the target project. In some embodiments, first acquisition module 810 can be used to perform operation S210 in the green project classification method described above, which is not further described here.

[0082] The first prediction module 820 is used to predict a classification identifier for the target data, where the classification identifier corresponds to one of a plurality of statistical identification fields related to the green industry. In some embodiments, the first prediction module 820 can be used to perform operation S220 in the green project classification method described above, which will not be described in detail here.

[0083] The classification model is obtained based on target sample training, and the target sample includes at least a first sample and a second sample; the first sample includes at least financial data related to the green industry; the second sample includes at least classification basis information related to the green industry, and / or descriptive information related to the green industry.

[0084] According to an embodiment of the present application, the green project classification device may include a first generation module.

[0085] The first generation module is used to generate text information describing the target project based on the name, and / or the loan project name, and / or the loan purpose. In some embodiments, the preset module 6 can be used to perform operation S310 in the above-mentioned green project classification method, which is not further described here.

[0086] The first prediction module may include a second generation module.

[0087] The second generation module is used to input the text information into the classification model to generate a classification identifier for the target data. In some embodiments, the preset 7 module can be used to perform operation S320 in the above-mentioned green project classification method, which will not be described in detail here.

[0088] According to an embodiment of the present application, the classification model includes an embedding layer, an encoding layer, and a classification layer, and the second generation module may include a conversion module, an encoding module, and a classification module.

[0089] The conversion module is used to convert the text information into a word embedding vector based on the embedding layer. In some embodiments, the conversion module can be used to perform operation S410 in the above-mentioned green project classification method, which will not be described in detail here.

[0090] The encoding module is used to perform context encoding on the word embedding vector based on the encoding layer to generate a dynamic semantic representation vector. In some embodiments, the encoding module can be used to perform operation S420 in the above-mentioned green project classification method, which will not be described in detail here.

[0091] The classification module is used to generate at least one classification identifier based on the dynamic semantic representation vector based on the classification layer. In some embodiments, the classification module can be used to perform operation S430 in the green project classification method described above, which will not be described in detail here.

[0092] Figure 9 A block diagram of a classification model training device according to an embodiment of the present application is schematically shown.

[0093] like Figure 9 As shown, according to an embodiment of the present application, classification model training may include a second acquisition module 910, a third acquisition module 920, and a training module 930.

[0094] The second acquisition module is configured to acquire a first sample, where the first sample is derived from a financial data table related to financial transactions, where the financial data table includes at least financial data related to the green industry. In some embodiments, the second acquisition module can be configured to perform operation S510 in the aforementioned classification model training method, which is not further described herein.

[0095] The third acquisition module is used to obtain a second sample. The second sample is sourced from a standard document on the green industry, including classification basis information related to the green industry and / or descriptive information related to the green industry. In some embodiments, the third acquisition module can be used to perform operation S520 in the above-mentioned classification model training method, which is not further described here.

[0096] The training module is used to train the target model according to the first sample and / or the second sample to obtain a classification model. In some embodiments, the training module can be used to perform operation S530 in the above classification model training method, which will not be described in detail here.

[0097] According to an embodiment of the present application, the second acquisition module may include a first submodule and a second submodule.

[0098] The first submodule is used to obtain a financial data table related to financial transactions. In some embodiments, the first submodule can be used to perform operation S610 in the above classification model training method, which will not be described in detail here.

[0099] The second submodule is configured to generate a first sample based on the data in the financial data table. The first sample includes at least one of a client name, a project name, a loan purpose, and a classification identifier, where the classification identifier is used to identify the green industry category to which the first sample belongs. In some embodiments, the second submodule can be configured to perform operation S620 in the aforementioned classification model training method, which is not further described here.

[0100] According to an embodiment of the present application, the third acquisition module may include a third submodule and a fourth submodule.

[0101] The third submodule is used to obtain standard documents of the green industry. In some embodiments, the third submodule can be used to perform operation S710 in the above classification model training method, which will not be described in detail here.

[0102] The fourth submodule is configured to generate at least one second sample based on the names of the statistical identification fields recorded in the standard document as training sample labels, the classification basis of the statistical identification fields, and / or the descriptive text as input in the training sample. In some embodiments, the fourth submodule can be configured to perform operation S720 in the aforementioned classification model training method, which is not further described here.

[0103] According to the embodiments of the present application, any number of modules, submodules, units, and subunits, or at least part of the functions of any number of them, can be implemented in one module. According to the embodiments of the present application, any one or more of the modules, submodules, units, and subunits can be split into multiple modules for implementation. According to the embodiments of the present application, any one or more of the modules, submodules, units, and subunits can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application-specific integrated circuit (ASIC), or can be implemented by hardware or firmware in any other reasonable way of integrating or packaging the circuit, or can be implemented in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, according to the embodiments of the present application, one or more of the modules, submodules, units, and subunits can be at least partially implemented as a computer program module, which can perform the corresponding functions when the computer program module is executed.

[0104] For example, any of the first acquisition module 810, the first prediction module 820, and / or the second acquisition module 910, the third acquisition module 920, and the training module 930 can be combined into one module / unit / sub-unit for implementation, or any of these modules / units / sub-units can be split into multiple modules / units / sub-units. Alternatively, at least part of the functionality of one or more of these modules / units / sub-units can be combined with at least part of the functionality of other modules / units / sub-units and implemented in one module / unit / sub-unit. According to an embodiment of the present application, at least one of the first acquisition module 810, the first prediction module 820, and / or the second acquisition module 910, the third acquisition module 920, and the training module 930 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application-specific integrated circuit (ASIC), or can be implemented by hardware or firmware through any other reasonable means of integrating or packaging circuits, or can be implemented in any one of the three implementation methods of software, hardware, and firmware, or in an appropriate combination of any of them. Alternatively, at least one of the first acquisition module 810, the first prediction module 820, and / or the second acquisition module 910, the third acquisition module 920, and the training module 930 can be at least partially implemented as a computer program module, which can perform the corresponding function when executed.

[0105] It should be noted that the data processing system part in the embodiments of the present application corresponds to the data processing method part in the embodiments of the present application. The description of the data processing system part specifically refers to the data processing method part and will not be repeated here.

[0106] Figure 10 A block diagram of an electronic device suitable for implementing the method described above according to an embodiment of the present application is schematically shown. Figure 10 The electronic device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0107] like Figure 10As shown, the electronic device 1000 according to an embodiment of the present application includes a processor 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage portion 1008 into a random access memory (RAM) 1003. The processor 1001 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 1001 may also include onboard memory for caching purposes. The processor 1001 may include a single processing unit or multiple processing units for performing different actions of the method flow according to the embodiment of the present application.

[0108] Various programs and data required for the operation of the electronic device 1000 are stored in the RAM 1003. The processor 1001, the ROM 1002, and the RAM 1003 are connected to each other via a bus 1004. The processor 1001 performs various operations of the method flow according to the embodiment of the present application by executing the programs in the ROM 1002 and / or the RAM 1003. It should be noted that the programs may also be stored in one or more memories other than the ROM 1002 and the RAM 1003. The processor 1001 may also perform various operations of the method flow according to the embodiment of the present application by executing the programs stored in the one or more memories.

[0109] According to an embodiment of the present application, electronic device 1000 may further include an input / output (I / O) interface 1005, which is also connected to bus 1004. Electronic device 1000 may also include one or more of the following components connected to I / O interface 1005: an input section 1006 including a keyboard, mouse, etc.; an output section 1007 including devices such as a cathode ray tube (CRT), liquid crystal display (LCD), and speakers; a storage section 1008 including a hard disk; and a communication section 1009 including a network interface card such as a LAN card or modem. Communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to I / O interface 1005 as needed. Removable media 1011, such as a magnetic disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed in drive 1010 as needed, so that computer programs read from the removable media can be installed into storage section 1008 as needed.

[0110] According to an embodiment of the present application, the method flow according to the embodiment of the present application can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1009, and / or installed from the removable medium 1011. When the computer program is executed by the processor 1001, the above-mentioned functions defined in the system of the embodiment of the present application are executed. According to an embodiment of the present application, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0111] This application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments, or may exist independently and not be incorporated into the device / apparatus / system. The computer-readable storage medium carries one or more programs, and when the one or more programs are executed, the method according to the embodiments of this application is implemented.

[0112] According to embodiments of the present application, a computer-readable storage medium may be a non-volatile computer-readable storage medium. Examples include, but are not limited to, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0113] For example, according to an embodiment of the present application, the computer-readable storage medium may include the ROM 1002 and / or the RAM 1003 described above and / or one or more memories other than the ROM 1002 and the RAM 1003 .

[0114] An embodiment of the present application also includes a computer program product, which includes a computer program, which contains program code for executing the method provided by the embodiment of the present application. When the computer program product is run on an electronic device, the program code is used to enable the electronic device to implement the method provided by the embodiment of the present application.

[0115] When the computer program is executed by the processor 1001, the above functions defined in the system / device of the embodiment of the present application are performed. According to the embodiment of the present application, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0116] In one embodiment, the computer program can rely on tangible storage media such as optical storage devices and magnetic storage devices. In another embodiment, the computer program can also be transmitted and distributed in the form of a signal on a network medium, and downloaded and installed through the communication part 1009, and / or installed from a removable medium 1011. The program code contained in the computer program can be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above. According to an embodiment of the present application, the program code for executing the computer program provided by the embodiment of the present application can be written in any combination of one or more programming languages. Specifically, these computer programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include but are not limited to Java, C++, Python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. Where a remote computing device is involved, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0117] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It will be understood by those skilled in the art that the features described in the various embodiments of the present application can be combined and / or combined in various ways, even if such combinations or combinations are not explicitly described in the present application. In particular, without departing from the spirit and teachings of the present application, the features described in the various embodiments of the present application may be combined and / or coupled in various ways, and all of these combinations and / or couplings fall within the scope of the present application.

[0118] The embodiments of the present application have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present application. Although each embodiment has been described separately above, this does not mean that the measures in each embodiment cannot be advantageously used in combination. Without departing from the scope of the present application, those skilled in the art may make various substitutions and modifications, and these substitutions and modifications should all fall within the scope of the present application.

Claims

1. A green project classification method, characterized in that: The method comprises: Acquire target data, where the target data includes at least one of a name of a target project, a loan project in the target project, and a purpose of the loan in the target project; Predicting a classification identifier for the target data based on the classification model, where the classification identifier corresponds to one of a plurality of statistical identifier fields related to the green industry; The classification model is obtained by training based on target samples, and the target samples include at least a first sample and a second sample; Wherein, the first sample includes at least financial data related to the green industry; The second sample at least includes classification basis information related to the green industry and / or description information related to the green industry.

2. The method according to claim 1, characterized in that Before predicting the classification identification for the target data, the method further includes: Generate text information describing the target project based on the name, and / or the name of the loan project, and / or the purpose of the loan; The prediction classifies and identifies the target data, including: The text information is input into the classification model to generate a classification identifier for the target data.

3. The method according to claim 2, characterized in that The classification model includes an embedding layer, an encoding layer, and a classification layer. The predicting of a classification identifier for the target text includes: Based on the embedding layer, converting the text information into a word embedding vector; Based on the encoding layer, context encoding is performed on the word embedding vector to generate a dynamic semantic representation vector; Based on the classification layer, at least one classification identifier is generated according to the dynamic semantic representation vector.

4. A classification model training method, characterized in that: include: Obtaining a first sample, where the source of the first sample is a financial data table related to financial transactions, and the financial data table at least includes financial data related to the green industry; Obtaining a second sample, where the source of the second sample is a standard document of the green industry, including classification basis information related to the green industry and / or description information related to the green industry; A target model is trained based on the first sample and / or the second sample to obtain the classification model.

5. The method according to claim 4, characterized in that The obtaining of the first sample includes: Obtain financial data tables related to financial transactions; The first sample is generated based on the data in the financial data table, and the first sample includes at least one of the customer name, project name, loan purpose and classification identification, and the classification identification is used to mark the green industry classification to which the first sample belongs.

6. The method according to claim 5, characterized in that The obtaining of the second sample includes: Obtain standard documents for green industries; At least one second sample is generated based on the names of the statistical identification fields recorded in the standard document as training sample labels, the classification basis of the statistical identification fields, and / or the description text as input in the training sample.

7. A green project classification device, characterized in that: include: A first acquisition module is configured to acquire target data, wherein the target data includes at least one of a name of a target project, a loan project in the target project, and a purpose of the loan in the target project; and a first prediction module, configured to predict a classification identifier for the target data, the classification identifier corresponding to one of a plurality of statistical identifier fields related to the green industry; The classification model is obtained by training based on target samples, and the target samples include at least a first sample and a second sample; Wherein, the first sample includes at least financial data related to the green industry; The second sample at least includes classification basis information related to the green industry and / or description information related to the green industry.

8. A classification model training device, characterized in that: include: a second acquisition module, configured to acquire a first sample, where the source of the first sample is a financial data table related to financial transactions, wherein the financial data table at least includes financial data related to the green industry; A third module is configured to obtain a second sample, where the source of the second sample is a standard document of the green industry, including classification basis information related to the green industry and / or description information related to the green industry; as well as A training module is used to train a target model based on the first sample and / or the second sample to obtain the classification model.

9. An electronic device, characterized in that: include: at least one processor; as well as a memory connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 6.

10. A computer-readable storage medium storing computer instructions, characterized in that: in, The computer instructions are used to cause the computer to execute the method according to any one of claims 1-6.

11. A computer program product, characterized in that The computer program / instruction comprises a computer program / instruction, characterized in that when the computer program / instruction is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images