Search

CN-122021881-A - Data message generation method, device, equipment and storage medium

CN122021881ACN 122021881 ACN122021881 ACN 122021881ACN-122021881-A

Abstract

The application belongs to the technical field of artificial intelligence, and relates to a data message generation method, a device, equipment and a storage medium, wherein data to be analyzed are obtained; inputting the data to be analyzed into a preset joint evaluation component to obtain a preliminary evaluation result, distributing a target inference engine for the data to be analyzed according to the preliminary evaluation result, driving the target inference engine to generate a plurality of inference interpretation paths, obtaining an optimal inference interpretation path, analyzing the optimal inference interpretation path to obtain all inference nodes contained in the optimal inference interpretation path, and generating an inference report in a natural language text form aiming at the data to be analyzed by combining all the inference nodes and the preliminary evaluation result. The method is applied to a data message generation scene of carrying out natural language text form on the identified financial business risk data, so that when the financial business risk data are identified, an interpretable data message is generated, and the method can assist business handling personnel in understanding and persuade business handling users.

Inventors

  • QU XIAOYANG

Assignees

  • 平安科技（深圳）有限公司

Dates

Publication Date
20260512
Application Date
20260113

Claims (10)

  1. 1. The data message generating method is characterized by comprising the following steps: Acquiring data to be analyzed; inputting the data to be analyzed into a preset joint evaluation component to obtain a preliminary evaluation result; distributing a target inference engine for the data to be analyzed according to the preliminary evaluation result; Driving the target reasoning engine to generate a plurality of reasoning interpretation paths; Sequentially inputting the plurality of reasoning interpretation paths into a preset reasoning process scoring model to obtain an optimal reasoning interpretation path; analyzing the optimal reasoning interpretation path to obtain all reasoning nodes contained in the optimal reasoning interpretation path; and combining all the inference nodes with the preliminary evaluation result to generate an inference report in a natural language text form aiming at the data to be analyzed.
  2. 2. The method for generating a data message according to claim 1, wherein the preset joint evaluation component includes a risk level evaluation sub-component and a confidence level evaluation sub-component, and the step of inputting the data to be analyzed into the preset joint evaluation component to obtain a preliminary evaluation result specifically includes: acquiring a risk level output by the risk level evaluation sub-component aiming at the data to be analyzed; acquiring an evaluation confidence level generated by the confidence level evaluation sub-component aiming at the risk level; and filling the risk level and the evaluation confidence as preliminary evaluation parameters into a preset evaluation result output template to generate the preliminary evaluation result.
  3. 3. The method for generating a data packet according to claim 2, wherein the step of obtaining the risk level outputted by the risk level evaluation sub-component for the data to be analyzed specifically comprises: extracting risk characteristic factors from the data to be analyzed to obtain all risk characteristic factors contained in the data to be analyzed; According to the risk weights respectively corresponding to different risk characteristic factors, comprehensive calculation is carried out to obtain comprehensive risk weights corresponding to the data to be analyzed; and determining the risk level of the data to be analyzed based on the comprehensive risk weight.
  4. 4. The method for generating a data message according to claim 2, wherein the step of obtaining the evaluation confidence level generated by the confidence level evaluation sub-component for the risk level specifically comprises: Identifying the component number information of the risk level evaluation sub-component from a preset risk evaluation confidence coefficient table, wherein the preset risk evaluation confidence coefficient table comprises the component number information and the evaluation confidence coefficient information of all sub-components used for risk level evaluation, and the component number information and the evaluation confidence coefficient information are in a key value pair cache form; And identifying the evaluation confidence of the risk level evaluation sub-component from a preset risk evaluation confidence list according to the component number information.
  5. 5. The method for generating a data packet according to claim 1, wherein the step of allocating a target inference engine to the data to be analyzed according to the preliminary evaluation result specifically comprises: Identifying the risk level and the evaluation confidence level contained in the preliminary evaluation result; and distributing a target inference engine to the data to be analyzed based on the combination of the risk level and the evaluation confidence.
  6. 6. The method for generating a data message according to claim 5, wherein the step of assigning a target inference engine to the analysis data based on a combination of the risk level and the evaluation confidence level specifically comprises: When the risk level is high and the evaluation confidence level is low, allocating an inference engine with high-inference computing resources to the data to be analyzed, wherein the inference engine with high-inference computing resources comprises an inference engine with more inference path generation amount; Otherwise, allocating an inference engine with low-inference computing resources to the data to be analyzed, wherein the inference engine with low-inference computing resources comprises an inference engine with less inference path generation amount.
  7. 7. The data message generating method according to claim 1, wherein before the step of sequentially inputting the plurality of inference interpretation paths into a preset inference process scoring model to obtain an optimal inference interpretation path is performed, the method further comprises: Acquiring a pre-arranged reasoning explanation resource, wherein the reasoning explanation resource consists of business handling logic corresponding to a target business, business explanation criteria, and explanatory rules and policy files of industries to which the business belongs; generating a total standard interpretation path according to the reasoning interpretation resources, and setting a reasoning score for each standard interpretation path by combining historical business handling experience data; taking the total standard interpretation path and the corresponding reasoning score as reasoning knowledge to construct the reasoning process scoring model; the step of sequentially inputting the plurality of reasoning interpretation paths into a preset reasoning process scoring model to obtain an optimal reasoning interpretation path specifically comprises the following steps: sequentially inputting the plurality of reasoning interpretation paths into the reasoning process scoring model; identifying a standard interpretation path corresponding to each reasoning interpretation path respectively in a comparison mode; Determining the reasoning scores of the plurality of reasoning explanation paths according to the reasoning scores of each standard explanation path; and screening the reasoning explanation path with the highest reasoning score from the plurality of reasoning explanation paths, and marking the reasoning explanation path as the optimal reasoning explanation path.
  8. 8. A data message generating apparatus, comprising: the data acquisition module to be analyzed is used for acquiring data to be analyzed; the preliminary evaluation result acquisition module is used for inputting the data to be analyzed into a preset joint evaluation assembly to acquire a preliminary evaluation result; The target inference engine distribution module is used for distributing a target inference engine to the data to be analyzed according to the preliminary evaluation result; The reasoning interpretation path generation module is used for driving the target reasoning engine to generate a plurality of reasoning interpretation paths; The optimal reasoning interpretation path acquisition module is used for sequentially inputting the plurality of reasoning interpretation paths into a preset reasoning process scoring model to acquire the optimal reasoning interpretation paths; The reasoning node acquisition module is used for analyzing the optimal reasoning interpretation path and acquiring all reasoning nodes contained in the optimal reasoning interpretation path; and the reasoning report generation module is used for combining all the reasoning nodes and the preliminary evaluation result to generate a reasoning report in a natural language text form aiming at the data to be analyzed.
  9. 9. A computer device comprising a memory and a processor, the memory having stored therein computer readable instructions which when executed by the processor implement the steps of the data message generation method of any of claims 1 to 7.
  10. 10. A computer readable storage medium having stored thereon computer readable instructions which when executed by a processor implement the steps of the data message generation method of any of claims 1 to 7.

Description

Data message generation method, device, equipment and storage medium Technical Field The application relates to the technical field of artificial intelligence, and relates to a data message generation method, device, equipment and storage medium, which are applied to a data message generation scene for carrying out natural language text on identified business risk data. Background In an intelligent financial wind control scene, tasks such as transaction fraud detection, credit application approval and the like have extremely high requirements on the accuracy and the interpretability of the model. The traditional model generally only outputs final judgment, such as 'pass' or 'reject', lacks clear and auditable decision logic, is difficult to meet the compliance requirement of financial supervision, has variable financial fraud means, is difficult to quickly adapt to a novel risk mode, and cannot realize the generation of an interpretable business decision data message on the premise of low reasoning cost and high response speed. Disclosure of Invention The embodiment of the application aims to provide a data message generation method, a device, equipment and a storage medium, which are used for solving the technical problem that an interpretable business decision data message cannot be quickly generated during the current financial business risk identification. In a first aspect, an embodiment of the present application provides a data packet generating method, which adopts the following technical scheme: A data message generation method comprises the following steps: Acquiring data to be analyzed; inputting the data to be analyzed into a preset joint evaluation component to obtain a preliminary evaluation result; distributing a target inference engine for the data to be analyzed according to the preliminary evaluation result; Driving the target reasoning engine to generate a plurality of reasoning interpretation paths; Sequentially inputting the plurality of reasoning interpretation paths into a preset reasoning process scoring model to obtain an optimal reasoning interpretation path; analyzing the optimal reasoning interpretation path to obtain all reasoning nodes contained in the optimal reasoning interpretation path; and combining all the inference nodes with the preliminary evaluation result to generate an inference report in a natural language text form aiming at the data to be analyzed. In a second aspect, an embodiment of the present application further provides a data packet generating device, which adopts the following technical scheme: A data message generation apparatus, comprising: the data acquisition module to be analyzed is used for acquiring data to be analyzed; the preliminary evaluation result acquisition module is used for inputting the data to be analyzed into a preset joint evaluation assembly to acquire a preliminary evaluation result; The target inference engine distribution module is used for distributing a target inference engine to the data to be analyzed according to the preliminary evaluation result; The reasoning interpretation path generation module is used for driving the target reasoning engine to generate a plurality of reasoning interpretation paths; The optimal reasoning interpretation path acquisition module is used for sequentially inputting the plurality of reasoning interpretation paths into a preset reasoning process scoring model to acquire the optimal reasoning interpretation paths; The reasoning node acquisition module is used for analyzing the optimal reasoning interpretation path and acquiring all reasoning nodes contained in the optimal reasoning interpretation path; and the reasoning report generation module is used for combining all the reasoning nodes and the preliminary evaluation result to generate a reasoning report in a natural language text form aiming at the data to be analyzed. In a third aspect, an embodiment of the present application further provides a computer device, which adopts the following technical scheme: a computer device comprising a memory and a processor, said memory having stored therein computer readable instructions which when executed by said processor implement the steps of the data message generation method described above. In a fourth aspect, an embodiment of the present application further provides a computer readable storage medium, which adopts the following technical solutions: a computer readable storage medium having stored thereon computer readable instructions which when executed by a processor perform the steps of a data message generation method as described above. Compared with the prior art, the embodiment of the application has the following main beneficial effects: The data message generation method includes the steps of obtaining data to be analyzed, inputting the data to be analyzed into a preset joint evaluation component to obtain a preliminary evaluation result, distributing a target inference engine for the dat