Digital economy dynamic analysis method and system based on multi-modal large model

Abstract

The application relates to a kind of digital economic dynamic analysis method and system based on multimodal big model, method includes: real-time acquisition of multimodal multi-source data;Multiple source data is substituted into analysis module, and economic activity characteristic set is updated;In response to the update of economic activity characteristic set, economic activity characteristic set is substituted into economic dynamic analysis module, and the dynamic update economic analysis template and industry knowledge graph corresponding to each business circle are obtained;In response to the industry analysis instruction of user terminal, based on the latest edition of economic analysis template and industry knowledge graph, generate estimated graphic-text report and feedback to user terminal.It realizes the automatic analysis and fusion of multi-modal data information, improves the processing efficiency of multi-modal data, realizes the dynamic update of economic analysis template and the automatic construction of knowledge graph, and timely feedback to the user real-time estimated graphic-text report, and then improves the accuracy and dynamic estimation ability of estimated graphic-text report, ensures the accuracy of digital economic dynamic analysis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to the technical field of digital economy dynamic analysis, and in particular to a digital economy dynamic analysis method and system based on a multi-modal large model. BACKGROUND

[0002] With the vigorous development of the digital economy, enterprises as market mainstays, their dynamic changes have become an important window for observing the development of the digital economy. However, different modal data information is scattered in various systems, forming "data islands", which makes it difficult to associate cross-field knowledge, and the data analysis technology in the prior art has the problems of business process solidification, low collaboration efficiency, and link connection relying on manual intervention, which is prone to information delay and decision deviation, directly affecting the timeliness of economic analysis.

[0003] Therefore, a digital economy dynamic analysis method and system based on a multi-modal large model are needed, which can analyze and integrate cross-modal data, and enable real-time dynamic updating of digital economy analysis and industry knowledge. SUMMARY

[0004] Therefore, the present application proposes a digital economy dynamic analysis method and system based on a multi-modal large model, which solves the technical problems of business process solidification, low collaboration efficiency, and link connection relying on manual intervention in the prior art, which is prone to information delay and decision deviation, directly affecting the timeliness of economic analysis.

[0005] According to a first aspect of the present application, a digital economy dynamic analysis method based on a multi-modal large model is provided, the method comprising the following steps:

[0006] real-time acquisition of multi-modal multi-source data, wherein the multi-source data at least includes satellite remote sensing images, enterprise financial reports, and social media texts;

[0007] substituting the multi-source data into an analysis module to update an economic activity feature set, wherein the economic activity feature set at least includes factory start-up features, commercial district flow density features, traffic features, service facility distribution features, and commercial district market features;

[0008] in response to the update of the economic activity feature set, substituting the economic activity feature set into an economic dynamic analysis module to obtain a dynamically updated economic analysis template and an industry knowledge graph corresponding to each commercial district;

[0009] in response to an industry analysis instruction of a user terminal, generating an estimated graphic-text report based on the latest version of the economic analysis template and the industry knowledge graph and feeding back to the user terminal.

[0010] According to an embodiment of the present invention, substituting the multi-source data into the parsing module to update the economic activity feature set includes:

[0011] Based on satellite remote sensing images and corporate financial reports, the characteristics of factory operation were obtained;

[0012] Based on satellite remote sensing images and social media text, the characteristics of pedestrian density, traffic, and distribution of service facilities in the business district were obtained.

[0013] Based on corporate financial reports and social media texts, the characteristics of the business district market were obtained;

[0014] The economic activity feature set is updated based on the factory operation characteristics, business district pedestrian density characteristics, traffic characteristics, service facility distribution characteristics, and business district market characteristics.

[0015] According to an embodiment of the present invention, substituting the economic activity feature set into the economic dynamic analysis module yields a dynamically updated economic analysis template and industry knowledge graph corresponding to each business district, including:

[0016] Based on the latest version of the economic activity feature set, the comprehensive scores of basic features for different industries are obtained;

[0017] Obtain the current industry guidance feature set, wherein the feature types of the industry guidance feature set include at least policy response features, technological breakthrough features, industry linkage features, and industrial chain transmission features;

[0018] Based on the comprehensive scores of basic characteristics of different industries and the current industry guidance feature set, the recommendation index corresponding to different industries in each business district is obtained;

[0019] Based on the recommendation index corresponding to different industries within each business district, the economic analysis template and industry knowledge graph corresponding to each business district are updated.

[0020] According to an embodiment of the present invention, based on the latest version of the economic activity feature set, the comprehensive scores of basic features for different industries are obtained, including:

[0021] The comprehensive score of basic characteristics for different industries is obtained using the following formula:

[0022]

[0023] in, Representing the A comprehensive score based on the fundamental characteristics of each industry. Representing the The first industry Standardized values ​​of each feature Representing the The first industry Dynamic weights of each feature The number of representative features, here ;

[0024] Among them, the The first industry The dynamic weights of each feature are obtained using the following formula:

[0025]

[0026] in, Represents PCA weights. Representing the The first industry The original feature weights of each feature. Representing the The first industry The industry-specific correlation parameters of each feature.

[0027] According to an embodiment of the present invention, obtaining the current industry guidance feature set includes:

[0028] Based on social media text, determine the policy response characteristic values ​​for each industry;

[0029] Obtain current technology data for each industry;

[0030] Based on the current technology data of each industry, determine the characteristic values ​​of technological breakthroughs in each industry;

[0031] Acquire data on supply chain connections between different industries;

[0032] Based on supply chain correlation data and corporate financial reports across different industries, the industry correlation characteristic value of each industry is determined.

[0033] Based on social media texts, corporate financial reports, and supply chain linkage data between different industries, the characteristic value of supply chain transmission for each industry is determined.

[0034] To obtain the current industry guidance feature set.

[0035] According to an embodiment of the present invention, based on the comprehensive scores of basic characteristics of different industries and the current industry guidance feature set, the recommendation index corresponding to different industries within each business district is obtained, including:

[0036] The recommendation index for different industries within each business district is obtained using the following formula:

[0037]

[0038] in, Representing the The first business district Recommendation index for each industry , , , , All represent the first Dynamic weighting coefficients that match each industry. Representing the Policy response characteristic values ​​of each industry Representing the Characteristic values ​​of technological breakthroughs in various industries Representing the Industry-related characteristic values ​​of each industry Representing the The characteristic value of supply chain transmission in each industry Representing the The first business district The comprehensive score of the basic characteristics of each industry.

[0039] According to an embodiment of the present invention, updating the economic analysis template and industry knowledge graph corresponding to each business district based on the recommendation index corresponding to different industries within each business district includes:

[0040] Based on the recommendation index corresponding to different industries within each business district, generate the corresponding prompt word engineering parameters for each business district;

[0041] Based on the recommendation index corresponding to different industries within each business district and the corresponding prompt word engineering parameters, update the economic analysis template corresponding to each business district.

[0042] Based on the recommendation index corresponding to different industries within each business district and the engineering parameters of the prompt words corresponding to each business district, the industry knowledge graph corresponding to each business district is updated.

[0043] According to an embodiment of the present invention, based on the recommendation index corresponding to different industries within each business district, the engineering parameters for generating prompt words corresponding to each business district include:

[0044] Based on the recommendation index corresponding to different industries within each business district, the multiple industries within each business district are sorted to obtain the ranking of multiple industries within each business district;

[0045] Based on the arrangement position of each industry within each business district, determine the number of prompt word parameters corresponding to each industry within each business district;

[0046] Based on the number of prompt word parameters corresponding to each industry within each business district, generate the corresponding prompt word engineering parameters for each business district.

[0047] According to an embodiment of the present invention, the recommendation index corresponding to different industries within each business district, based on the comprehensive score of basic characteristics of different industries and the current industry guidance feature set, further includes:

[0048] The number is obtained through the following formula. Policy response characteristics of individual industries :

[0049]

[0050] in, Representing the The policy strength parameter value corresponding to each policy. Representing the The industry and the first The degree of alignment between policy objectives This represents the value of the policy's timeliness decay parameter. These represent the weight values ​​corresponding to different preset policy levels;

[0051] The number is obtained through the following formula. Technological Breakthrough Characteristics of Individual Industries :

[0052]

[0053] in, Represents dynamic weighting coefficients. Represents the patent quality parameter value. Represents the R&D output parameter value;

[0054] The number is obtained through the following formula. The industry and the first Industry-related characteristic values ​​of individual industries :

[0055]

[0056] in, Representing the The industry and the first Covariance of returns for each industry Representing the The industry and the first The industrial chain linkage coefficient of each industry Representing the Variance of returns for each industry Representing the The profitability of individual industries Representing the Variance of returns for each industry Representing the The rate of return of each industry.

[0057] According to a second aspect of the present application, a digital economic dynamic analysis system based on a multi-modal large model is provided, the system comprising a communication device, a collection device, an analysis module, an economic dynamic analysis module and a control device, the communication device being configured to realize information interaction between a plurality of user terminals and the system, the collection device being configured to acquire multi-modal multi-source data in real time, acquire current industry guidance feature sets, acquire current technical data of each industry and acquire industry chain correlation data between different industries, the analysis module being configured to analyze the multi-source data to update economic activity feature sets, the economic dynamic analysis module being configured to obtain a dynamically updated economic analysis template corresponding to each business circle and an industry knowledge graph according to the economic activity feature sets, and the control device comprising a processor and a memory, the memory being adapted to store a plurality of program codes, the program codes being adapted to be loaded and run by the processor to execute any one of the technical solutions of the above-mentioned multi-modal large model-based digital economic dynamic analysis method.

[0058] As can be seen from the above technical solutions, the multi-modal large model-based digital economic dynamic analysis method provided by the present application has the following beneficial effects:

[0059] The multi-modal large model-based digital economic dynamic analysis method provided by the present application realizes automatic analysis and fusion of multi-modal data information by updating economic activity feature sets through multi-modal multi-source data acquired in real time into the analysis module, improves the processing efficiency of multi-modal data, and in response to the update of the economic activity feature sets, the economic activity feature sets are substituted into the economic dynamic analysis module to obtain a dynamically updated economic analysis template corresponding to each business circle and an industry knowledge graph, realizes dynamic update of the economic analysis template and automatic construction of the knowledge graph, ensures the timeliness of the economic analysis template and the industry knowledge graph, and in response to the industry analysis instruction of the user terminal, generates an estimated graphic-text report according to the latest economic analysis template and the industry knowledge graph and feeds back to the user terminal, realizes timely feedback of real-time estimated graphic-text report to the user, and further improves the accuracy and dynamic estimation ability of the estimated graphic-text report, ensures the accuracy of the digital economic dynamic analysis, avoids the technical problems of existing technologies, such as business process solidification, low coordination efficiency, link connection relying on manual intervention, easy to produce information delay and decision deviation, and directly affecting the timeliness of economic analysis. BRIEF DESCRIPTION OF DRAWINGS

[0060] With reference to the accompanying drawings, the disclosure of the present application will become more readily apparent. It will be readily understood to those skilled in the art that the drawings are for purposes of illustration only and are not intended to limit the scope of the present application. In addition, similar reference numerals are used to designate similar parts throughout the drawings.

[0061] Figure 1 is a main step flow diagram of a digital economy dynamic analysis method based on a multi-modal large model according to an embodiment of the present application;

[0062] Figure 2 is a main structure block diagram of a digital economy dynamic analysis system based on a multi-modal large model according to an embodiment of the present application.

[0063] List of reference signs:

[0064] 200: digital economy dynamic analysis system; 201: control device; 2011: processor; 2012: memory; 2013: program code; 202: analysis module; 203: acquisition device; 204: economic dynamic analysis module; 205: communication device. DETAILED DESCRIPTION

[0065] Some embodiments of the present application will be described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present application, and are not intended to limit the protection scope of the present application.

[0066] In the description of the present application, "module" and "processor" can include hardware, software or a combination of both. A module can include hardware circuit, various suitable sensors, communication port, memory, and can also include software part such as program code, and can be a combination of software and hardware. The processor can be a central processor, microprocessor, image processor, digital signal processor or any other suitable processor. The processor has data and / or signal processing functions. The processor can be implemented in software, hardware or a combination of both. The non-transitory computer readable storage medium includes any suitable medium that can store program code, such as magnetic disk, hard disk, optical disk, flash memory, read-only memory, random access memory, etc. The term "A and / or B" means all possible combinations of A and B, such as only A, only B or A and B. The term "at least one of A or B" or "at least one of A and B" has similar meaning as "A and / or B", which can include only A, only B or A and B. The singular form of the term "one", "this" can also include plural forms.

[0067] Some terms related to the present application will be explained first.

[0068] PCA, Principal Component Analysis, principal component analysis.

[0069] Referring to the accompanying drawings Figure 1 , Figure 1is a main step flow diagram of a digital economy dynamic analysis method based on a multi-modal large model according to an embodiment of the present application. As shown in Figure 1 , the digital economy dynamic analysis method based on a multi-modal large model in the embodiment of the present application mainly includes the following steps S101-S104.

[0070] Step S101: Real-time acquisition of multi-modal multi-source data, wherein the multi-source data at least includes satellite remote sensing images, enterprise financial reports, and social media texts;

[0071] Step S102: Substituting the multi-source data into the analysis module to update the economic activity feature set, wherein the economic activity feature set at least includes factory start-up features, commercial district passenger flow density features, traffic features, service facility distribution features, and commercial district market features;

[0072] Specifically, substituting the multi-source data into the analysis module to update the economic activity feature set includes:

[0073] Based on satellite remote sensing images and enterprise financial reports, the factory start-up features are obtained;

[0074] Based on satellite remote sensing images and social media texts, the commercial district passenger flow density features, traffic features, and service facility distribution features are obtained;

[0075] Based on enterprise financial reports and social media texts, the commercial district market features are obtained;

[0076] Based on the factory start-up features, commercial district passenger flow density features, traffic features, service facility distribution features, and commercial district market features, the economic activity feature set is updated.

[0077] Specifically, based on satellite remote sensing images and enterprise financial reports, the factory start-up features are obtained, including:

[0078] Based on satellite remote sensing images, the in-and-out vehicle parameters inside each factory are determined;

[0079] Based on enterprise financial reports, the capacity utilization rate of each factory is determined;

[0080] Based on the in-and-out vehicle parameters inside each factory and the capacity utilization rate of each factory, the factory start-up features are determined, wherein the factory start-up features = the in-and-out vehicle parameters inside the factory * a preset vehicle weight parameter corresponding to the factory + the capacity utilization rate of the factory * a preset capacity utilization rate parameter corresponding to the factory.

[0081] Specifically, based on satellite remote sensing images and social media texts, the commercial district passenger flow density features, traffic features, and service facility distribution features are obtained, including:

[0082] The commercial district passenger flow density features is obtained by the following formula:

[0083]

[0084] wherein, represents the social media check-in data parameter, represents the mobile signaling data parameter, represents the preset weight coefficient of the commercial district flow density feature, wherein, , represents the preset commercial district area parameter, represents the preset time window, represents the weather correction factor, represents the decay coefficient, wherein is 0.1;

[0085] Traffic feature is obtained by the following formula:

[0086]

[0087] wherein, represents the number of vehicles detected in the preset time period in the satellite image, represents the maximum capacity of the road section, represents the number of negative road condition feedbacks in the social media text, represents the total number of feedbacks in the social media text, represents the preset weight coefficient of the traffic feature, wherein ;

[0088] Service facility distribution feature is obtained by the following formula:

[0089]

[0090] wherein, represents the surrounding flow density parameter of the i-th facility, represents the check-in data in the social media text, represents the preset weight coefficient of the service facility distribution feature, wherein . Specifically, based on the enterprise financial report and the social media text, the commercial district market feature includes:

[0091] Commercial district market feature

[0092] is obtained by the following formula:

[0093] ​

[0094] in, This represents the gross profit margin parameter within the business district as stated in the company's financial report. This parameter represents the frequency of different keywords consumed in social media text. Represents the total number of keywords. The weighting coefficients representing the pre-defined market characteristics of the business district are shown here. 6.

[0095] Step S103: In response to the update of the economic activity feature set, the economic activity feature set is substituted into the economic dynamic analysis module to obtain the dynamically updated economic analysis template and industry knowledge graph corresponding to each business district.

[0096] Specifically, by substituting the economic activity feature set into the economic dynamic analysis module, the dynamically updated economic analysis template and industry knowledge graph corresponding to each business district are obtained, including:

[0097] Based on the latest version of the economic activity feature set, the comprehensive scores of basic features for different industries are obtained;

[0098] Obtain the current industry guidance feature set, wherein the feature types of the industry guidance feature set include at least policy response features, technological breakthrough features, industry linkage features, and industrial chain transmission features;

[0099] Based on the comprehensive scores of basic characteristics of different industries and the current industry guidance feature set, the recommendation index corresponding to different industries in each business district is obtained;

[0100] Based on the recommendation index corresponding to different industries within each business district, the economic analysis template and industry knowledge graph corresponding to each business district are updated.

[0101] Specifically, based on the latest version of the economic activity feature set, the comprehensive scores of basic features for different industries include:

[0102] The comprehensive score of basic characteristics for different industries is obtained using the following formula:

[0103]

[0104] in, Representing the A comprehensive score based on the fundamental characteristics of each industry. Representing the The first industry Standardized values ​​of each feature Representing the The first industry Dynamic weights of each feature The number of representative features, here ;

[0105] The dynamic weight of the first feature of the first industry is obtained by the following formula:

[0106]

[0107] wherein, represents the PCA weight,

[0108] Specifically, the original feature weight of the first feature of the first industry is obtained by the following formula:

[0109]

[0110] wherein, represents the total number of retained principal components,

[0111] Specifically, the industry feature correlation parameter of the first feature of the first industry is obtained by the following formula:

[0112]

[0113] wherein, represents the target variable of the first industry,

[0114] Specifically, the standardized value of the first feature of the first industry is obtained by the following formula:

[0115]

[0116] wherein,​​​​​​​​​​​​​​​​​​​​​​​​​​​​​ a raw value representing the first feature of the first industry, a raw value representing the first feature of the first industry, a raw value representing the first feature of the first industry, a minimum value representing the first feature of all industries, a minimum value representing the first feature of all industries, a maximum value representing the first feature of all industries, a maximum value representing the first feature of all industries.

[0117] Specifically, obtaining the current industry guidance feature set comprises:

[0118] determining a policy response feature value of each industry based on social media text;

[0119] obtaining current technical data of each industry;

[0120] determining a technology breakthrough feature value of each industry based on the current technical data of each industry;

[0121] obtaining industrial chain correlation data between different industries;

[0122] determining an industry correlation feature value of each industry based on the industrial chain correlation data between different industries and enterprise financial reports;

[0123] determining an industrial chain transmission feature value of each industry based on social media text, enterprise financial reports and industrial chain correlation data between different industries;

[0124] to obtain the current industry guidance feature set.

[0125] Specifically, based on the basic feature comprehensive score of different industries and the current industry guidance feature set, a recommended index corresponding to different industries in each business circle is obtained, comprising:

[0126] The recommended index corresponding to different industries in each business circle is obtained by the following formula:

[0127]

[0128] wherein, a recommended index corresponding to the first industry in the first business circle, 、 、 、 、 、 、 all represent dynamic weight coefficients matched with the first industry, a policy response feature value of the first industry, a technology breakthrough feature value of the first industry, a policy response feature value of the first industry, a technology breakthrough feature value of the first industry, a policy response feature value of the first industry, Representing the Industry-related characteristic values ​​of each industry Representing the The characteristic value of supply chain transmission in each industry Representing the The first business district The comprehensive score of the basic characteristics of each industry.

[0129] Specifically, based on the recommendation index corresponding to different industries within each business district, the economic analysis template and industry knowledge graph corresponding to each business district are updated, including:

[0130] Based on the recommendation index corresponding to different industries within each business district, generate the corresponding prompt word engineering parameters for each business district;

[0131] Based on the recommendation index corresponding to different industries within each business district and the corresponding prompt word engineering parameters, update the economic analysis template corresponding to each business district.

[0132] Based on the recommendation index corresponding to different industries within each business district and the engineering parameters of the prompt words corresponding to each business district, the industry knowledge graph corresponding to each business district is updated.

[0133] Specifically, based on the recommendation index corresponding to different industries within each business district, the engineering parameters for generating prompt words for each business district include:

[0134] Based on the recommendation index corresponding to different industries within each business district, the multiple industries within each business district are sorted to obtain the ranking of multiple industries within each business district;

[0135] Based on the arrangement position of each industry within each business district, determine the number of prompt word parameters corresponding to each industry within each business district;

[0136] Based on the number of prompt word parameters corresponding to each industry within each business district, generate the corresponding prompt word engineering parameters for each business district.

[0137] Specifically, the prompt word engineering parameters for each industry are generated. This can be done through existing prompt word engineering parameter generation methods or through a trained neural network model. The choice of the prompt word engineering parameter acquisition method here is only an example. In actual testing, those skilled in the art can choose according to actual needs, which will not be elaborated here.

[0138] Specifically, based on the comprehensive scores of basic characteristics of different industries and the current industry guidance feature set, the recommendation index corresponding to different industries within each business district also includes:

[0139] The number is obtained through the following formula. Policy response characteristics of individual industries :

[0140]

[0141] in, Representing the The policy strength parameter value corresponding to each policy. Representing the The industry and the first The degree of alignment between policy objectives This represents the value of the policy's timeliness decay parameter. These represent the weight values ​​corresponding to different preset policy levels;

[0142] The number is obtained through the following formula. Technological Breakthrough Characteristics of Individual Industries :

[0143]

[0144] in, Represents dynamic weighting coefficients. Represents the patent quality parameter value. Represents the R&D output parameter value;

[0145] The number is obtained through the following formula. The industry and the first Industry-related characteristic values ​​of individual industries :

[0146]

[0147] in, Representing the The industry and the first Covariance of returns for each industry Representing the The industry and the first The industrial chain linkage coefficient of each industry, among which... , Representing the The industry for the first Intermediate inputs in various industries Represents the total output value, Representing the Variance of returns for each industry Representing the The profitability of individual industries Representing the Variance of returns for each industry Representing the The rate of return of each industry.

[0148] Specifically, the industry chain transmission eigenvalue of the i-th industry is obtained by the following formula:

[0149]

[0150] represents the price transmission efficiency, wherein, , represents the price change of the i-th industry, represents the price change of the upstream industry of the i-th industry, represents the inventory influence fluctuation parameter value, represents the network center parameter value. Specifically, the patent quality parameter value is obtained by the following formula:

[0151]

[0152]

[0153] , , , represents the dynamic weight coefficient corresponding to each data volume, represents the number of citations, represents the number of international patent applications, represents the innovation parameter value.

[0154] Specifically, each of the above parameter values can be obtained by a trained machine learning model or a trained neural network model. The selection of the acquisition method of each parameter value herein is only exemplary and in actual testing, a person skilled in the art can select according to actual needs, which will not be repeated here.

[0155] Specifically, the economic dynamic analysis module also stores a real-time updated dynamic weight data table, which stores dynamic weight coefficients corresponding to each different data. The dynamic weight coefficients in the dynamic weight data table can be obtained by a trained machine learning model or a trained neural network model. The selection of the acquisition method of each dynamic weight coefficient herein is only exemplary and in actual testing, a person skilled in the art can select according to actual needs, which will not be repeated here.

[0156] Step S104: In response to the industry analysis instruction of the user terminal, a predicted graphic-text report is generated based on the latest version of the economic analysis template and the industry knowledge graph and is fed back to the user terminal.

[0157] ​​​​​Specifically, the generation of the prediction text report based on the latest economic analysis template and industry knowledge graph and the feedback to the user terminal include:

[0158] The latest economic analysis template and industry knowledge graph are substituted into the economic dynamic analysis module to generate a prediction text report and feedback to the user terminal.

[0159] Specifically, the latest economic analysis template and industry knowledge graph are substituted into the economic dynamic analysis module to generate a prediction text report and feedback to the user terminal, which includes:

[0160] Based on the latest economic analysis template and industry knowledge graph, a prediction parameter set for each business circle is obtained, wherein the prediction parameter set at least includes trend prediction parameters and risk prediction parameters.

[0161] Based on the prediction parameter set for each business circle, a prediction text report is generated.

[0162] Specifically, the prediction parameter set can be obtained by predicting the economic situation in each business circle using existing economic prediction technology, or by using a trained neural network model. The selection of the prediction parameter set acquisition method is only exemplary, and in actual testing, a person skilled in the art can select according to actual needs, and will not be repeated here.

[0163] Specifically, the industry knowledge graph and the generation method of the prediction text report can use the existing generation method, or can use a trained neural network model for generation. The selection of the generation method of the industry knowledge graph and the prediction text report is only exemplary, and in actual testing, a person skilled in the art can select according to actual needs, and will not be repeated here.

[0164] Specifically, the construction method of the economic analysis template can be obtained by training a machine learning model, or by training a neural network model. The selection of the construction method of the economic analysis template is only exemplary, and in actual testing, a person skilled in the art can select according to actual needs, as long as the economic analysis template is constructed according to the recommended index corresponding to each industry in each business circle and the prompt word engineering parameter corresponding to each business circle. This will not be repeated here.

[0165] Based on the steps S101-S104, the economic activity feature set is updated by substituting the real-time acquired multi-modal multi-source data into the analysis module, automatic analysis and fusion of multi-modal data information are realized, the processing efficiency of multi-modal data is improved, in response to the update of the economic activity feature set, the economic activity feature set is substituted into the economic dynamic analysis module, the dynamically updated economic analysis template and industry knowledge graph corresponding to each business circle are obtained, dynamic update of the economic analysis template and automatic construction of the knowledge graph are realized, the timeliness of the economic analysis template and the industry knowledge graph is ensured, in response to the industry analysis instruction of the user terminal, the latest economic analysis template and the industry knowledge graph are used to generate an estimated graphic-text report and feed back to the user terminal, real-time estimated graphic-text reports are fed back to the user in a timely manner, the accuracy and dynamic estimation ability of the estimated graphic-text report are improved, the accuracy of digital economic dynamic analysis is ensured, and the technical problems of existing technologies, such as business process solidification, low coordination efficiency, and link connection relying on manual intervention, are avoided, which can easily cause information delay and decision deviation, and directly affect the timeliness of economic analysis.

[0166] It should be noted that, although the steps in the above embodiments are described in a specific order, those skilled in the art can understand that, in order to achieve the effect of the present application, the different steps do not necessarily have to be executed in such an order, they can be executed simultaneously (in parallel) or in other orders, and these changes are within the protection scope of the present application.

[0167] Further, the present application also provides a digital economic dynamic analysis system based on a multi-modal large model.

[0168] Referring to the accompanying Figure 2 , Figure 2 is the main structure block diagram of a digital economic dynamic analysis system based on a multi-modal large model according to an embodiment of the present application. As Figure 2As shown, the digital economy dynamic analysis system 200 based on a multi-modal large model in the embodiment of the present application mainly comprises a communication device 205, a plurality of collection devices 203 of different types of collected data, an analysis module 202, an economic dynamic analysis module 204, and a control device 201. The communication device 205 is used to realize information interaction between a plurality of user terminals and the system. The collection device 203 is used to obtain multi-modal multi-source data in real time, obtain the current industry guidance feature set, obtain the current technical data of each industry, and obtain the industrial chain correlation data between different industries. The analysis module 202 is used to analyze multi-source data to update the economic activity feature set. The economic dynamic analysis module 204 is used to obtain the dynamically updated economic analysis template and the industry knowledge graph corresponding to each business circle according to the economic activity feature set. The control device 201 comprises a processor 2011 and a memory 2012. The memory 2012 can be configured to store program codes 2013 of the multi-modal large model-based digital economy dynamic analysis method of the above-mentioned method embodiments. The processor 2011 can be configured to execute the program codes 2013 in the memory 2012, which includes but is not limited to the program codes 2013 of the multi-modal large model-based digital economy dynamic analysis method of the above-mentioned method embodiments. For the convenience of description, only the parts related to the embodiments of the present application are shown, and the specific technical details not disclosed are referred to the method part of the embodiments of the present application. The control device 201 can be a control device device formed by various electronic devices.

[0169] Specifically, the collection device 203 can be various collection devices of different types, and can also be satellite remote sensing devices, enterprise financial report acquisition devices, and Internet acquisition devices for collecting social media texts. The selection of the collection device 203 here is only exemplary, and those skilled in the art can select according to actual needs in actual testing, as long as the collection device 203 can realize real-time acquisition of multi-modal multi-source data, acquisition of the current industry guidance feature set, acquisition of the current technical data of each industry, and acquisition of the industrial chain correlation data between different industries. Here, no further description is given.

[0170] Specifically, the communication device 205 can be connected through wifi, Bluetooth, or wired connection. The selection of the communication device 205 here is only exemplary, and those skilled in the art can select according to actual use needs, as long as the communication device 205 can realize mutual communication connection with a plurality of user terminals and the system, and further realize information interaction between the plurality of user terminals and the system. Here, no further description is given.

[0171] In one embodiment, the description of the specific implementation function can be referred to steps S101-S104.

[0172] The above-mentioned digital economy dynamic analysis system 200 based on a multi-modal large model is used to execute Figure 1 The technical principles, technical problems solved, and technical effects of the multi-modal large model-based digital economy dynamic analysis method embodiment shown are similar. For the convenience and brevity of description, the specific working process and related description of the multi-modal large model-based digital economy dynamic analysis system 200 can refer to the description of the multi-modal large model-based digital economy dynamic analysis method embodiment, which will not be described here.

[0173] Those skilled in the art can understand that all or part of the processes in the method of the above-mentioned embodiment can also be completed by a computer program instructing related hardware. The computer program can be stored in a computer readable storage medium. When the computer program is executed by the processor 2011, the steps of the above-mentioned various method embodiments can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable storage medium can include any entity or device, medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory, random access memory, electrical carrier signal, telecommunication signal, and software distribution medium, etc. that can carry the computer program code. It should be noted that the contents included in the computer readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer readable storage medium does not include electrical carrier signals and telecommunication signals.

[0174] Further, the multi-modal large model-based digital economy dynamic analysis system 200 of the present application further includes a computer readable storage medium. In one computer readable storage medium embodiment according to the present application, the computer readable storage medium can be configured to store program code 2013 for executing the multi-modal large model-based digital economy dynamic analysis method of the above-mentioned method embodiment. The program code 2013 can be loaded and run by the processor 2011 to implement the above-mentioned multi-modal large model-based digital economy dynamic analysis method. For the convenience of description, only the part related to the embodiment of the present application is shown, and the specific technical details that are not disclosed are referred to the method part of the embodiment of the present application. The computer readable storage medium can be a memory device formed by various electronic devices. Optionally, the computer readable storage medium in the embodiment of the present application is a non-transitory computer readable storage medium.

[0175] Further, it should be understood that, since the setting of each module is only for illustrating the functional units of the device of the present application, the corresponding physical device of the module can be the processor 2011 itself, or a part of the software in the processor 2011, a part of the hardware, or a part of the combination of the software and the hardware. Therefore, the number of each module in the figure is only illustrative.

[0176] Those skilled in the art can understand that each module in the device can be adaptively split or combined. Such splitting or combining of the specific module does not cause the technical solution to deviate from the principles of the present application, and therefore, the technical solution after splitting or combining will fall within the protection scope of the present application.

[0177] So far, the technical solution of the present application has been described in combination with the preferred embodiments shown in the drawings, but those skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without deviating from the principles of the present application, and the technical solution after the changes or replacements will fall within the protection scope of the present application.

Claims

1. A method for digital economy dynamic analysis based on a multi-modal large model, characterized in that, The method comprises the following steps: Real-time acquisition of multi-modal multi-source data, wherein the multi-source data at least comprises satellite remote sensing images, enterprise financial reports, and social media texts; Substituting the multi-source data into an analysis module to update an economic activity feature set, wherein the economic activity feature set at least comprises factory start-up features, commercial district passenger flow density features, traffic features, service facility distribution features, and commercial district market features; In response to the update of the economic activity feature set, substituting the economic activity feature set into an economic dynamic analysis module to obtain a dynamically updated economic analysis template and an industry knowledge graph corresponding to each commercial district; In response to an industry analysis instruction of a user terminal, generating an estimated graphic-text report based on the latest version of the economic analysis template and the industry knowledge graph and feeding back to the user terminal; The step of substituting the economic activity feature set into the economic dynamic analysis module to obtain the dynamically updated economic analysis template and the industry knowledge graph corresponding to each commercial district comprises: Obtaining a basic feature comprehensive score of different industries based on the latest version of the economic activity feature set; Obtaining a current industry guidance feature set, wherein the feature types of the industry guidance feature set at least comprise policy response features, technology breakthrough features, industry correlation features, and industry chain transmission features; Obtaining a recommendation index corresponding to different industries in each commercial district based on the basic feature comprehensive score of different industries and the current industry guidance feature set; Updating the economic analysis template and the industry knowledge graph corresponding to each commercial district based on the recommendation index corresponding to different industries in each commercial district.

2. The digital economy dynamic analysis method based on a multi-modal large model according to claim 1, characterized in that, The step of substituting the multi-source data into the analysis module to update the economic activity feature set comprises: Obtaining factory start-up features based on satellite remote sensing images and enterprise financial reports; Obtaining commercial district passenger flow density features, traffic features, and service facility distribution features based on satellite remote sensing images and social media texts; Obtaining commercial district market features based on enterprise financial reports and social media texts; Updating the economic activity feature set based on the factory start-up features, the commercial district passenger flow density features, the traffic features, the service facility distribution features, and the commercial district market features.

3. The digital economy dynamic analysis method based on a multi-modal large model according to claim 2, characterized in that, The step of obtaining a basic feature comprehensive score of different industries based on the latest version of the economic activity feature set comprises: The basic feature comprehensive score of different industries is obtained by the following formula: wherein, represents a base feature composite score for the industry, represents a standardized value for the feature for the industry, represents a dynamic weight for the feature for the industry, represents the number of features, here ; wherein the dynamic weight of the i-th feature of the j-th industry is obtained by the following formula: wherein the dynamic weight of the i-th feature of the j-th industry is obtained by the following formula:​ wherein, represents a PCA weight, represents a raw feature weight of a first feature of a first industry, represents a raw feature weight of a first feature of a first industry, represents a raw feature weight of a first feature of a first industry, represents an industry feature correlation parameter of a first feature of a first industry, represents an industry feature correlation parameter of a first feature of a first industry, represents an industry feature correlation parameter of a first feature of a first industry.

4. The digital economy dynamic analysis method based on a multi-modal large model according to claim 3, characterized in that, The step of obtaining a current industry guidance feature set comprises: Determining a policy response feature value of each industry based on social media texts; Obtaining current technical data of each industry; Determining a technology breakthrough feature value of each industry based on the current technical data of each industry; Obtaining industry chain correlation data between different industries; Determining an industry correlation feature value of each industry based on the industry chain correlation data between different industries and enterprise financial reports; Determining an industry chain transmission feature value of each industry based on social media texts, enterprise financial reports, and the industry chain correlation data between different industries; To obtain the current industry guidance feature set.

5. The digital economy dynamic analysis method based on a multi-modal large model according to claim 4, characterized in that, The step of obtaining a recommendation index corresponding to different industries in each commercial district based on the basic feature comprehensive score of different industries and the current industry guidance feature set comprises: The recommendation index corresponding to different industries in each commercial district is obtained by the following formula: wherein, represent the corresponding recommended index of the first industry in the first commercial circle, , , , , all represent the dynamic weight coefficient matched with the first industry, represent the policy response characteristic value of the first industry, represent the technology breakthrough characteristic value of the first industry, represent the industry correlation characteristic value of the first industry, represent the industry chain conduction characteristic value of the first industry, represent the basic characteristic comprehensive score of the first industry in the first commercial circle.

6. The digital economy dynamic analysis method based on a multi-modal large model according to claim 5, characterized in that, The updating of the economic analysis template and the industry knowledge graph corresponding to each business circle based on the recommendation index corresponding to different industries in each business circle comprises: Generating the prompt word engineering parameter corresponding to each business circle based on the recommendation index corresponding to different industries in each business circle; Updating the economic analysis template corresponding to each business circle based on the recommendation index corresponding to different industries in each business circle and the prompt word engineering parameter corresponding to each business circle; Updating the industry knowledge graph corresponding to each business circle based on the recommendation index corresponding to different industries in each business circle and the prompt word engineering parameter corresponding to each business circle.

7. The digital economy dynamic analysis method based on a multi-modal large model according to claim 6, characterized in that, The generating of the prompt word engineering parameter corresponding to each business circle based on the recommendation index corresponding to different industries in each business circle comprises: Ranking the multiple industries in each business circle based on the recommendation index corresponding to different industries in each business circle to obtain the arrangement of the multiple industries in each business circle; Determining the number of prompt word parameters corresponding to each industry in each business circle based on the arrangement position of each industry in each business circle; Generating the prompt word engineering parameter corresponding to each business circle based on the number of prompt word parameters corresponding to each industry in each business circle.

8. The digital economy dynamic analysis method based on a multi-modal large model according to claim 7, characterized in that, The obtaining of the recommendation index corresponding to different industries in each business circle based on the comprehensive score of the basic features of different industries and the current industry guidance feature set further comprises: The number is obtained through the following formula. Policy response characteristics of individual industries : wherein, represents a policy strength parameter value corresponding to the first policy, represents a policy strength parameter value corresponding to the first policy, represents a matching degree between the first industry and the first policy target, represents a matching degree between the first industry and the first policy target, represents a policy timeliness decay parameter value, represents a policy timeliness decay parameter value, represents a weight value corresponding to a preset different policy level; The number is obtained through the following formula. Technological Breakthrough Characteristics of Individual Industries : wherein, represents a dynamic weight coefficient, represents a patent quality parameter value, represents a research output parameter value; The industry correlation eigenvalue of the first industry and the second industry is obtained by the following formula: Industry correlation eigenvalue of the first industry and the second industry :​ ,in, Representing the The industry and the first Covariance of returns for each industry Representing the The industry and the first The industrial chain linkage coefficient of each industry Representing the Variance of returns for each industry Representing the The profitability of individual industries Representing the Variance of returns for each industry Representing the The rate of return of each industry. 9.A digital economy dynamic analysis system based on a multi-modal large model, characterized in that, The system comprises a communication device, a collection device, an analysis module, an economic dynamic analysis module, and a control device. The communication device is used to realize information interaction between a plurality of user terminals and the system. The collection device is used to obtain multi-modal multi-source data in real time, obtain the current industry guidance feature set, obtain the current technical data of each industry, and obtain the industry chain association data between different industries. The analysis module is used to analyze the multi-source data to update the economic activity feature set. The economic dynamic analysis module is used to obtain the dynamically updated economic analysis template and the industry knowledge graph corresponding to each business circle based on the economic activity feature set. The control device comprises a processor and a memory. The memory is adapted to store a plurality of program codes. The program codes are adapted to be loaded and run by the processor to execute the digital economic dynamic analysis method based on the multi-modal large model according to any one of claims 1 to 8.

Images