Automatic generation method and device of ESG suggestion report, equipment and storage medium

Abstract

The invention relates to the technical field of data processing, and discloses an automatic generation method and device of an ESG suggestion report, equipment and a storage medium, and an initial ESG report is obtained by automatically screening enterprise original data and combining data of a life cycle evaluation database for analysis. And inputting the initial ESG report into the generative intelligent model to obtain an ESG suggested theme. And analyzing and processing the ESG suggested theme to obtain a search result, and analyzing the search result by using the generative intelligent model to obtain a guidance strategy. And optimizing the guidance strategy by using the evaluation model, and generating a professional and authoritative ESG suggestion report. According to the method provided by the invention, the analysis processing efficiency of the initial ESG report is greatly improved, an enterprise is helped to understand the ESG performance deeper, and powerful support is provided for formulating and implementing a sustainable development strategy, so that the enterprise is helped to cope with ESG challenges better, and the sustainable development goal is achieved.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and particularly to an automatic generation method, device, equipment and storage medium for ESG recommendation reports. Background Art

[0002] In the current era of rapid informatization development, all industries are facing the need for processing and analyzing massive amounts of data. Whether it is the internal operation management and market research of enterprises, or the academic research and policy formulation of scientific research institutions, it is inseparable from the collection, collation and in-depth analysis of a large number of documents and data. These processes often involve various formats of files (such as documents, tables, images, etc.) and complex and diverse data sources, requiring the ability to extract key information from them, summarize, compare and analyze, in order to support decision-making and report writing.

[0003] However, due to the high complexity of the data processing technology field, most of the existing solutions rely on highly manual processes to handle these tasks. Staff need to manually open and read files, enter useful information into a dedicated database or table, and then use various independent software tools for data integration, cleaning and analysis. This series of steps is not only cumbersome, but also highly dependent on personal experience and judgment. Moreover, at the end of data processing, the writing of reports and the putting forward of relevant recommendations also often rely on manual synthesis of all analysis results, which is not only time-consuming and laborious, but may also lead to deviations in conclusions due to subjective factors.

[0004] Therefore, how to improve the efficiency and accuracy of data processing, reduce manual intervention, avoid deviations caused by human subjective judgment, and provide efficient and intelligent data analysis and decision-making support for the industry has become an important problem that needs to be solved urgently. Summary of the Invention

[0005] In view of this, the embodiments of this application provide an automatic generation method, device, computer equipment and storage medium for ESG recommendation reports, which can effectively solve the core problems of the existing technology adopting a "manual + tool" hybrid mode, such as large consumption of human resources, high error rate, slow processing speed and over-reliance on manual work.

[0006] In a first aspect, the embodiments of this application provide an automatic generation method for ESG recommendation reports, including: Processing the original enterprise data to generate an initial ESG report; Inputting the initial ESG report into a generative intelligent model for analysis to obtain ESG recommendation topics; Analyzing and processing the ESG recommendation topics to obtain search results; Using the generative intelligent model to perform policy analysis on the search results to obtain guiding policies; Optimize the guidance strategy using a preset evaluation model to generate an ESG recommendation report.

[0007] In some embodiments, processing the enterprise's original data and analyzing it in combination with the life cycle assessment database to generate an initial ESG report, including: Filter the enterprise's original data according to business type, geographical location, and time range to obtain eligible enterprise data; Fill the eligible enterprise data into a preset initial ESG report template, and refer to the relevant data in the life cycle assessment database during the filling process; Use the preset formula in the initial ESG report template to calculate the eligible enterprise data to generate the initial ESG report.

[0008] In some embodiments, inputting the initial ESG report into a generative intelligent model for analysis to obtain ESG recommendation themes, including: Inputting the initial ESG report into a generative intelligent model for analysis to obtain ESG recommendation themes, including: Input the initial ESG report into a generative intelligent model for analysis to generate guidance recommendation information on the enterprise's ESG status; Process the guidance recommendation information on the enterprise's ESG status to obtain ESG recommendation themes.

[0009] In some embodiments, analyzing and processing the ESG recommendation themes to obtain search results, including: Analyze the ESG recommendation themes to generate theme embeddings corresponding to the ESG recommendation themes; Use the theme embeddings as index values, merge them with the index values pre-generated in the life cycle assessment database and the index values pre-generated from the enterprise's original data to form embedding vectors and perform a search to obtain search results related to the ESG recommendation themes.

[0010] In some embodiments, using the generative intelligent model to perform strategy analysis on the search results to obtain a guidance strategy, including: Input the search results related to the ESG recommendation themes into the generative intelligent model for analysis to obtain the guidance strategy expressed by professional tones and professional words.

[0011] In some embodiments, optimizing the guidance strategy using a preset evaluation model to generate an ESG recommendation report, including: Optimize the guidance strategy represented by the preset professional intonation and professional words using a preset evaluation model to obtain the quality score of the guidance strategy; When the quality score is lower than a preset threshold, adjust the parameters of the generative intelligent model to regenerate the guidance strategy until the quality score of the guidance strategy is higher than or equal to the preset threshold.

[0012] In some embodiments, after the step of when the quality score is lower than a preset threshold, adjusting the parameters of the generative intelligent model to regenerate the guidance strategy until the quality score of the guidance strategy is higher than or equal to the preset threshold, further includes: Convert the guidance strategy that is higher than or equal to the preset threshold into a preset format to generate an ESG recommendation report.

[0013] In a second aspect, an embodiment of the present application provides an automatic generation device for an ESG recommendation report, including: An initial report generation module, configured to process enterprise raw data, analyze it in combination with data in a life cycle assessment database, and generate an initial ESG report; A theme generation module, configured to input the initial ESG report into a generative intelligent model for analysis to obtain an ESG recommendation theme; A search result acquisition module, configured to analyze and process the ESG recommendation theme to obtain search results; A guidance strategy generation module, configured to perform strategy analysis on the search results using the generative intelligent model to obtain a guidance strategy; A recommendation report generation module, configured to optimize the guidance strategy using a preset evaluation model to generate an ESG recommendation report.

[0014] In a third aspect, an embodiment of the present application provides a computer device, which includes a processor and a memory. The memory stores a computer program, and the processor is configured to execute the computer program to implement the automatic generation method of the ESG recommendation report in the first aspect above.

[0015] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, and when the computer program is executed on a processor, it implements the automatic generation method of the ESG recommendation report in the first aspect above.

[0016] The embodiments of the present application have the following beneficial effects: An automatic generation method, device, computer device, and storage medium for an ESG recommendation report of the present application significantly improve the management efficiency and decision-making quality of enterprises in the aspects of environment, society, and governance (ESG). First, the original data of the enterprise is efficiently processed, and combined with the data in the life cycle assessment database, a detailed initial ESG report is generated. This process not only reduces the time and labor costs of manual data collation but also ensures the comprehensiveness and accuracy of the data. Then, the initial ESG report is input into a generative intelligent model for in-depth analysis to extract key ESG recommendation themes. This step uses advanced natural language processing technology to quickly identify and focus on the core improvement points of the enterprise in the ESG field. Subsequently, further analysis and processing are performed on the extracted ESG recommendation themes to generate relevant search results, providing rich reference materials and industry best practices. The generative intelligent model is used to perform strategic analysis on these search results to generate practical guiding strategies, thereby providing strong support for the enterprise to formulate a scientific ESG improvement plan. Finally, a preset evaluation model optimizes the guiding strategies to ensure that the proposed strategies are not only operable but also can maximize the ESG performance of the enterprise, and ultimately generate a comprehensive and targeted ESG recommendation report. The method of the present application not only improves the efficiency and accuracy of data processing, reduces manual intervention, and avoids biases caused by subjective judgment but also provides efficient and intelligent ESG data analysis and decision-making support for enterprises, helping enterprises achieve greater breakthroughs and progress on the road of sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 FIG. shows an application scenario diagram of an automatic generation method for an ESG recommendation report according to an embodiment of the present application; Figure 2 FIG. shows a flowchart of an automatic generation method for an ESG recommendation report according to an embodiment of the present application; Figure 3 FIG. shows another flowchart of an automatic generation method for an ESG recommendation report according to an embodiment of the present application; Figure 4 FIG. shows a schematic diagram of an initial ESG report of an automatic generation method for an ESG recommendation report according to an embodiment of the present application; Figure 5Shows another flowchart of a method for automatically generating an ESG recommendation report according to an embodiment of the present application; Figure 6 Shows another flowchart of a method for automatically generating an ESG recommendation report according to an embodiment of the present application; Figure 7 Shows another flowchart of a method for automatically generating an ESG recommendation report according to an embodiment of the present application; Figure 8 Shows a schematic diagram of the power consumption of building material suppliers involved in a method for automatically generating an ESG recommendation report according to an embodiment of the present application; Figure 9 Shows a schematic structural diagram of an apparatus for automatically generating an ESG recommendation report according to an embodiment of the present application. Detailed implementation manners

[0019] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments.

[0020] Generally, the components of the embodiments of the present application described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application claimed, but merely represents selected embodiments of the present application. All other embodiments obtained by those skilled in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application.

[0021] Hereinafter, the terms "including", "having" and their cognates that can be used in various embodiments of the present application are only intended to represent specific features, numbers, steps, operations, elements, components or combinations of the foregoing items, and should not be construed as first excluding the existence of one or more other features, numbers, steps, operations, elements, components or combinations of the foregoing items or increasing the possibility of one or more features, numbers, steps, operations, elements, components or combinations of the foregoing items. In addition, the terms "first", "second", "third", etc. are only used for differentiating descriptions and cannot be construed as indicating or implying relative importance.

[0022] Unless otherwise limited, all terms (including technical terms and scientific terms) used here have the same meaning as those commonly understood by those of ordinary skill in the art to which various embodiments of the present application belong. The terms (such as those defined in a general-use dictionary) will be interpreted as having the same meaning as the contextual meaning in the relevant technical field and will not be interpreted as having an idealized meaning or an overly formal meaning unless clearly defined in various embodiments of the present application.

[0023] The following will, in conjunction with the accompanying drawings, elaborate on some embodiments of the present application. Without conflict, the following embodiments and the features in the embodiments may be combined with each other.

[0024] Considering that the existing technology adopts a "manual + tool" hybrid mode, with core problems such as high consumption of human resources, high error rate, slow processing speed, and over - reliance on manual work, therefore, an automatic generation method for ESG recommendation reports is proposed. By automatically screening the original enterprise data, an initial ESG report is obtained. Subsequently, the initial ESG report is input into a generative intelligent model to obtain ESG recommendation topics. Then, the ESG recommendation topics are analyzed and processed to obtain search results, and the generative intelligent model is used to deeply analyze the search results to obtain guiding strategies. After that, an evaluation model is used to optimize the guiding strategies, and finally a professional and authoritative ESG recommendation report is generated. The method of the present application not only greatly improves the analysis and processing efficiency of the initial ESG report, but also helps enterprises to more deeply understand their ESG performance, provides strong support for formulating and implementing sustainable development strategies, and helps enterprises better cope with ESG challenges and achieve sustainable development goals.

[0025] An automatic generation method for ESG recommendation reports provided by an embodiment of the present application can be applied in an application environment such as Figure 1 Specifically, the automatic generation method for ESG recommendation reports of the present application is applied in a computer system, and the system includes such as Figure 1The client and server shown, where the client communicates with the server over a network. The client, also known as the user side, refers to a program that provides local services to clients corresponding to the server. The client can be installed on, but not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The server can be implemented with an independent server or a server cluster composed of multiple servers. Specifically, first, the original data of the enterprise is processed on the server side and analyzed in combination with the data in the life cycle assessment database to generate an initial ESG report. Subsequently, the initial ESG report is transmitted to the generative intelligent model on the server side for in-depth analysis to extract ESG recommendation topics. The extracted ESG recommendation topics are then further analyzed and processed on the server side to obtain relevant search results. These search results are again analyzed through the generative intelligent model for strategy analysis to generate guiding strategies. Finally, a preset evaluation model optimizes these guiding strategies to generate the final ESG recommendation report. Throughout the process, the client is mainly responsible for user interaction operations, such as data upload, report viewing, and strategy feedback, while complex data processing, intelligent analysis, and report generation are efficiently completed by the server side. By reasonably integrating the functions of the client and the server, not only is the efficiency and accuracy of data processing improved, manual intervention is reduced, and the deviation caused by human subjective judgment is avoided, but also efficient and intelligent ESG data analysis and decision support are provided for enterprises, significantly enhancing the enterprise's management level and sustainable development ability in the aspects of environment, society, and governance (ESG).

[0026] Figure 2 FIG. shows a flowchart of a method for automatically generating an ESG recommendation report according to an embodiment of the present application. The method for automatically generating the ESG recommendation report includes the following steps: Step S100, process the original data of the enterprise and analyze it in combination with the life cycle assessment database to generate an initial ESG report.

[0027] Exemplarily, the original data of the enterprise refers to a dataset that has been preliminarily screened and sorted and complies with industry norms and enterprise internal standards. These datasets cover various information of the enterprise, such as production activity records, energy consumption details, waste emission statistics, employee welfare status, and corporate governance structure. After obtaining the original data of the enterprise, a series of processing is required: for example, first, error and redundant information are removed through data cleaning; second, data conversion is performed to ensure the consistency and standardization of the data format; finally, data analysis is carried out to dig out the deep meaning and value behind the data. After this series of processing, the standard enterprise data is obtained, and the obtained standard enterprise data is filled into a pre-designed initial ESG report template to generate an initial ESG report.

[0028] Among them, the Life Cycle Assessment (LCA) database aims to comprehensively evaluate the environmental impact of a product throughout its entire life cycle, namely from raw material acquisition, production and manufacturing, use stage to waste treatment. The LCA database covers key information such as environmental impact assessment data of various raw materials, energy consumption and emissions in different production processes. By combining and applying the relevant data of the LCA database, enterprises can more deeply evaluate the comprehensive environmental impact of their operations, thereby further enhancing the comprehensiveness and accuracy of the initial ESG report.

[0029] The initial ESG report is an important document for evaluating an enterprise's performance in three dimensions: Environmental, Social, and Governance. Through the initial ESG report, an enterprise can clearly demonstrate to the outside world its positive actions in environmental protection, social responsibility fulfillment, and improvement of corporate governance structure, thereby enhancing investors' confidence and improving the enterprise's social image and brand value.

[0030] In an alternative embodiment, as Figure 3 shown, in step S100, the enterprise's original data is processed and analyzed in combination with the life cycle assessment database to generate an initial ESG report, including: Step S110, screening the enterprise's original data according to business type, geographical location, and time range to obtain eligible enterprise data.

[0031] For example, the enterprise's original data refers to the data extracted from documents such as financial reports, invoices, and receipts. Data screening is divided into several dimensions: First, screening according to the main business or industry classification of the enterprise, such as manufacturing, service, new energy, and environmental protection. Second, screening according to the geographical location of the enterprise, including countries, regions, cities, or more specific geographical locations (such as industrial parks, business districts, etc.). Finally, screening according to the time attribute of the data, covering enterprise data in the past year, past quarter, or a specific time period. That is, screening the enterprise's original data according to the three dimensions of business type, geographical location, and time range to obtain data rows, data columns, and specific values directly related to the initial ESG report and meeting specific conditions.

[0032] Step S120, filling the eligible enterprise data into a preset initial ESG report template and referring to the relevant data of the life cycle assessment database during the filling process.

[0033] Exemplarily, during the process of populating the initial ESG report template with eligible enterprise data, relevant data in the Life Cycle Assessment (LCA) database is referred to. That is, when analyzing the environmental impact of an enterprise, not only the data provided by the enterprise is relied on, but also the background information or industry standards provided by external databases are utilized. For example, information on aspects such as the carbon footprint of products, resource usage, and life cycle costs helps to more comprehensively evaluate the ESG performance of the enterprise.

[0034] For example, the initial ESG report template is a pre-designed document or form that contains multiple sections and sub-sections. Each section corresponds to one or more key aspects of the initial ESG report and is used to organize and present the content of the initial ESG report. For example, the initial ESG report template contains multiple sections, and each section has a dedicated area for data display and explanation, covering information such as type, source, usage amount, emission indicators, emissions volume, total emissions volume, etc., and may include comparisons with industry standards, competitors, or historical data, thereby providing key decision-making basis for investors, regulatory agencies, consumers, and other stakeholders. These templates help enterprises to more comprehensively understand their ESG performance and formulate improvement strategies accordingly.

[0035] Step S130, calculate the eligible enterprise data using the preset formula in the initial ESG report template to generate the initial ESG report.

[0036] Exemplarily, the initial ESG report template usually also contains a series of preset formulas and metrics. The preset formulas are used to calculate the enterprise data that has been screened and corrected with reference data, and then the calculation results are summarized in the initial ESG report, thus forming a complete report to provide a quantitative analysis of the enterprise's sustainability and social responsibility.

[0037] For example, for relevant data such as electricity bills and raw material purchase orders submitted by building material suppliers, enterprise data is filtered and screened according to business type, geographical location, and time range. Next, the filtered and screened data is filled into the corresponding initial ESG report template. For example, the electricity consumption data of ovens and generators of an enterprise headquarters from 2023 to 2024 is filled into the template part of the fixed emission sources of the enterprise headquarters for that year; at the same time, the raw material purchase quantity data for that year is filled into the template part of the indirect emission sources for that year. During the data filling process, relevant data from the LCA database is combined for quantitative analysis, including greenhouse gas emission indicators of different electrical appliances and raw materials during their life cycles. Subsequently, the initial ESG report template integrated with enterprise data and LCA database information is processed computationally, following the formulas preset in the initial ESG report template. For example, the electricity consumption of the oven is multiplied by the emission factors of various greenhouse gases respectively according to the energy composition of the electricity and its corresponding emission factors (e.g., coal, natural gas, hydropower, nuclear energy, renewable energy, etc.), and then the emissions of each greenhouse gas are accumulated to finally obtain the total emissions. The total emissions data obtained from these computations is filled into the corresponding positions in the initial ESG report template, thus generating a complete initial ESG report, as Figure 4 shown.

[0038] Step S200, input the initial ESG report into a generative intelligent model for analysis to obtain ESG recommendation topics.

[0039] Demonstratively, generative artificial intelligence (Generative AI) models, such as the GPT series (e.g., GPT-3, GPT-4), T5 (Text-To-Text Transfer Transformer), etc., these models can understand and generate natural language text, and use machine learning and deep learning technologies to generate new content by analyzing historical data patterns. The generative intelligent model can learn the information in the database and generate new content based on the prompts or conditions input by the user.

[0040] For example, before inputting the initial ESG report into the generative intelligent model for analysis, first perform data cleaning on the collected initial ESG report to remove irrelevant information, noisy data, and duplicate content. Then, convert the text data in the report into a format suitable for model processing, including steps such as word segmentation, stop word removal, and stemming. Subsequently, select a suitable generative intelligent model, such as the GPT series, T5, or other large language models, according to the analysis requirements and data processing capabilities.

[0041] To improve the accuracy and relevance of obtaining ESG recommendation topics, this application optimizes the RAG (Retrieval-Augmented Generation) method. Specifically, during the chunking and embedding process, background information is input together with each paragraph and considered before being converted into an embedding representation. In addition to using embedding search, TF-IDF (Term Frequency-Inverse Document Frequency) encoding is also combined to enhance the keyword search ability of the RAG database. Before generating the final recommendation topics, a re-ranker network is introduced to optimize the ranking of the search results. Subsequently, the model is further trained by adjusting parameters such as model parameters, optimizers, and learning rates to improve its overall performance and accuracy. To evaluate the stability and reliability of the model, various methods such as cross-validation are also adopted to ensure its performance consistency on different datasets.

[0042] After the model training is completed, the initial ESG report is used as input data and input into the trained generative intelligent model to extract features of the text data in the initial ESG report. For example, term frequency, TF-IDF values, word vectors, etc. These features can reflect the key information and topics in the text. The model will extract key topic features related to ESG from the report, and these features may cover aspects such as the enterprise's environmental performance indicators, social responsibility projects, and governance practices.

[0043] In other embodiments, the model can be further optimized according to the quality and accuracy of the generated topics. For example, methods such as adjusting the model structure, increasing training data, and using more advanced natural language processing techniques. To verify the performance of the model, an independent initial ESG report dataset can be used to evaluate the model, and metrics such as the accuracy, recall rate, and F1 score of the model can be calculated to evaluate its accuracy and reliability in different scenarios. Finally, the initial ESG report is input into the deployed model to generate ESG recommendation topics, providing strong support for the formulation and improvement of the enterprise's ESG strategy.

[0044] In an alternative embodiment, as Figure 5 shown, in step S200, the initial ESG report is input into the generative intelligent model for analysis to obtain ESG recommendation topics, including: Step S210, input the initial ESG report into the generative intelligent model for analysis to generate guiding advice information on the enterprise's ESG status; Exemplarily, the guidance and recommendation information includes generalization directions and insights. Among them, the generalization direction refers to the general or overall direction regarding the enterprise's ESG performance and development trends extracted from the initial ESG report. For example, the enterprise's progress in environmental protection, the fulfillment of social responsibilities, and the optimization of corporate governance structure, etc. Insights refer to the in-depth understanding and insights obtained based on the analysis of the initial ESG report data and the generative intelligent model. The generalization directions and insights reveal the key information, patterns or trends in the initial ESG report, as well as their potential impact on the enterprise's future development.

[0045] Step S220: Process the guidance and recommendation information on the enterprise's ESG status to obtain ESG recommendation topics.

[0046] Exemplarily, after obtaining the guidance and recommendation information on the enterprise's ESG status, search for all relevant ESG recommendation topics among the United Nations' seventeen Sustainable Development Goals based on the guidance and recommendation information. For example: (1) Recommendations in the environmental aspect: Improve energy efficiency and reduce carbon emissions; Promote the use of renewable energy and reduce dependence on fossil energy; Implement green supply chain management to ensure the environmental compliance of suppliers. (2) Recommendations in the governance aspect: Improve the corporate governance structure, enhance decision-making efficiency and transparency; Strengthen risk management to ensure the stable operation of the enterprise; Promote the implementation of sustainable development strategies to ensure long-term value creation.

[0047] That is, based on the guidance and recommendation information analyzed from the initial ESG report, the topics of improvement suggestions or strategies for the enterprise's ESG status can be determined, which helps the enterprise better identify the strengths and weaknesses in its ESG performance and formulate effective strategies for improvement and enhancement.

[0048] Step S300: Analyze and process the ESG recommendation topics to obtain search results.

[0049] Exemplarily, each ESG recommendation topic will be analyzed independently to generate a topic embedding. These embedding vectors are used to construct an index, which is combined with the pre-generated index vectors in the LCA database and the original data index vectors provided by the enterprise for embedding vector search, so as to obtain relevant search results.

[0050] In an alternative embodiment, as Figure 6 shown, in step S300, analyzing and processing the ESG recommendation topics to obtain search results includes: Step S310: Analyze the ESG recommendation topics to generate topic embeddings corresponding to the ESG recommendation topics.

[0051] Demonstratively, topic embedding is a technique that represents topics in the form of vectors that can capture semantic relationships and similarities between topics. So that in the subsequent steps, through topic embedding, the ESG recommendation topics can be converted into a data format that is easy to process. The generated topic embedding is combined with the index vectors of the LCA database and the original data of the enterprise, and the embedded vector search technology is used to perform efficient search and matching in the initial ESG report, database or library. This process helps users quickly find data and information related to the ESG recommendation topics and improves the accuracy and efficiency of data retrieval.

[0052] Step S320, embedding the topic as an index value, merging it with the index value pre-generated by the LCA database and the index value pre-generated by the original enterprise data, forming an embedding vector and performing a search to obtain search results related to the ESG recommendation topic.

[0053] Exemplarily, the generated topic embedding is used as a keyword for search or indexing, the relevant data of the LCA database is converted into index values, and the original data of the enterprise is converted into index values. Then the topic embedding of the ESG recommendation topic, the index value of the LCA database, and the index value of the original data of the enterprise are merged to form a larger set of embedding vectors. The merged set of embedding vectors is used for the search task, that is, searching for the LCA data or original data of the enterprise that is most relevant to the specific ESG recommendation topic in the embedding vector space. That is, the search results include search results directly related to the ESG recommendation topic.

[0054] In short, using data analysis and processing technology, the ESG recommendation topics, LCA database and corporate raw data are closely integrated to provide strong support for the formulation of corporate sustainable development strategies. By searching and analyzing these relevant data, companies can better understand their ESG performance and formulate improvement strategies accordingly.

[0055] Step S400: Utilize the generative intelligent model to perform strategic analysis on the search results to obtain a guiding strategy.

[0056] Demonstratively, the data and information related to the ESG recommendation topic obtained through the embedding vector search is input into the generative intelligent model, so that the generative intelligent model performs strategic analysis on the input search results, for example, analyzing historical data and current trends, and predicting the future development direction of the ESG field. Evaluate the potential risks of different ESG issues to enterprises or industries, including changes in environmental regulations, social pressure, corporate governance issues, etc. And identify potential opportunities in the ESG field, such as green investment opportunities, social responsibility projects, etc. Then, based on the above analysis, the generative intelligent model generates specific ESG guidance strategies, such as improvement suggestions, action plans, goal setting, potential risk warnings, etc.

[0057] In an alternative embodiment, in step S400, a generative intelligent model is used to perform a policy analysis on the search results to obtain a guidance policy, including: The search results related to the ESG recommendation theme are input into the generative intelligent model for analysis to obtain an ESG guidance policy represented by professional tones and professional words.

[0058] Exemplarily, the data and information related to the ESG recommendation theme obtained through embedded vector search are input into the generative intelligent model for in-depth analysis. Based on the analysis of the relevant data and information, the generative intelligent model will provide a guidance policy described in professional words and professional tones. Among them, professional words include professional terms, industry jargons, and relevant keywords in related fields; while professional tones refer to a formal, objective, and rigorous expression method.

[0059] It should be noted that the generative intelligent model is an artificial intelligence model trained with a large number of professional recommendation documents. Therefore, when generating recommendations, it can use professional tones and words, and provide in-depth, specific, and practical guidance policies. For example, in the actual application process, the feedback provided by the enterprise can also be used to further optimize and improve the generative intelligent model to enhance the quality and accuracy of the guidance policy.

[0060] In step S500, a preset evaluation model is used to optimize the guidance policy to generate an ESG recommendation report.

[0061] Exemplarily, the evaluation model can be a classifier or a regressor based on machine learning, or a deep learning model, such as a BERT model, a GPT model, etc. According to the characteristics of the guidance policy, evaluation criteria are set, such as accuracy, relevance, integrity, readability, professionalism, etc., and weights are set for each criterion to reflect its importance in the overall evaluation. Then, according to the evaluation criteria and weights, a preset threshold for quality scoring is set. The guidance policy is input into the preset evaluation model, and the evaluation model scores the guidance policy according to the preset evaluation criteria and weights. According to the scoring results, the generative intelligent model is adjusted to generate an ESG recommendation report.

[0062] In an alternative embodiment, as Figure 7 shown, in step S500, a preset evaluation model is used to optimize the guidance policy to generate an ESG recommendation report, including: In step S510, a preset evaluation model is used to optimize the guidance policy represented by professional tones and professional words to obtain a quality score of the guidance policy.

[0063] Exemplarily, the preset evaluation model is the BERT evaluation model. BERT (Bidirectional Encoder Representations from Transformers) is an advanced natural language processing (NLP) technology that can deeply understand the context information of text, thereby performing more accurate analysis and evaluation. Through the BERT evaluation model, the tone and words in the guidance strategy can be adjusted and optimized. For example, change 'immoral' to the more accurate 'violate moral standards'; change 'waste' to the more professional 'low resource utilization rate', etc.

[0064] Subsequently, the BERT evaluation model will score the guidance strategy according to the preset evaluation criteria and weights. The evaluation criteria usually cover multiple dimensions such as the accuracy, relevance, integrity, readability, and professionalism of the suggestions to ensure the comprehensiveness and objectivity of the evaluation. Each criterion will be assigned corresponding weights to reflect its importance in the overall evaluation.

[0065] Finally, the evaluation results will include the scores for each criterion and the overall quality score, which will serve as an important basis for us to adjust and optimize the generative intelligence model to further improve the quality and accuracy of the guidance strategy.

[0066] Step S520, when the quality score is lower than the preset threshold, adjust the parameters of the generative intelligence model to regenerate the guidance strategy until the quality score of the guidance strategy is higher than or equal to the preset threshold.

[0067] Exemplarily, the setting of the preset threshold is based on multiple metrics, including but not limited to readability and tone. For example, readability is evaluated by the Flesch Reading Ease, reading time, number of words, number of sentences, and number of difficult words. And the tone is analyzed by a tone analysis tool to analyze the confidence level and analytical nature of the article. These metrics will be comprehensively considered to generate a quality score.

[0068] To ensure high standards, the preset threshold for the quality score is set as the upper quartile of the historical report scores. If the quality score is lower than the preset threshold, it indicates that there is room for improvement in the quality of the guidance strategy and further adjustment and optimization are needed, such as adjusting the temperature parameter until the quality score of the guidance strategy is equal to or higher than the preset threshold. It should be noted that adjusting the temperature parameter can make the generative intelligence model pay more attention to diversity or accuracy when generating suggestions, thereby improving the quality of the suggestions.

[0069] Meanwhile, the generative intelligent model also extracts the vocabulary that can be improved and places them in the context window. The context window allows the model to consider the specific meanings and usages of these words in the context when regenerating suggestions, so as to generate more accurate and coherent guidance strategies. Through iterative generation and evaluation, it can be ensured that the finally generated guidance strategy has high quality and accuracy.

[0070] In an alternative embodiment, in step S520, when the quality score is lower than the preset threshold, the parameters of the generative intelligent model are adjusted to regenerate the guidance strategy until the quality score of the guidance strategy is higher than or equal to the preset threshold, and further includes: Convert the guidance strategy that is higher than or equal to the preset threshold into a preset format to generate an ESG recommendation report.

[0071] Exemplarily, arrange the generated guidance strategies in a logical order to form coherent paragraphs. Optimize and adjust the content in the paragraphs to ensure the accuracy and completeness of the information. Combine all the paragraphs into the final ESG recommendation report. Among them, the ESG recommendation report should include multiple parts such as an introduction, ESG recommendation classification, recommendation details, evaluation results, etc., to comprehensively reflect the content and quality of the ESG recommendation report. For example: It is recommended that this building material supplier refer to the best practices of its peers, give priority to purchasing and using environmentally friendly building materials with lower greenhouse gas emissions, and also select efficient and energy-saving production equipment and machines to improve electricity efficiency; for example, in an ESG recommendation report related to a building material supplier, a line chart showing the monthly electricity consumption changes of a certain building material supplier in a year can be presented, such as Figure 8As shown; and the following specific suggestions can also be presented to help building material suppliers reduce electricity use and improve overall operational efficiency: (1) Optimize the production process: Conduct an in-depth analysis of the existing production process to identify and eliminate unnecessary energy-wasting links. By introducing energy-saving equipment and advanced technologies, significantly enhance the energy efficiency of the production line. (2) Implement an energy consumption monitoring system: Install energy consumption monitoring equipment to track electricity usage in real time. This will help identify high-energy-consuming equipment and its usage periods, thereby formulating targeted energy-saving measures. (3) Invest in efficient equipment: Prioritize investment in high-efficiency machinery and equipment, such as variable frequency drives and energy-saving motors. These devices can significantly reduce electricity consumption during operation while maintaining production efficiency. (4) Employee training and awareness improvement: Regularly conduct energy-saving training to enhance employees' awareness of energy use. Encourage employees to put forward energy-saving suggestions and actively take energy-saving measures in their daily work. (5) Strengthen equipment maintenance: Regularly maintain equipment to ensure it operates in the best condition. Good maintenance can not only reduce failure rates but also effectively reduce energy waste. (6) Utilize renewable energy: Consider installing solar panels or other renewable energy systems on the facility to reduce dependence on external electricity. This can not only reduce electricity bills but also enhance the company's environmental image. (7) Make good use of natural light: When there is sufficient natural light near the windows, turn off the indoor lighting. Consider installing light sensors to reduce energy waste caused by human error. (8) Manage electrical appliance use: For electrical appliances that will not be used within the next hour, turn them off or set them to standby mode in a timely manner to avoid unnecessary power consumption. (9) Optimize air conditioner use: Set the air conditioner temperature between 24°C and 26°C to ensure a balance between comfort and energy efficiency. (10) Implement a separate switch system: Adopt a separate switch system in areas such as corridors and flexibly adjust lighting according to actual usage to improve energy utilization efficiency.

[0072] By implementing the above suggestions, building material suppliers can effectively reduce electricity use, improve operational efficiency, and make a positive contribution to sustainable development.

[0073] The method for automatically generating an ESG recommendation report according to the embodiments of the present application. First, by processing the original data of the enterprise, an initial ESG report is generated. Then, the initial ESG report is input into a generative intelligent model for analysis to obtain ESG recommendation topics. Next, the ESG recommendation topics are processed to obtain search results, and the search results are input into the generative intelligent model for analysis to generate guiding strategies. Finally, a preset evaluation model is used to optimize the guiding strategies to generate an ESG recommendation report. The method for automatically generating an ESG recommendation report of the present application not only improves the analysis and processing efficiency of the initial ESG report but also provides an enterprise with a comprehensive, accurate, and standardized initial ESG report to help the enterprise better understand its ESG performance and formulate corresponding improvement strategies.

[0074] The present application also provides an automatic generation device for an ESG recommendation report. As Figure 9 shown, the automatic generation device for the ESG recommendation report includes: An initial report generation module 91, configured to process the original data of an enterprise and generate an initial ESG report; A theme generation module 92, configured to input the initial ESG report into a generative intelligent model for analysis to obtain an ESG recommendation theme; A search result acquisition module 93, configured to analyze and process the ESG recommendation theme to obtain a search result; A guidance strategy generation module 94, configured to perform strategy analysis on the search result by using the generative intelligent model to obtain a guidance strategy; A recommendation report generation module 95, configured to optimize the guidance strategy by using a preset evaluation model and generate an ESG recommendation report.

[0075] It can be understood that the device in this embodiment corresponds to an automatic generation method for an ESG recommendation report in the above embodiment. The optional items in the above embodiment are also applicable to this embodiment, so they will not be described repeatedly here.

[0076] The present application also provides a computer device. Exemplarily, the computer device includes a processor and a memory. Among them, the memory stores a computer program, and the processor runs the computer program to enable the computer device to execute the above automatic generation method for an ESG recommendation report or the functions of each module in the above automatic generation device for an ESG recommendation report.

[0077] Among them, the processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including a central processing unit (CPU), a graphics processing unit (GPU), a network processor (NP), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc., and can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present application.

[0078] The memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electric Erasable Programmable Read-Only Memory (EEPROM), etc. Among them, the memory is used to store a computer program, and after receiving an execution instruction, the processor can execute the computer program accordingly.

[0079] This application also provides a computer-readable storage medium for storing the computer program used in the above computer device. For example, the computer-readable storage medium can include, but is not limited to: various media such as USB flash drives, external hard drives, Read-Only Memory (ROM), Random Access Memory (RAM), magnetic disks, or optical discs that can store program codes.

[0080] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely illustrative. For example, the flowcharts and structure diagrams in the drawings show the possible architectures, functions, and operations of the devices, methods, and computer program products according to multiple embodiments of this application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in an alternative implementation, the functions marked in the blocks can occur in a different order than marked in the drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the structure diagram and / or flowchart, as well as the combination of blocks in the structure diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0081] In addition, each functional module or unit in various embodiments of this application can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part.

[0082] When the above-described function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a smart phone, a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present application.

[0083] As described above, the above are only specific implementation manners of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, and all of them should be covered within the protection scope of the present application.

Claims

1. A method for automatically generating an ESG recommendation report, characterized in that: The method comprises: Process the original data of the enterprise and analyze it in combination with the data from the life cycle assessment database to generate an initial ESG report; Input the initial ESG report into a generative intelligent model for analysis to obtain ESG recommendation topics; Analyze and process the ESG recommendation topic to obtain search results; Performing a strategy analysis on the search results using the generative intelligent model to obtain a guiding strategy; The guidance strategy is optimized using the preset evaluation model to generate an ESG recommendation report.

2. The method for automatically generating an ESG recommendation report according to claim 1, characterized in that: The original data of the enterprise is processed and analyzed in combination with the life cycle assessment database to generate an initial ESG report, including: Filter the original enterprise data according to business type, geographic location and time range to obtain enterprise data that meets the conditions; Fill the qualified enterprise data into the preset initial ESG report template, and refer to the relevant data of the life cycle assessment database during the filling process; The preset formula in the initial ESG report template is used to calculate the qualified enterprise data to generate the initial ESG report.

3. The method for automatically generating an ESG recommendation report according to claim 1, characterized in that: The initial ESG report is input into the generative intelligent model for analysis to obtain ESG recommendation topics, including: Input the initial ESG report into the generative intelligent model for analysis to generate guidance and suggestion information on the ESG status of the enterprise; The guidance and suggestion information on the ESG status of the enterprise is processed to obtain an ESG suggestion theme.

4. The method for automatically generating an ESG recommendation report according to claim 1, characterized in that: The analyzing and processing of the ESG recommendation topic to obtain search results includes: Analyze the ESG recommendation theme and generate a theme embedding corresponding to the ESG recommendation theme; The topic embedding is used as an index value, and is merged with the index value pre-generated in the life cycle assessment database and the index value pre-generated in the enterprise original data to form an embedding vector and perform a search to obtain search results related to the ESG recommendation topic.

5. The method for automatically generating an ESG recommendation report according to claim 1, characterized in that: The using the generative intelligent model to perform a strategy analysis on the search results to obtain a guiding strategy includes: The search results related to the ESG recommendation topic are input into the generative intelligent model for analysis to obtain the guidance strategy expressed by professional tone and professional words.

6. The method for automatically generating an ESG recommendation report according to claim 5, characterized in that: The use of a preset evaluation model to optimize the guidance strategy and generate an ESG recommendation report includes: Optimizing the guidance strategy expressed by the professional tone and professional words by using a preset evaluation model to obtain a quality score of the guidance strategy; When the quality score is lower than a preset threshold, the parameters of the generative intelligent model are adjusted to regenerate the guidance strategy until the quality score of the guidance strategy is higher than or equal to the preset threshold.

7. The method for automatically generating an ESG recommendation report according to claim 6, characterized in that: When the quality score is lower than a preset threshold, adjusting the parameters of the generative intelligent model to regenerate the guidance strategy until the quality score of the guidance strategy is higher than or equal to the preset threshold, further comprising: The guidance strategies that are higher than or equal to the preset threshold are converted according to a preset format to generate an ESG recommendation report.

8. An automatic generation device for an ESG recommendation report, characterized in that: The device comprises: The initial report generation module is used to process the original data of the enterprise and analyze it in combination with the data of the life cycle assessment database to generate an initial ESG report; A theme generation module, used for inputting the initial ESG report into a generative intelligent model for analysis to obtain ESG recommended themes; A search result acquisition module, used to analyze and process the ESG recommendation topic to obtain search results; A guidance strategy generation module, used to perform a strategy analysis on the search results using the generative intelligent model to obtain a guidance strategy; The recommendation report generation module is used to optimize the guidance strategy using a preset evaluation model and generate an ESG recommendation report.

9. A computer device, characterized in that: The computer device comprises a processor and a memory, wherein the memory stores a computer program, and the processor is configured to execute the computer program to implement the method for automatically generating an ESG recommendation report according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: It stores a computer program, which, when executed on a processor, implements the method for automatically generating an ESG recommendation report according to any one of claims 1-7.

Images