Multi-mode enterprise competitiveness condition intelligent diagnosis modeling method and system and medium

By performing feature matching and convolutional neural network model optimization on enterprise multimodal operation data, the limitations of multimodal data processing in the existing technology are solved, and accurate and real-time diagnosis of enterprise competitiveness is achieved, and accuracy is provided. Competitiveness Assessment.

CN119089799APending Publication Date: 2024-12-06BEIJING BAIYI ORIENTAL EDUCATION TECH CO LTD

Patent Information

Application Number
CN202411579138.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-07
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing multimodal large model based on deep learning is difficult to achieve accurate information extraction and real-time diagnosis of enterprise competitiveness when processing enterprise operation-related data, especially in the category processing and feature matching of multimodal data. limitation.

Method used

By acquiring and preprocessing the multimodal operation data of historical enterprises, performing feature matching, building a convolutional neural network model, optimizing the model and verifying feature parameters, calculating and evaluating index coefficients to judge the competitiveness of the enterprise, and using TF-IDF and recurrent neural networks The model extracts topic semantic information and combines linear analysis and convolutional neural network for parameter optimization.

Benefits of technology

It realizes accurate screening of enterprise multimodal data and intelligent diagnostic modeling information extraction to judge competitiveness status in real time, which can accurately evaluate the strength of enterprise competitiveness and provide real-time diagnostic results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089799A_ABST
    Figure CN119089799A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of multi-dimensional data analysis, particularly discloses a multi-modal enterprise competitiveness condition intelligent diagnosis modeling method and system and a medium, and realizes an intelligent diagnosis modeling information extraction process of accurate screening of enterprise multi-modal data and real-time judgment of enterprise competitiveness conditions. Comprising the following steps: acquiring historical enterprise multi-modal operation data, and cleaning, labeling and preprocessing the historical enterprise multi-modal operation data; performing feature matching on the preprocessed multi-dimensional operation text data and the preprocessed multi-dimensional operation numerical data to obtain operation feature parameters; constructing a convolutional neural network model, inputting the set of operation characteristic parameters as a sample into the convolutional neural network model for training, obtaining a parameter optimization model, and outputting an optimization standard value; according to the parameter optimization model, operation characteristic parameters actually collected by the current enterprise are input into the parameter optimization model for verification, and evaluation parameters are output; and extracting an evaluation index coefficient according to the evaluation parameter, and judging an enterprise competitiveness result according to the evaluation index coefficient.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of multidimensional data analysis, and in particular to a multi-modal intelligent diagnosis modeling method, system and medium for enterprise competitiveness status. Background Art

[0002] Enterprise competitiveness refers to the superiority of an enterprise over its competitors in the market, which involves multiple dimensions such as products, services, costs and marketing. This ability is not only related to the profitability and market share of an enterprise, but also an important cornerstone for the survival and development of an enterprise; multimodal data includes text, images, audio and other forms, which makes the processing and analysis of multimodal data of an enterprise more complicated. How to effectively integrate and analyze these different types of data is an important challenge.

[0003] Currently, multimodal data from a wide range of sources are analyzed and processed by building multimodal large models. With the development of deep learning technology, building dynamic change models for various types of data has also become a common processing method. However, the current modeling methods of these multimodal large models based on deep learning have limitations in the intelligent analysis of the competitiveness of enterprises.

[0004] In particular, when processing data related to enterprise operations, there are problems with the classification of multimodal data and the precise matching of different data features; with massive multimodal data, it is difficult to achieve accurate information extraction and diagnosis of the enterprise's competitiveness in real time through common methods. Summary of the invention

[0005] The purpose of the present invention is to provide a multi-modal intelligent diagnosis modeling method, system and medium for enterprise competitiveness status, and solve the following technical problems: How to achieve accurate screening of enterprise multimodal data and real-time judgment of enterprise competitiveness through intelligent diagnostic modeling information extraction process.

[0006] The purpose of the present invention can be achieved by the following technical solutions: Multi-modal intelligent diagnostic modeling method for enterprise competitiveness, including: S1. Obtain historical enterprise multimodal operational data for cleaning, labeling and preprocessing; Multimodal operational data includes multidimensional operational text data and multidimensional operational numerical data; S2. Perform feature matching on the preprocessed multidimensional operation text data and multidimensional operation numerical data to obtain operation feature parameters; S3. Construct a convolutional neural network model, and input the set of operation characteristic parameters as samples into the convolutional neural network model for training, obtain a parameter optimization model, and output the optimization standard value; S4. According to the parameter optimization model, the actual operation characteristic parameters collected by the current enterprise are input into the parameter optimization model for verification, and the evaluation parameters are output; S5. Extract evaluation index coefficients according to the evaluation parameters, and determine the enterprise competitiveness results according to the evaluation index coefficients.

[0007] Preferably, the feature matching method in S2 is: S21, converting the pre-processed multi-dimensional operation text data into a word frequency vector set, weighting the word frequency vector set by TF-IDF, inputting the weighted word frequency vector set into a recurrent neural network model for training, and obtaining topic semantic information; S22. Collect topic semantic information, match and filter multi-dimensional operation numerical data, count the matching amount of multi-dimensional operation numerical data of different types of topics and perform linear analysis to determine whether the operation distribution meets the requirements: If yes, then output the operational characteristic parameters; If not, eliminate the current feature.

[0008] Preferably, the specific analysis method of the operation distribution is: By formula Calculate the operation distribution coefficient ;in, is the total number of subject types, and ∈ ; For the The matching amount of multi-dimensional operational numerical data of the subject matter; For the Multi-dimensional operational numerical data standard matching quantity of class subject; For the Preset deviation value of multi-dimensional operational numerical data matching quantity of class subject; For the Preset weight coefficients for class topics; is a preset function.

[0009] Preferably, the operation characteristic parameters are obtained in the following manner: Operational Distribution Factor and preset operation distribution coefficient threshold For comparison: like ≤ , then the multi-dimensional operation numerical data matching amount corresponding to the current topic type is output, and the multi-dimensional operation numerical data matching amount corresponding to each topic type is used as the component set Elements , filter the collection The average matching amount and the highest matching amount of all elements in the , and the average value of the sum of the average matching amount and the highest matching amount of all elements is recorded as the operation characteristic parameter.

[0010] Preferably, the construction process of the parameter optimization model in S3 is: S31, collecting a number of normalized operational characteristic parameters and semantic texts of their corresponding topic types to generate training samples; S32. Input the training samples into the convolutional neural network model for optimization training and update to obtain the parameter optimization model, and output the optimization standard value according to the regression task.

[0011] Preferably, the evaluation parameters in S4 include: By formula Get evaluation parameters ;in, is the optimization function; To optimize the influence coefficient; Optimize forecast values ​​for operational characteristic parameters; Optimize standard values ​​for operational characteristic parameters.

[0012] Preferably, in S5: By formula Calculate the evaluation index coefficient ; in, is the total number of evaluation indicators, and ∈ ; For the The evaluation parameters of the evaluation indicators; Parameters for standard evaluation; is the average value of historical evaluation parameters.

[0013] Preferably, it also includes: The evaluation index coefficient The threshold interval of the coefficient of the standard evaluation index To compare: like ∈ , then the evaluation indicators are normal and the enterprise competitiveness is judged to be normal; like < , then the evaluation index is low and the enterprise competitiveness is judged to be weak; like > , then the evaluation index is low and the enterprise competitiveness is judged to be strong.

[0014] Multi-modal intelligent diagnosis modeling system for enterprise competitiveness, including: Data collection module, used to obtain historical enterprise multimodal operational data for cleaning, labeling and preprocessing; A feature extraction module is used to perform feature matching on the pre-processed multi-dimensional operation text data and multi-dimensional operation numerical data to obtain operation feature parameters; The model training module is used to build a convolutional neural network model, and input the set of operation characteristic parameters as samples into the convolutional neural network model for training, obtain the parameter optimization model, and output the optimization standard value; The model evaluation module, based on the parameter optimization model, inputs the actual operation characteristic parameters collected by the current enterprise into the parameter optimization model for verification and outputs the evaluation parameters; The result evaluation module extracts the evaluation index coefficients according to the evaluation parameters, and judges the enterprise competitiveness results according to the evaluation index coefficients.

[0015] A computer-readable storage medium having a computer program, which, when executed by a processor, implements the steps of a multi-modal intelligent diagnosis modeling method for enterprise competitiveness status.

[0016] Beneficial effects of the present invention: The present invention obtains multidimensional operation text data and multidimensional operation numerical data by obtaining historical enterprise multimodal operation data, performs preprocessing and feature matching, obtains operation characteristic parameters, and realizes accurate screening of enterprise multimodal data. The set of operation characteristic parameters is input as a sample into the convolutional neural network model for training, a parameter optimization model is obtained, and the optimization standard value is output; according to the parameter optimization model, the operation characteristic parameters actually collected by the current enterprise are input into the parameter optimization model for verification, and the evaluation parameters are output; the evaluation index coefficients are extracted according to the evaluation parameters, and the intelligent diagnostic modeling information extraction process for real-time judgment of the enterprise's competitiveness status is realized according to the evaluation index coefficients, and then the competitiveness of the enterprise is judged according to the evaluation index coefficient results.

[0017] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0019] Figure 1 A step diagram of the multi-modal enterprise competitiveness intelligent diagnosis modeling method of the present invention; Figure 2 This is a step diagram of the feature matching method in step S2 of the present invention; Figure 3 This is a module diagram of the multi-modal enterprise competitiveness status intelligent diagnosis modeling system of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] See also Figure 1 As shown, the present invention is a multi-modal intelligent diagnosis modeling method for enterprise competitiveness, comprising: S1. Obtain historical enterprise multimodal operational data for cleaning, labeling and preprocessing; Multimodal operational data includes multidimensional operational text data and multidimensional operational numerical data; S2. Perform feature matching on the preprocessed multidimensional operation text data and multidimensional operation numerical data to obtain operation feature parameters; S3. Construct a convolutional neural network model, and input the set of operation characteristic parameters as samples into the convolutional neural network model for training, obtain a parameter optimization model, and output the optimization standard value; S4. According to the parameter optimization model, the actual operation characteristic parameters collected by the current enterprise are input into the parameter optimization model for verification, and the evaluation parameters are output; S5. Extract evaluation index coefficients according to the evaluation parameters, and determine the enterprise competitiveness results according to the evaluation index coefficients.

[0022] In the above technical solution, in this embodiment, a multi-modal intelligent diagnosis modeling method for enterprise competitiveness is designed to accurately judge the competitiveness of an enterprise, thereby realizing the intelligent diagnosis modeling information extraction process for accurate screening of enterprise multi-modal data and real-time judgment of enterprise competitiveness. The specific methods are: The first step is to collect multimodal operational data from historical enterprise big data databases for cleaning, labeling and preprocessing; and use artificial intelligence models to learn and reason about multimodal data (such as text, images, sounds, videos, etc.). Unlike traditional data, which is usually collected from a single source or mode (such as a database or spreadsheet), this type of data is usually structured and easier to analyze, but may not capture the full complexity of the situation; the multimodal data designed in this paper emphasizes the complementarity and integration of data in different modes, and realizes the connection and deepening of cross-modal data by integrating data of multiple modes and using technologies such as representation learning, modal fusion and alignment.

[0023] Among them, the multimodal operation data in this design includes multidimensional operation text data and multidimensional operation numerical data; there is a large amount of text and numerical information in the images, sounds and videos, which can be screened through artificial intelligence models.

[0024] In the second step, feature matching is then performed on the pre-processed multi-dimensional operational text data and multi-dimensional operational numerical data to obtain operational feature parameters; data integration is required during the feature matching process, and data is filtered through text field features. Finally, the filtered data is standardized to obtain operational feature parameters that meet the requirements.

[0025] For details, please refer to Figure 2 As shown, as an implementation of the present invention, the method of performing feature matching in step S2 is: S21, converting the pre-processed multi-dimensional operation text data into a word frequency vector set, weighting the word frequency vector set by TF-IDF, inputting the weighted word frequency vector set into a recurrent neural network model for training, and obtaining topic semantic information; S22. Collect topic semantic information, match and filter multi-dimensional operation numerical data, count the matching amount of multi-dimensional operation numerical data of different types of topics and perform linear analysis to determine whether the operation distribution meets the requirements: If yes, then output the operational characteristic parameters; If not, eliminate the current feature.

[0026] The above-mentioned feature matching method in step S2 is designed. First, the pre-processed multi-dimensional operation text data is converted into a word frequency vector set to realize the integration of information extraction from the text data, determine the word frequency of the screened text data, and weight the word frequency vector set through TF-IDF. The weighted word frequency vector set is input into the recurrent neural network model for training to obtain topic semantic information; the topic that conforms to the current operation status is obtained through the training of the word frequency vector of the recurrent neural network model, and its topic semantic information is collected as the output target information.

[0027] Secondly, the multi-dimensional operational numerical data is matched and screened for the subject semantic information. Compared with directly screening multi-dimensional operational text data and multi-dimensional operational numerical data, it lacks matching and connection, and a large number of matching analysis processes are required before screening. The steps required are relatively cumbersome, and text fields and features must be aligned as well as data screening and cleaning processes must be performed. In this design, the screened subject semantic information is matched and screened for the corresponding operational numerical data. After screening, the matching amount of multi-dimensional operational numerical data of different types of topics is statistically analyzed, and the matching amount of multi-dimensional operational numerical data after statistics is linearly analyzed to determine whether the operational distribution meets the requirements. If so, the operational feature parameters are output; if not, the current features are eliminated.

[0028] It should be noted that feature matching is an iterative process, and the data preprocessing steps may need to be adjusted back and forth based on the results of the preliminary analysis; considering timeliness, the latest data processing technologies and best practices should be used first.

[0029] The technical solution of the above steps S21-S22 realizes semantic analysis of text based on operation-related features, screens the values ​​of operation-related features that meet the semantics, and accumulates the matching quantity, thereby realizing the determination and analysis process of the key and main parameters of enterprise competitiveness and realizing the standardized screening process of operation features, thereby accurately narrowing the screening scope related to the main operations of enterprise competitiveness.

[0030] The third step is to build a convolutional neural network model, analyze the operational characteristic parameters and train the current model. Through task construction and inputting a set of operational characteristic parameters that meet the requirements as samples into the convolutional neural network model for training, a parameter optimization model is obtained through multiple training processes, and the optimization standard value is output to realize the construction of the parameter optimization model, ensuring that the output results are more in line with the optimization task requirements.

[0031] Specifically, as an implementation of the present invention, the construction process of the parameter optimization model in step S3 is: S31, collecting a number of normalized operational characteristic parameters and semantic texts of their corresponding topic types to generate training samples; S32. Input the training samples into the convolutional neural network model for optimization training and update to obtain the parameter optimization model, and output the optimization standard value according to the regression task.

[0032] The construction process of the parameter optimization model is realized through steps S31-S32, mainly by collecting a number of normalized operational feature parameters and the semantic texts of the corresponding subject types to generate training samples. Since different semantic texts represent different business operations, there are positive and negative ones, it is necessary to build positive and negative operation types in advance, determine whether the current operational feature parameters are positive or negative compared to the enterprise, assign prerequisites, and output the required results according to the regression task, that is, by inputting the training samples into the convolutional neural network model for optimization training, confirming the update according to the optimal parameters obtained by training, obtaining the trained parameter optimization model, and outputting the optimization standard value according to the regression task, to achieve standardized processing of the operational feature parameters, so as to facilitate the next step of verification and analysis.

[0033] The fourth step is to input the actual operating characteristic parameters collected by the current enterprise into the parameter optimization model for verification based on the parameter optimization model in the previous step, and output the evaluation parameters, thus realizing the accurate acquisition process of the evaluation parameters.

[0034] Specifically, as an implementation of the present invention, the acquisition and calculation process of the evaluation parameters in S4 is: By formula Get evaluation parameters ;in, is the optimization function; To optimize the influence coefficient; Optimize forecast values ​​for operational characteristic parameters; Optimize standard values ​​for operational characteristic parameters.

[0035] The above-mentioned calculation and acquisition of the evaluation parameters in step S4 realizes providing a detailed evaluation basis for the competitiveness of the enterprise, specifically through the formula Get evaluation parameters , reflect the overall situation of the entire enterprise characteristic theme evaluation according to the size of the evaluation parameters, and make further adjustments based on the results.

[0036] It needs further explanation that the optimization function It is the result of fitting historical data, which is used to adjust the current calculation results to ensure that the calculation and prediction results are within the reasonable evaluation range. It will not be described in detail here. It is obtained by simulation based on historical experience data and will not be described in detail here; Optimization influence coefficient It is set based on the degree of optimization of the impact of operating characteristic parameters on the overall competitiveness of the enterprise. It is acquired through machine simulation learning and will not be elaborated here.

[0037] The fifth step is to extract the evaluation index coefficient according to the evaluation parameters, and judge the enterprise competitiveness results according to the evaluation index coefficient. The evaluation index coefficient is calculated by the following formula to achieve accurate judgment of the strength of enterprise competitiveness.

[0038] Specifically, as an implementation of the present invention, in step S5, the enterprise competitiveness is judged by obtaining the evaluation index coefficient: By formula Calculate the evaluation index coefficient ; in, is the total number of evaluation indicators, and ∈ ; For the The evaluation parameters of the evaluation indicators; Parameters for standard evaluation; is the average value of historical evaluation parameters.

[0039] The evaluation index coefficients obtained by analyzing the cumulative changes of the evaluation parameters of different evaluation indicators are calculated by the formula Size, based on the comparison of size, reflects the competitiveness of the enterprise. Among them, the standard evaluation parameters It is obtained by simulation based on historical data and will not be described in detail here.

[0040] As an embodiment of the present invention, it also includes: The evaluation index coefficient The threshold interval of the coefficient of the standard evaluation index To compare: like ∈ , then the evaluation indicators are normal and the enterprise competitiveness is judged to be normal; like < , then the evaluation index is low and the enterprise competitiveness is judged to be weak; like > , then the evaluation index is low and the enterprise competitiveness is judged to be strong.

[0041] As an implementation mode of the present invention, a specific analysis method of the operation distribution situation is as follows: By formula Calculate the operation distribution coefficient ;in, is the total number of subject types, and ∈ ; For the The matching amount of multi-dimensional operational numerical data of the subject matter; For the Multi-dimensional operational numerical data standard matching quantity of class subject; For the Preset deviation value of multi-dimensional operational numerical data matching quantity of class subject; For the Preset weight coefficients for class topics; is a preset function.

[0042] In the above technical solution, the specific analysis method of the operation distribution in this embodiment includes obtaining the operation distribution coefficient, according to the formula Calculate the operation distribution coefficient , Operational Distribution Coefficient It is obtained by further calculation and analysis based on the matching amount of multi-dimensional operational numerical data of different subject types. , It is obtained by simulating historical data and will not be described in detail here; the preset function It is an adjustment function set according to historical data to ensure that the result of adjusting the operation distribution coefficient is within a specific and reasonable range.

[0043] As an implementation mode of the present invention, the operation characteristic parameter acquisition method is: Operational Distribution Factor and preset operation distribution coefficient threshold For comparison: like ≤ , then the multi-dimensional operation numerical data matching amount corresponding to the current topic type is output, and the multi-dimensional operation numerical data matching amount corresponding to each topic type is used as the component set Elements , filter the collection The average matching amount and the highest matching amount of all elements in the , and the average value of the sum of the average matching amount and the highest matching amount of all elements is recorded as the operation characteristic parameter.

[0044] In the above technical solution, the collection of the operation characteristic parameters in this embodiment is based on the operation distribution coefficient and preset operation distribution coefficient threshold For comparison, the camp distribution coefficient Less than or equal to the preset operation distribution coefficient threshold In the case of, output the multi-dimensional operation numerical data matching amount corresponding to the current theme type; then further process the multi-dimensional operation numerical data matching amount corresponding to the previous theme type: through the collection , filter the elements The average number of matches and the element with the highest number of matches in this collection The value of is used as the operation characteristic parameter.

[0045] One calculation method in this embodiment is to record the average matching amount of all elements and the average value of the sum of the highest matching amount as the operation characteristic parameter. Of course, in practice, according to the characteristics and processing of the enterprise operation parameters, the weight relationship calculation and analysis are carried out specifically.

[0046] As an embodiment of the present invention, please refer to Figure 3 As shown in the figure, a multi-modal intelligent diagnosis modeling system for enterprise competitiveness is also designed, including: Data collection module, used to obtain historical enterprise multimodal operational data for cleaning, labeling and preprocessing; A feature extraction module is used to perform feature matching on the pre-processed multi-dimensional operation text data and multi-dimensional operation numerical data to obtain operation feature parameters; The model training module is used to build a convolutional neural network model, and input the set of operation characteristic parameters as samples into the convolutional neural network model for training, obtain the parameter optimization model, and output the optimization standard value; The model evaluation module, based on the parameter optimization model, inputs the actual operation characteristic parameters collected by the current enterprise into the parameter optimization model for verification and outputs the evaluation parameters; The result evaluation module extracts the evaluation index coefficients according to the evaluation parameters, and judges the enterprise competitiveness results according to the evaluation index coefficients.

[0047] The above five modules, namely, data collection module, feature extraction module, model training module, model evaluation module and result evaluation module, realize a multimodal intelligent diagnosis method for enterprise competitiveness and provide basic modeling functions.

[0048] The present invention also provides a computer-readable storage medium having a computer program. The computer-readable storage medium may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device.

[0049] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by an electronic device, the electronic device implements the steps of the multi-modal enterprise competitiveness status intelligent diagnosis modeling method.

[0050] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, equipment, and non-volatile computer storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0051] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined in this application, they shall all fall within the protection scope of the present invention.

Claims

1. A multi-modal intelligent diagnosis modeling method for enterprise competitiveness, characterized by: The method comprises: S1. Obtain historical enterprise multimodal operational data for cleaning, labeling and preprocessing; The multimodal operation data includes multidimensional operation text data and multidimensional operation numerical data; S2. Perform feature matching on the preprocessed multidimensional operation text data and multidimensional operation numerical data to obtain operation feature parameters; The method of performing feature matching in S2 is: S21, converting the pre-processed multi-dimensional operation text data into a word frequency vector set, weighting the word frequency vector set by TF-IDF, inputting the weighted word frequency vector set into a recurrent neural network model for training, and obtaining topic semantic information; S22. Collect topic semantic information, match and filter multi-dimensional operation numerical data, count the matching amount of multi-dimensional operation numerical data of different types of topics and perform linear analysis to determine whether the operation distribution meets the requirements: If yes, then output the operational characteristic parameters; If not, eliminate the current feature; S3. Construct a convolutional neural network model, and input the set of operation characteristic parameters as samples into the convolutional neural network model for training, obtain a parameter optimization model, and output the optimization standard value; S4. According to the parameter optimization model, the actual operation characteristic parameters collected by the current enterprise are input into the parameter optimization model for verification, and the evaluation parameters are output; S5. Extract evaluation index coefficients according to the evaluation parameters, and determine the enterprise competitiveness results according to the evaluation index coefficients.

2. The multi-modal enterprise competitiveness intelligent diagnosis modeling method according to claim 1 is characterized in that: The specific analysis method of the operation distribution is as follows: By formula Calculate the operation distribution coefficient ;in, is the total number of subject types, and ∈ ; For the The matching amount of multi-dimensional operational numerical data of the subject matter; For the Multi-dimensional operational numerical data standard matching quantity of class subject; For the Preset deviation value of multi-dimensional operational numerical data matching quantity of class subject; For the Preset weight coefficients for class topics; is a preset function.

3. The multi-modal enterprise competitiveness intelligent diagnosis modeling method according to claim 2 is characterized in that: The method for obtaining the operation characteristic parameters is as follows: The operation distribution coefficient and preset operation distribution coefficient threshold For comparison: like ≤ , then the multi-dimensional operation numerical data matching amount corresponding to the current topic type is output, and the multi-dimensional operation numerical data matching amount corresponding to each topic type is used as the component set Elements , filter the collection The average matching amount and the highest matching amount of all elements in the , and the average value of the sum of the average matching amount and the highest matching amount of all elements is recorded as the operation characteristic parameter.

4. The multi-modal enterprise competitiveness intelligent diagnosis modeling method according to claim 1 is characterized in that: The construction process of the parameter optimization model in S3 is as follows: S31, collecting a number of normalized operational characteristic parameters and semantic texts of their corresponding topic types to generate training samples; S32. Input the training samples into the convolutional neural network model for optimization training and update to obtain the parameter optimization model, and output the optimization standard value according to the regression task.

5. The multi-modal enterprise competitiveness intelligent diagnosis modeling method according to claim 1 is characterized in that: The evaluation parameters in S4 include: By formula Get evaluation parameters ;in, is the optimization function; To optimize the influence coefficient; Optimize forecast values ​​for operational characteristic parameters; Optimize standard values ​​for operational characteristic parameters.

6. The multi-modal enterprise competitiveness intelligent diagnosis modeling method according to claim 5 is characterized in that: In S5: By formula Calculate the evaluation index coefficient ; in, is the total number of evaluation indicators, and ∈ ; For the The evaluation parameters of the evaluation indicators; Parameters for standard evaluation; is the average value of historical evaluation parameters.

7. The multi-modal enterprise competitiveness intelligent diagnosis modeling method according to claim 6 is characterized in that: Also includes: The evaluation index coefficient The threshold interval of the coefficient of the standard evaluation index To compare: like ∈ , then the evaluation indicators are normal and the enterprise competitiveness is judged to be normal; like < , then the evaluation index is low and the enterprise competitiveness is judged to be weak; like > , then the evaluation index is low and the enterprise competitiveness is judged to be strong.

8. A multi-modal enterprise competitiveness intelligent diagnosis modeling system, used to implement the multi-modal enterprise competitiveness intelligent diagnosis modeling method according to any one of claims 1 to 7, characterized in that: The system comprises: Data collection module, used to obtain historical enterprise multimodal operational data for cleaning, labeling and preprocessing; A feature extraction module is used to perform feature matching on the pre-processed multi-dimensional operation text data and multi-dimensional operation numerical data to obtain operation feature parameters; The model training module is used to build a convolutional neural network model, and input the set of operation characteristic parameters as samples into the convolutional neural network model for training, obtain the parameter optimization model, and output the optimization standard value; The model evaluation module, based on the parameter optimization model, inputs the actual operation characteristic parameters collected by the current enterprise into the parameter optimization model for verification and outputs the evaluation parameters; The result evaluation module extracts the evaluation index coefficients according to the evaluation parameters, and judges the enterprise competitiveness results according to the evaluation index coefficients.

9. A computer-readable storage medium, the computer-readable storage medium having a computer program, characterized in that: When the computer program is executed by a processor, the steps of the multi-modal enterprise competitiveness status intelligent diagnosis modeling method as described in any one of claims 1 to 7 are implemented.

Citation Information

Patent Citations

  • Enterprise risk analysis method and system based on big data, and medium

    CN116468271A

  • Data prediction model training method, data prediction model using method, data prediction model training device, data prediction model using device, equipment and medium

    CN117010542A

  • Business evaluation optimization method for digital catering and server

    CN117455583A

  • Enterprise management method and system based on big data

    CN118446415A

  • Systems and methods of multimodal clustering using machine learning

    US20230065870A1

Cited By

  • Adjusting and optimizing method based on input and output characteristic parameters and device applying method

    CN119758937A

  • Enterprise environment responsibility evaluation method, device, equipment, medium and program product

    CN120579715A