An innovation chain and industrial chain collaborative diagnosis method and system fusing multi-source data

Abstract

This invention discloses a method and system for collaborative diagnosis of innovation chains and industrial chains that integrates multi-source data, belonging to the fields of regional economic evaluation and artificial intelligence data processing technology. It aims to address the problem that existing evaluation methods rely on single structured data and lack in-depth decision-making capabilities. The method includes: collecting structured indicators of scientific research and industry in the target region and unstructured data of policy texts; quantifying the policy synergy effectiveness index using natural language processing technology; introducing market mechanism parameters to construct a nonlinear dual-track correction model to dynamically correct the original data and calculate the true degree of fusion and coordination; constructing structured prompt words based on the corrected multi-dimensional features to drive a pre-trained large language model to automatically generate a collaborative weakness diagnosis report and decision optimization suggestions. This invention improves the comprehensiveness and accuracy of evaluation, reduces the impact of pseudo-fusion, and achieves integrated intelligent diagnosis of measurement, attribution, and suggestions.

Description

Technical Field

[0001] This invention relates to the fields of data processing, artificial intelligence and industrial economics, and more specifically, to a method and system for collaborative diagnosis of innovation chains and industrial chains that integrates multi-source data. Background Technology

[0002] The deep integration of the innovation chain and the industrial chain is a crucial support for driving high-quality regional economic development. Regarding the measurement and evaluation of the integration level of these two chains, existing technologies generally adopt a quantitative evaluation approach driven by statistical indicators: structured indicator data (such as R&D expenditure, patent applications, number of high-tech enterprises, and industrial added value) are collected within a preset time scale. These indicators are then normalized, linearly weighted, and combined with comprehensive models such as coupling coordination degree to calculate an integration score or level for regional comparison, performance evaluation, and policy assessment. In recent years, some technical solutions have also begun to introduce machine learning or deep learning models to learn and model the features of parameters from multiple stakeholders, including governments, enterprises, and universities, to monitor, predict, or optimize collaborative efficiency. This has formed a technical route of "parameter collection—feature extraction—output of evaluation / optimization results."

[0003] Existing technologies still have the following significant shortcomings in the diagnostic of collaboration between the innovation chain and the industrial chain: First, existing evaluation systems rely heavily on annual or periodic statistical data, with low data update frequency and time lag, making it difficult to reflect short-term changes in policy adjustments, industrial fluctuations, and innovation activities. Second, unstructured information such as policy documents, planning texts, and implementation details contain key guidance on industrial direction, tool combinations, and constraints, but traditional statistical evaluation methods lack the ability to computably process unstructured texts, making it difficult to extract quantifiable and comparable effectiveness information from policy texts and incorporate it into the integration degree measurement framework, resulting in difficulty in characterizing the "policy guidance effect." Third, linear weighting and single-coupled models are easily influenced by scale-based indicators, leading to inflated results, and are unable to effectively identify "apparent integration" or "pseudo-integration" caused by structural mismatches between scientific research supply and industrial demand, thus causing the evaluation results to deviate from the true state of collaboration. Fourth, most evaluation systems still output scores or grades, lacking automated attribution analysis of the causes of low integration and textual decision-making suggestions for governance. Management departments often find it difficult to further determine whether the main reasons for low scores are due to over-supplied scientific research, insufficient industrial capacity, or supply-demand mismatch, thus affecting the accuracy of policy formulation and resource allocation.

[0004] Chinese patent CN117391646A discloses a collaborative innovation management system, which uses deep neural networks to extract parameter features from government, enterprises, and universities to optimize collaborative innovation efficiency. While this technology involves parameter modeling of multiple stakeholders in the innovation and industrial chains, its core logic primarily focuses on the digitization of management processes and parameter monitoring. Although it employs deep neural network modeling, the model lacks interpretability in its calculation process and cannot understand unstructured science and technology policy texts. This results in its inability to calculate "policy guidance effectiveness" and difficulty in identifying situations where policy implementation and evaluation results are disconnected.

[0005] Chinese patent CN116757885A discloses a corporate intellectual property dimension review system, proposing an evaluation method based on indicators such as "proportion of patents in strategic emerging industries" and "patent maintenance period." However, this method remains at a superficial frequency statistics stage, neglecting the structural matching problem between scientific research supply and industrial demand. More importantly, existing review systems of this kind typically only output scoring results and rankings, lacking explanations of causes and decision-making suggestions, and lacking intelligent attribution capabilities based on large language models. When faced with low scores, decision-makers cannot determine whether it is due to "over-capacity scientific research supply" or "insufficient industrial carrying capacity," and cannot directly obtain text-based decision-making suggestions generated by AI with logical reasoning capabilities.

[0006] Therefore, there is an urgent need to design an innovation chain and industrial chain collaborative diagnosis method and system that can deeply integrate unstructured policy texts and structured indicator data, possess a nonlinear dual-track correction mechanism, and utilize generative artificial intelligence to achieve automated diagnosis.

[0007] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0008] The purpose of this invention is to provide a method and system for collaborative diagnosis of innovation chain and industrial chain by integrating multi-source data. By constructing a three-level data fusion architecture and combining it with an improved artificial intelligence diagnostic model, the invention can achieve accurate assessment of the collaborative status of innovation chain and industrial chain, bottleneck location and optimization suggestions, thereby overcoming the defects in the prior art.

[0009] To achieve the above objectives, this invention provides a collaborative diagnostic method for innovation chains and industrial chains that integrates multi-source data, comprising the following steps:

[0010] Step S01: Multi-source data acquisition;

[0011] The system collects structured index data on the research supply side and structured index data on the industry demand side of the target region within a preset time window, as well as unstructured science and technology policy text data related to the target region. The structured index data on the research supply side and the structured index data on the industry demand side are normalized and then weighted and summed according to preset weights to obtain the original research supply score. Demand Score for Primary Industries ;

[0012] Step S02: Data standardization processing;

[0013] Natural language processing technology is used to parse, extract features, and perform quantitative analysis on the unstructured science and technology policy text data. Based on the policy coordination effectiveness scoring model, the policy coordination effectiveness index UOS within the preset time window is calculated.

[0014] Step S03: Data fusion processing;

[0015] Introducing market mechanism parameter S m Using the policy synergy effectiveness index UOS obtained in step S02 as a correction factor, a nonlinear dual-track correction model is constructed; the original scientific research supply score is then adjusted using the nonlinear dual-track correction model. and the score of demand for primary industries Dynamic adjustments are made to obtain the corrected true research supply score S. final Score I of actual industry demand final Then, based on the original research supply score... and the score of demand for primary industries The original fusion coordination degree D was calculated. raw According to the actual scientific research supply score S final Score I of actual industry demand final The corrected fusion coordination degree D was calculated. final ;

[0016] Step S04: Construct structured prompts;

[0017] Extracting the original fusion coordination degree D raw With the corrected fusion coordination degree D final The differences in characteristics, and the actual scientific research supply score S final Score I of actual industry demand final Based on the comparative characteristics and combined with the UOS policy coordination effectiveness index, a structured cue word containing role setting, data context and diagnostic objectives is constructed.

[0018] Step S05: Output the analysis results and decision recommendations in text format;

[0019] The structured prompts are input into the pre-trained DeepSeek-V3 large language model, which then outputs the analysis results and decision-making suggestions for the shortcomings of innovation chain and industrial chain collaboration in the target region, and generates a visual diagnostic report.

[0020] Preferably, in the technical solution, in step S01, the unstructured science and technology policy text data includes the full text or parsable text of science and technology plans, implementation rules, regulations, and guiding opinions.

[0021] Preferably, in the technical solution, in step S02, text extraction and cleaning are performed on each unstructured science and technology policy text to obtain unstructured science and technology policy text content for analysis; the keyword density score of each unstructured science and technology policy text is calculated based on a preset innovation chain and industry chain keyword library; when there is no available unstructured science and technology policy text within the preset time window, a preset default benchmark value is assigned; the average keyword density score of all valid unstructured science and technology policy texts within the preset time window is calculated to obtain the policy synergy effectiveness index UOS.

[0022] Preferably, in the technical solution, in step S3, the actual scientific research supply score S final The calculation formula is:

[0023] S final =S raw ×(1+α×UOS),

[0024] Real Industry Demand Score I final The calculation formula is:

[0025] I final =I raw ×(1+β×S m ),

[0026] in, This is the preset adjustment weight coefficient.

[0027] Preferably, in the technical solution, in step S03, the market mechanism parameter S m The calculation process is as follows: Construct a set of research supply texts and a set of industry demand texts; use a pre-trained semantic representation model based on deep neural networks to map the research supply texts and industry demand texts into high-dimensional semantic vectors respectively, and calculate the cosine similarity matrix between the supply semantic vectors and the demand semantic vectors; aggregate the cosine similarity matrix to obtain the mean or weighted mean of the supply-demand semantic fit, and use the mean or weighted mean of the supply-demand semantic fit as the market mechanism parameter S. m .

[0028] Preferably, in the technical solution, the process of constructing structured prompt words in step S04 is as follows: The task role of the DeepSeek-V3 large language model is preset as a regional industrial economic analysis expert in the prompt words; the target area identifier, time window identifier, and original fusion coordination degree D are then included. raw With the corrected fusion coordination degree D final The difference, and the real scientific research supply score S final Score I of actual industry demand final The difference is used as the data context to fill in the prompt word template; the prompt words limit the output to include the integration weakness attribution type and corresponding suggestions; the integration weakness attribution type includes at least scientific research advanced type, industry lagging type and supply and demand mismatch type.

[0029] Preferably, in the technical solution, in step S05, the visual diagnostic report includes a fusion degree time-series evolution diagram, a comparison diagram before and after correction, and a supply-demand matching bottleneck diagram; the fusion degree time-series evolution diagram plots the corrected fusion coordination degree D with time as the horizontal axis. final The change curve is displayed, and the fusion level passing line is marked; a comparison chart before and after correction is drawn, and the original fusion degree D is plotted in the same coordinate system. raw With the corrected fusion coordination degree D final The comparison curves are used, and the difference area between the two curves is filled to show the degree of supply-demand mismatch; the supply-demand matching bottleneck chart uses a bar chart to compare the actual scientific research supply score S at each time point. final Score I of actual industry demand final It automatically labels whether scientific research is ahead of schedule or industry is lagging behind based on the difference.

[0030] An innovation chain and industrial chain collaborative diagnostic system that integrates multi-source data includes a multi-source data acquisition module, a policy index calculation module, a dual-track correction calculation module, a prompt word construction module, a model inference calling module, and a visualization report generation module;

[0031] The multi-source data acquisition module is used to acquire structured indicator data from the scientific research supply side, structured indicator data from the industry demand side, and unstructured science and technology policy text data.

[0032] The policy index calculation module is used to parse and extract features from unstructured science and technology policy text data, and to calculate the policy synergy effectiveness index UOS.

[0033] The dual-track correction calculation module is used to receive the market mechanism parameter S. m It also incorporates the policy synergy effectiveness index UOS to perform nonlinear dual-track correction, outputting the true scientific research supply score S. final Real Industry Demand Score I final and the original fusion coordination degree D rawAnd the corrected fusion coordination degree D final ;

[0034] The prompt word construction module is used to construct prompt words based on D. raw D final S final I final and UOS build structured prompts;

[0035] The model inference calling module is used to input structured prompt words into the pre-trained large language model DeepSeek-V3 and obtain the collaborative weakness analysis results and decision suggestion text;

[0036] The visualization report generation module is used to generate reports based on D. raw D final S final I final Generate a diagnostic report that includes a time-series evolution diagram of fusion degree, a comparison diagram before and after correction, and a supply-demand matching shortcoming diagram, and integrate decision-making suggestion text output.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] By integrating structured indicator data on scientific research supply and industry demand with unstructured science and technology policy text data, natural language processing is used to quantify the policy synergy effectiveness index UOS, and market mechanism parameter S is introduced. m A nonlinear dual-track correction model is constructed to dynamically correct the original supply and demand scores, reducing the inflated integration degree bias caused by scale-based indicators and obtaining a more objective integration coordination degree result. Further, structured prompt words are constructed by extracting correction differences and supply-demand comparison features to drive the large language model to output integration shortcomings attribution and decision-making suggestions. It can also generate time-series, comparative, and shortcoming-based visual diagnostic reports, improving decision support efficiency. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the innovation chain and industrial chain collaborative diagnosis method that integrates multi-source data according to the present invention;

[0040] Figure 2 This is a structural block diagram of the innovation chain and industrial chain collaborative diagnostic system that integrates multi-source data according to the present invention. Detailed Implementation

[0041] The specific embodiments of the present invention will be described in detail below, but it should be understood that the scope of protection of the present invention is not limited to the specific embodiments.

[0042] Unless otherwise expressly stated, throughout the specification and claims, the term "comprising" or its variations such as "including" or "comprises" shall be understood to include the stated elements or components without excluding other elements or other components.

[0043] like Figure 1 As shown, a collaborative diagnostic method for innovation chains and industrial chains that integrates multi-source data includes the following steps:

[0044] Step S01: Multi-source data acquisition;

[0045] The system collects structured indicators of scientific research supply and industry demand within a preset time window for the target region, while also collecting unstructured science and technology policy text data related to the target region. The structured indicators of scientific research supply include indicators such as R&D investment, researchers, and knowledge output; the structured indicators of industry demand include indicators such as the number of high-tech enterprises, industry output, profits, employment, and new product sales; and the unstructured science and technology policy text data includes the full text or parsable text of science and technology plans, implementation rules, regulations, and guiding opinions. The structured indicators of scientific research supply and industry demand are normalized to eliminate dimensional differences, and then weighted and summed according to preset weights to obtain the original scientific research supply score. Demand Score for Primary Industries The normalization method can be range normalization, standardization, or other methods commonly used in this field; the preset weights can be determined by entropy weighting, analytic hierarchy process, or manual configuration.

[0046] Step S02: Data standardization processing;

[0047] The unstructured science and technology policy text data is parsed, features are extracted, and quantitative analysis is performed using natural language processing (NLP) technology. Based on the policy synergy effectiveness scoring model, the policy synergy effectiveness index (UOS) within the preset time window is calculated. The NLP technology includes word segmentation, entity recognition, keyword extraction, vector representation, topic analysis, text classification, or a combination thereof.

[0048] For each unstructured science and technology policy text, text extraction and cleaning are performed to obtain the unstructured science and technology policy text content for analysis; the keyword density score of each unstructured science and technology policy text is calculated based on a preset innovation chain and industrial chain keyword library; when there is no available unstructured science and technology policy text within the preset time window, a preset default benchmark value is assigned; the average keyword density score of all valid unstructured science and technology policy texts within the preset time window is calculated to obtain the policy synergy effectiveness index UOS;

[0049] Step S03: Data fusion processing;

[0050] Introducing market mechanism parameter S m Using the policy synergy effectiveness index UOS obtained in step S02 as a correction factor, a nonlinear dual-track correction model is constructed; the original scientific research supply score is then adjusted using the nonlinear dual-track correction model. and the score of demand for primary industries Dynamic adjustments are made to obtain the corrected true research supply score S. final Score I of actual industry demand final Then, based on the original research supply score... and the score of demand for primary industries The original fusion coordination degree D was calculated. raw According to the actual scientific research supply score S final Score I of actual industry demand final The corrected fusion coordination degree D was calculated. final ;

[0051] Real Research Supply Score final The calculation formula is:

[0052] S final =S raw ×(1+α×UOS),

[0053] Real Industry Demand Score I final The calculation formula is:

[0054] I final =I raw ×(1+β×S m ),

[0055] in, The preset adjustment weight coefficient;

[0056] Market mechanism parameter S m The calculation process is as follows: Construct a set of research supply texts and a set of industry demand texts; use a pre-trained semantic representation model based on deep neural networks (such as paraphrase-multilingual-MiniLM-L12-v2) to map the research supply texts and industry demand texts into high-dimensional semantic vectors respectively, and calculate the cosine similarity matrix between the supply semantic vector and the demand semantic vector; aggregate the cosine similarity matrix to obtain the mean or weighted mean of the supply-demand semantic fit, and use the mean or weighted mean of the supply-demand semantic fit as the market mechanism parameter S. m ;

[0057] Step S04: Construct structured prompts;

[0058] Extracting the original fusion coordination degree Draw With the corrected fusion coordination degree D final The differences in characteristics, and the actual scientific research supply score S final Score I of actual industry demand final Based on the comparative characteristics and combined with the UOS policy coordination effectiveness index, a structured cue word containing role setting, data context and diagnostic objectives is constructed.

[0059] The process of constructing structured prompts is as follows: The task role of the DeepSeek-V3 large language model is preset as a regional industrial economic analysis expert within the prompts; the target region identifier, time window identifier, and original fusion coordination degree D are then incorporated. raw With the corrected fusion coordination degree D final The difference, and the real scientific research supply score S final Score I of actual industry demand final The difference is used as the data context to fill in the prompt word template; the prompt words limit the output to include the integration weakness attribution type and corresponding suggestions; the integration weakness attribution type includes at least scientific research advanced type, industry lagging type and supply and demand mismatch type;

[0060] Step S05: Output the analysis results and decision recommendations in text format;

[0061] The structured prompts are input into the pre-trained large language model DeepSeek-V3, which then outputs the analysis results and decision-making suggestions for the innovation chain and industrial chain synergy shortcomings of the target region, and generates a visual diagnostic report.

[0062] The visual diagnostic report includes a time-series evolution chart of fusion degree, a comparison chart before and after correction, and a supply-demand matching bottleneck chart; the time-series evolution chart of fusion degree plots the fusion coordination degree D after correction with time as the horizontal axis. final The change curve is displayed, and the fusion level passing line is marked; a comparison chart before and after correction is drawn, and the original fusion degree D is plotted in the same coordinate system. raw With the corrected fusion coordination degree D final The comparison curves are used, and the difference area between the two curves is filled to show the degree of supply-demand mismatch; the supply-demand matching bottleneck chart uses a bar chart to compare the actual scientific research supply score S at each time point. final Score I of actual industry demand final It automatically labels whether scientific research is ahead of schedule or industry is lagging behind based on the difference.

[0063] like Figure 2As shown, an innovation chain and industrial chain collaborative diagnostic system integrating multi-source data includes a multi-source data acquisition module, a policy index calculation module, a dual-track correction calculation module, a prompt word construction module, a model inference invocation module, and a visualization report generation module. The multi-source data acquisition module is used to acquire structured indicator data from the scientific research supply side, structured indicator data from the industry demand side, and unstructured science and technology policy text data. It can also normalize and weight the structured indicator data to output the original supply score. Compared with the original demand score The policy index calculation module is used to parse and extract features from unstructured science and technology policy text data, and calculate the policy synergy effectiveness index UOS; the dual-track correction calculation module is used to receive the market mechanism parameter S. m It also incorporates the policy synergy effectiveness index UOS to perform nonlinear dual-track correction, outputting the true scientific research supply score S. final Real Industry Demand Score I final and the original fusion coordination degree D raw And the corrected fusion coordination degree D final The prompt word construction module is used to construct prompt words based on D. raw D final S final I final The UOS constructs structured prompt words; the model inference calling module is used to input the structured prompt words into the pre-trained large language model DeepSeek-V3 and obtain the collaborative weakness analysis results and decision suggestion text; the visualization report generation module is used to generate reports based on D... raw D final S final I final Generate a diagnostic report that includes a time-series evolution diagram of fusion degree, a comparison diagram before and after correction, and a supply-demand matching shortcoming diagram, and integrate decision-making suggestion text output.

[0064] The foregoing description of specific exemplary embodiments of the invention is for illustrative and explanatory purposes. These descriptions are not intended to limit the invention to the precise forms disclosed, and it will be apparent that many changes and variations can be made in accordance with the foregoing teachings. The exemplary embodiments were chosen and described in order to explain the specific principles of the invention and its practical application, thereby enabling those skilled in the art to implement and utilize various different exemplary embodiments of the invention, as well as various different choices and variations. The scope of the invention is intended to be defined by the claims and their equivalents.

Claims

1. A collaborative diagnostic method for innovation chains and industrial chains that integrates multi-source data, characterized in that, Includes the following steps: Step S01: Multi-source data acquisition; The system collects structured index data on the research supply side and structured index data on the industry demand side of the target region within a preset time window, as well as unstructured science and technology policy text data related to the target region. The structured index data on the research supply side and the structured index data on the industry demand side are normalized and then weighted and summed according to preset weights to obtain the original research supply score. Demand Score for Primary Industries ; Step S02: Data standardization processing; Natural language processing technology is used to parse, extract features, and perform quantitative analysis on the unstructured science and technology policy text data. Based on the policy coordination effectiveness scoring model, the policy coordination effectiveness index UOS within the preset time window is calculated. Step S03: Data fusion processing; Introducing market mechanism parameter S m Using the policy synergy effectiveness index UOS obtained in step S02 as a correction factor, a nonlinear dual-track correction model is constructed; the original scientific research supply score is then adjusted using the nonlinear dual-track correction model. and the score of demand for primary industries Dynamic adjustments are made to obtain the corrected true research supply score S. final Score I of actual industry demand final Then, based on the original research supply score... and the score of demand for primary industries The original fusion coordination degree D was calculated. raw According to the actual scientific research supply score S final Score I of actual industry demand final The corrected fusion coordination degree D was calculated. final ; Step S04: Construct structured prompts; Extracting the original fusion coordination degree D raw With the corrected fusion coordination degree D final The differences in characteristics, and the actual scientific research supply score S final Score I of actual industry demand final Based on the comparative characteristics and combined with the UOS policy coordination effectiveness index, a structured cue word containing role setting, data context and diagnostic objectives is constructed. Step S05: Output the analysis results and decision recommendations in text format; The structured prompts are input into the pre-trained DeepSeek-V3 large language model, which then outputs the analysis results and decision-making suggestions for the shortcomings of innovation chain and industrial chain collaboration in the target region, and generates a visual diagnostic report.

2. The innovation chain and industrial chain collaborative diagnosis method integrating multi-source data as described in claim 1, characterized in that: In step S01, unstructured science and technology policy text data includes the full text or parsable text of science and technology plans, implementation rules, regulations, and guiding opinions.

3. The innovation chain and industrial chain collaborative diagnosis method integrating multi-source data as described in claim 2, characterized in that: In step S02, text extraction and cleaning are performed on each unstructured science and technology policy text to obtain the unstructured science and technology policy text content for analysis; the keyword density score of each unstructured science and technology policy text is calculated based on a preset innovation chain and industry chain keyword library; when there is no available unstructured science and technology policy text within the preset time window, a preset default benchmark value is assigned; the average keyword density score of all valid unstructured science and technology policy texts within the preset time window is calculated to obtain the policy synergy effectiveness index UOS.

4. The innovation chain and industrial chain collaborative diagnosis method integrating multi-source data as described in claim 1, characterized in that: In step S03, the actual scientific research supply score S final The calculation formula is: S final =S raw ×(1+α×UOS), Real Industry Demand Score I final The calculation formula is: I final =I raw ×(1+β×S m ), in, This is the preset adjustment weight coefficient.

5. The innovation chain and industrial chain collaborative diagnosis method integrating multi-source data according to claim 4, characterized in that: In step S03, the market mechanism parameter S m The calculation process is as follows: Construct a set of research supply texts and a set of industry demand texts; use a pre-trained semantic representation model based on deep neural networks to map the research supply texts and industry demand texts into high-dimensional semantic vectors respectively, and calculate the cosine similarity matrix between the supply semantic vectors and the demand semantic vectors; aggregate the cosine similarity matrix to obtain the mean or weighted mean of the supply-demand semantic fit, and use the mean or weighted mean of the supply-demand semantic fit as the market mechanism parameter S. m .

6. The innovation chain and industrial chain collaborative diagnosis method integrating multi-source data according to claim 1, characterized in that: In step S04, the process of constructing structured prompt words is as follows: the task role of the DeepSeek-V3 large language model is preset as a regional industrial economic analysis expert in the prompt words; Target area identifier, time window identifier, and original fusion coordination degree D raw With the corrected fusion coordination degree D final The difference, and the real scientific research supply score S final Score I of actual industry demand final The difference is used as the data context to fill in the prompt word template; The prompts specify that the output should include the attribution type for the fusion of shortcomings and corresponding suggestions. The attribution types of integration shortcomings include at least three types: advanced scientific research, lagging industry, and supply-demand mismatch.

7. The innovation chain and industrial chain collaborative diagnosis method integrating multi-source data according to claim 1, characterized in that, In step S05, the visual diagnostic report includes a fusion degree time-series evolution diagram, a comparison diagram before and after correction, and a supply-demand matching bottleneck diagram; the fusion degree time-series evolution diagram plots the fusion coordination degree D after correction with time as the horizontal axis. final The change curve is displayed, and the fusion level passing line is marked; a comparison chart before and after correction is drawn, and the original fusion degree D is plotted in the same coordinate system. raw With the corrected fusion coordination degree D final The comparison curves are used, and the difference area between the two curves is filled to show the degree of supply-demand mismatch; the supply-demand matching bottleneck chart uses a bar chart to compare the actual scientific research supply score S at each time point. final Score I of actual industry demand final It automatically labels whether scientific research is ahead of schedule or industry is lagging behind based on the difference.

8. A collaborative diagnostic system for innovation chains and industrial chains that integrates multi-source data, characterized in that: The method described in any one of claims 1-7 shall be employed.

Images