CN117575827B - Intelligent visual management system and method for enterprise report - Google Patents

Intelligent visual management system and method for enterprise report Download PDF

Info

Publication number
CN117575827B
CN117575827B CN202410058497.6A CN202410058497A CN117575827B CN 117575827 B CN117575827 B CN 117575827B CN 202410058497 A CN202410058497 A CN 202410058497A CN 117575827 B CN117575827 B CN 117575827B
Authority
CN
China
Prior art keywords
enterprise
information
feature
matrix
financial
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202410058497.6A
Other languages
Chinese (zh)
Other versions
CN117575827A (en
Inventor
刘松国
范诗扬
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhijiang Laboratory Technology Holdings Co ltd
Original Assignee
Zhijiang Laboratory Technology Holdings Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zhijiang Laboratory Technology Holdings Co ltd filed Critical Zhijiang Laboratory Technology Holdings Co ltd
Priority to CN202410058497.6A priority Critical patent/CN117575827B/en
Publication of CN117575827A publication Critical patent/CN117575827A/en
Application granted granted Critical
Publication of CN117575827B publication Critical patent/CN117575827B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/12—Accounting
    • G06Q40/125—Finance or payroll
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/10—Office automation; Time management
    • G06Q10/105—Human resources

Abstract

The application relates to the field of intelligent evaluation, and particularly discloses an intelligent visual management system and method for enterprise reports.

Description

Intelligent visual management system and method for enterprise report
Technical Field
The application relates to the field of intelligent evaluation, in particular to an intelligent visual management system and method for enterprise report.
Background
In the process of seeking funds support for small and medium-sized scientific enterprises, scientific and reasonable estimation becomes a key. Investors want to know the value of an enterprise in order to decide whether to invest or provide funds support. However, traditional financial valuation methods, such as market rates, net rates, etc., may not accurately reflect the potential value of these enterprises. Therefore, in the evaluation process of small and medium-sized scientific enterprises, scientific and reasonable methods are needed to consider factors such as innovation capability, technical advantages, market potential, team strength and the like.
Thus, there is a need for an optimized intelligent visualization management scheme for enterprise reporting.
Disclosure of Invention
The present application has been made to solve the above-mentioned technical problems. The embodiment of the application provides an intelligent visual management system and method for enterprise reports, which adopt a text analysis technology based on artificial intelligence, respectively extract characteristics of human resource management reports of enterprises, financial reports of the enterprises and intellectual property related information of the enterprises to obtain talent structure characteristic vectors of the enterprises, financial analysis characteristic vectors of the enterprises and technical innovation characteristic vectors of the enterprises, then perform characteristic coding on characteristic matrixes formed by the characteristic vectors to obtain enterprise valuation characteristics, and finally input the enterprise valuation characteristics into a generator to obtain a visual enterprise valuation table.
According to one aspect of the present application, there is provided an intelligent visualization management system for enterprise reporting, comprising:
The enterprise report data acquisition module is used for acquiring human resource management reports, financial reports and intellectual property related information of the enterprise to be evaluated;
The human resource report preprocessing module is used for preprocessing the human resource management report of the enterprise to be evaluated to obtain a preprocessed human resource management report;
The financial report preprocessing module is used for preprocessing the financial report of the enterprise to be evaluated to obtain a preprocessed financial report;
the talent structure information extraction module is used for extracting relevant information of staff academic and relevant information of staff research and development staff duty ratio of the enterprise from the preprocessed human resource management report;
The talent structure feature coding module is used for inputting the related information of the staff academic and the related information of the staff research and development staff duty ratio of the staff to a human information text coder to obtain a talent structure feature vector of the enterprise;
The financial related information extraction module is used for extracting enterprise asset related information, liability related information, business income related information and research and development investment related information from the preprocessed financial report;
the financial information encoding module is used for inputting the enterprise asset related information, the liability related information, the business income related information and the research and development investment related information into a financial information text encoder to obtain enterprise financial analysis feature vectors;
The technical innovation information coding module is used for inputting the related information of the enterprise intellectual property rights into a technical innovation information text coder to obtain an enterprise technical innovation feature vector;
the comprehensive information feature construction module is used for constructing the enterprise talent structure feature vector, the enterprise financial analysis feature vector and the enterprise technical innovation feature vector into an enterprise comprehensive information feature matrix;
The enterprise estimation feature generation module is used for inputting the enterprise comprehensive information feature matrix into a feature extractor based on a bidirectional attention mechanism to obtain an enterprise estimation feature matrix;
The optimization module is used for optimizing the enterprise estimation characteristic matrix to obtain an optimized enterprise estimation characteristic matrix;
and the enterprise valuation visualization result generation module is used for enabling the optimized enterprise valuation feature matrix to pass through a generator to obtain a generation result, wherein the generation result is an enterprise valuation visualization table.
In the above-mentioned intelligent visual management system for enterprise report, the talent structure feature coding module includes: the human information word segmentation unit is used for carrying out word segmentation processing on the related information of the staff academic and research staff and the related information of the staff research and development staff of the enterprise so as to obtain a plurality of talent information related words; the human information word embedding unit is used for converting each talent information related word in the plurality of talent information related words into a talent information related word embedding vector through a word embedding layer to obtain a sequence of talent information related word embedding vectors, wherein the embedding layer uses a learnable embedding matrix to carry out embedded coding on each talent information related word; the human information context semantic coding unit is used for inputting the sequence of the talent information related words embedded vector into a converter module of the human information text encoder so as to obtain a plurality of human resource structural feature vectors; and the human information cascading unit is used for cascading the plurality of human resource structural feature vectors to obtain the enterprise talent structural feature vector.
In the above-mentioned intelligent visual management system for enterprise report, the financial information encoding module includes: the financial information word segmentation unit is used for word segmentation processing of the enterprise asset related information, the liability related information, the business income related information and the research and development investment related information to obtain a plurality of financial information related words; a financial information word embedding unit, configured to pass the plurality of financial information related words through a word embedding layer to convert each financial information related word in the plurality of financial information related words into a financial information related word embedding vector to obtain a sequence of financial information related word embedding vectors, where the embedding layer uses a learnable embedding matrix to perform embedding encoding on each financial information related word; a financial information context semantic coding unit for inputting a sequence of the financial information related word embedded vectors into a converter module of the financial information text encoder to obtain a plurality of financial analysis feature vectors; and the financial information cascading unit is used for cascading the plurality of financial analysis feature vectors to obtain the enterprise financial analysis feature vectors.
In the above-mentioned intelligent visual management system for enterprise report, the technical innovation information coding module includes: the invention creation information word segmentation unit is used for carrying out word segmentation processing on the enterprise intellectual property related information to obtain a plurality of invention creation information related words; an invention creation information word embedding unit, configured to pass the plurality of invention creation information related words through a word embedding layer to convert each of the plurality of invention creation information related words into an invention creation information related word embedding vector to obtain a sequence of the invention creation information related word embedding vector, where the embedding layer uses a learnable embedding matrix to perform embedded encoding on each of the invention creation information related words; the invention creation information context semantic coding unit is used for inputting a sequence of the invention creation information related word embedded vectors into a converter module of the technical innovation information text encoder so as to obtain a plurality of invention creation analysis feature vectors; the invention creation information cascading unit is used for cascading the plurality of invention creation analysis feature vectors to obtain the enterprise technical innovation feature vector.
In the above-mentioned intelligent visual management system for enterprise report, the enterprise valuation feature generation module includes: the enterprise comprehensive information feature matrix pooling unit is used for pooling the enterprise comprehensive information feature matrix along the horizontal direction and the vertical direction respectively to obtain a first enterprise comprehensive information feature pooling vector and a second enterprise comprehensive information feature pooling vector; the bidirectional feature association coding unit is used for carrying out association coding on the first enterprise comprehensive information feature pooling vector and the second enterprise comprehensive information feature pooling vector to obtain a bidirectional association matrix; the bidirectional association weight matrix generation unit is used for inputting the bidirectional association matrix into a Sigmoid activation function to obtain a bidirectional association weight matrix; and the matrix weighting calculation unit is used for calculating the point-by-point multiplication between the bidirectional association weight matrix and the enterprise comprehensive information feature matrix to obtain the enterprise estimation feature matrix.
In the above-mentioned intelligent visual management system for enterprise report, the optimization module includes: the probabilistic characteristic generating unit is used for inputting the enterprise estimated characteristic matrix into a Sigmoid function to obtain a probabilistic enterprise estimated characteristic matrix; and the optimized enterprise estimation feature generation unit is used for carrying out rigid consistency on the parameterized geometric relationship transition priori features on the probabilistic enterprise estimation feature matrix so as to obtain the optimized enterprise estimation feature matrix.
In the above-mentioned intelligent visual management system for enterprise report, the optimized enterprise valuation feature generating unit includes: carrying out rigid consistency of parameterized geometric relation transition priori features on the probabilistic enterprise estimation feature matrix by using the following optimization formula to obtain the optimized enterprise estimation feature matrix; wherein, the optimization formula is:
Wherein, Representing the first of the probabilistic enterprise valuation feature matricesThe characteristic value of the location is used to determine,Which represents a predetermined super-parameter that is to be used,The logarithmic function value is represented with a base of 2,Representing the first of the optimized enterprise valuation feature matricesCharacteristic values of the location.
According to another aspect of the present application, there is provided an intelligent visual management method for enterprise report, including:
Acquiring a human resource management report, a financial report and intellectual property related information of an enterprise to be evaluated;
Preprocessing the human resource management report of the enterprise to be evaluated to obtain a preprocessed human resource management report;
preprocessing the financial report of the enterprise to be evaluated to obtain a preprocessed financial report;
Extracting relevant information of staff academic and staff research and development staff duty ratio of the enterprise from the preprocessed human resource management report;
Inputting the related information of the staff academic and the related information of the staff occupancy rate of the staff research and development staff of the enterprise into a human information text encoder to obtain a staff structure feature vector of the enterprise;
Extracting enterprise asset-related information, liability-related information, business income-related information and research and development investment-related information from the preprocessed financial report;
Inputting the enterprise asset-related information, the liability-related information, the business income-related information and the research and development investment-related information into a financial information text encoder to obtain an enterprise financial analysis feature vector;
inputting the related information of the enterprise intellectual property into a technical innovation information text encoder to obtain an enterprise technical innovation feature vector;
constructing the talent structure feature vector of the enterprise, the financial analysis feature vector of the enterprise and the technical innovation feature vector of the enterprise into an enterprise comprehensive information feature matrix;
Inputting the enterprise comprehensive information feature matrix to a feature extractor based on a bidirectional attention mechanism to obtain an enterprise estimation feature matrix;
Optimizing the enterprise estimation feature matrix to obtain an optimized enterprise estimation feature matrix;
and the optimized enterprise estimation characteristic matrix is passed through a generator to obtain a generation result, wherein the generation result is an enterprise estimation visualization table.
Compared with the prior art, the intelligent visual management system and the method for the enterprise report, provided by the application, adopt a text analysis technology based on artificial intelligence, respectively extract characteristics of human resource management reports of enterprises, financial reports of the enterprises and intellectual property related information of the enterprises to obtain talent structure characteristic vectors of the enterprises, financial analysis characteristic vectors of the enterprises and technical innovation characteristic vectors of the enterprises, then perform characteristic coding on characteristic matrixes formed by the characteristic vectors to obtain enterprise valuation characteristics, and finally input the enterprise valuation characteristics into a generator to obtain a visual enterprise valuation table.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing embodiments of the present application in more detail with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and together with the embodiments of the application, do not limit the application. In the drawings, like reference numerals generally refer to like parts or steps.
FIG. 1 is a system block diagram of an intelligent visualization management system for enterprise reporting in accordance with an embodiment of the present application.
FIG. 2 is an architecture diagram of an intelligent visualization management system for enterprise reporting, in accordance with an embodiment of the present application.
FIG. 3 is a block diagram of a talent structural feature coding module in an intelligent visual management system for enterprise reporting in accordance with an embodiment of the present application.
FIG. 4 is a block diagram of a financial information encoding module in an intelligent visualization management system for enterprise reporting, in accordance with an embodiment of the present application.
FIG. 5 is a block diagram of a technical innovation coding module in an intelligent visual management system for enterprise reporting in accordance with an embodiment of the present application.
FIG. 6 is a block diagram of an enterprise valuation feature generation module in an intelligent visualization management system for enterprise reporting, in accordance with an embodiment of the present application.
FIG. 7 is a block diagram of an optimization module in an intelligent visualization management system for enterprise reporting, in accordance with an embodiment of the present application.
FIG. 8 is a flow chart of a method for intelligent visual management of enterprise reports in accordance with an embodiment of the application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more apparent, the embodiments of the present application will be described in further detail with reference to the accompanying drawings. These embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the concept of the example embodiments to those skilled in the art. The exemplary embodiments of the present application and their descriptions herein are for the purpose of explaining the present application, but are not to be construed as limiting the application.
Furthermore, the described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the application. One skilled in the relevant art will recognize, however, that the application can be practiced without one or more of the specific details, or with other methods, components, devices, steps, etc. In other instances, well-known methods, devices, implementations, or operations are not shown or described in detail to avoid obscuring aspects of the application.
The block diagrams shown in the figures are merely functional entities and do not necessarily correspond to physically separate entities. That is, the functional entities may be implemented in software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor devices and/or microcontroller devices.
The flow diagrams depicted in the figures are exemplary only, and do not necessarily include all of the elements and operations/steps, nor must they be performed in the order described. For example, some operations/steps may be decomposed, and some operations/steps may be combined or partially combined, so that the order of actual execution may be changed according to actual situations.
It should be understood that although the terms first, second, third, etc. may be used in the present application to describe various components, these components should not be limited by these terms. These terms are used to distinguish one element from another element. Accordingly, a first component discussed below could be termed a second component without departing from the teachings of the present inventive concept. As used in this disclosure, the term "and/or" and similar terms include all combinations of any, many, and all of the associated listed items.
FIG. 1 is a system block diagram of an intelligent visualization management system for enterprise reporting in accordance with an embodiment of the present application. FIG. 2 is an architecture diagram of an intelligent visualization management system for enterprise reporting, in accordance with an embodiment of the present application. As shown in fig. 1 and 2, in an intelligent visualization management system 100 for enterprise reporting, it includes: an enterprise report data acquisition module 110 for acquiring human resource management reports, financial reports, and intellectual property related information of an enterprise to be evaluated; a human resource report preprocessing module 120, configured to preprocess the human resource management report of the enterprise to be evaluated to obtain a preprocessed human resource management report; a financial report preprocessing module 130, configured to preprocess a financial report of the enterprise to be evaluated to obtain a preprocessed financial report; a talent structure information extraction module 140, configured to extract relevant information of staff learning of the enterprise from the preprocessed human resource management report, and relevant information of staff occupancy of the staff research and development of the enterprise; the talent structure feature coding module 150 is configured to input the related information of the staff learning of the enterprise and the related information of the duty ratio of the staff research and development personnel of the enterprise into a text encoder of human information to obtain a talent structure feature vector of the enterprise; a financial related information extraction module 160 for extracting enterprise asset related information, liability related information, business income related information, and research and development investment related information from the preprocessed financial report; a financial information encoding module 170, configured to input the enterprise asset related information, the liability related information, the business income related information, and the research and development investment related information into a financial information text encoder to obtain an enterprise financial analysis feature vector; a technical innovation information encoding module 180, configured to input the information related to the intellectual property of the enterprise into a technical innovation information text encoder to obtain an enterprise technical innovation feature vector; the comprehensive information feature construction module 190 is configured to construct the talent structure feature vector of the enterprise, the financial analysis feature vector of the enterprise, and the technical innovation feature vector of the enterprise into an enterprise comprehensive information feature matrix; the enterprise estimation feature generation module 200 is configured to input the enterprise comprehensive information feature matrix to a feature extractor based on a bidirectional attention mechanism to obtain an enterprise estimation feature matrix; an optimization module 210, configured to optimize the enterprise estimation feature matrix to obtain an optimized enterprise estimation feature matrix; and the enterprise valuation visualization result generating module 220 is configured to pass the optimized enterprise valuation feature matrix through a generator to obtain a generating result, where the generating result is an enterprise valuation visualization table.
Specifically, in the technical scheme of the application, firstly, a human resource management report, a financial report and intellectual property related information of an enterprise to be evaluated are obtained. It should be appreciated that by obtaining the human resource management report, information such as employee's academic, developer's duty, etc. of the enterprise may be known, which reflects talent quality and innovation ability of the enterprise. In the evaluation process, the technical innovation capability of small and medium-sized scientific enterprises is an important consideration, so that the human resource management report can provide talent structural characteristics of related enterprises and provide important references for evaluation. Financial reporting is an important basis for assessing corporate financial status and business performance. While small and medium-sized scientific enterprises may lack stable profitability patterns and reliable financial data, financial reports still provide critical information for the enterprises, such as assets, liabilities, business revenues, and research and development investment. Financial analysis features associated with valuations, such as business asset-related information, liability-related information, business revenue-related information, and research and development investment-related information, may be extracted from the financial reports by acquiring and processing the financial reports. These features can provide a financial reference for the valuation model. Intellectual property is an important asset for small and medium-sized enterprises in science and technology, including patents, trademarks, copyrights and the like. Acquiring intellectual property related information of an enterprise may reflect the technical innovation ability and market competitiveness of the enterprise. The human resource management report of the enterprise to be evaluated can be obtained from a human resource department of the enterprise, the financial report of the enterprise to be evaluated can be obtained from a financial department of the enterprise, and the intellectual property related information of the enterprise to be evaluated can be obtained from related information in the enterprise to know the intellectual property condition and strategy of the enterprise.
Human resource management reports typically contain large amounts of text information, which may be subject to format inconsistencies, wrongly written words, missing data, etc. Through preprocessing, the text can be cleaned and standardized, unnecessary symbols, spaces and special characters are removed, spelling errors are repaired, the format is unified, and the accuracy and consistency of data are ensured. That is, the human resource management report of the enterprise to be evaluated is preprocessed to obtain a preprocessed human resource management report.
In the embodiment of the present application, an implementation manner of preprocessing the human resource management report of the enterprise to be evaluated to obtain the preprocessed human resource management report may be: removing punctuation marks, spaces and other special characters in the text using character string processing functions, such as replace () or regular expressions; spell errors may be repaired using a spell check library or custom spell correction functions, for example, spell check and correction using the PYSPELLCHECKER library in Python; the normalization process may be performed on the text, for example, different academic representations may be unified, a dictionary or rule may be used to perform the normalization operation, for example, a dictionary including various academic representations may be defined, and a get () method may be used to find and return the normalized academic representations, and if no matching academic representations are found, the original text may be retained.
Similarly, financial reports often contain large amounts of text, including important information about the financial status of the business, business operations, and risk factors. Financial text data may also have non-uniform formats, missing values, misspellings, etc. The preprocessing can clean the data, remove unnecessary characters, punctuation marks and special symbols, repair spelling errors, and normalize the data to conform to a uniform format. This helps to improve the accuracy and consistency of the data and reduces errors in subsequent analysis. That is, the financial reports of the enterprise under evaluation are preprocessed to obtain preprocessed financial reports.
If related information of staff is not extracted, the educational background and professional skills of staff of the enterprise cannot be comprehensively known, and the quality and capability of human resources of the enterprise cannot be accurately estimated. Also, if the relevant information of the ratio of staff and research personnel of the enterprise is not extracted, the research and development power and innovation capability of the enterprise cannot be accurately evaluated, and reliable data support cannot be provided for investors and decision makers. Therefore, in the technical scheme of the application, the related information of the staff academic and the related information of the staff research and development staff duty ratio of the enterprise are extracted from the preprocessed human resource management report.
In this embodiment, one implementation method for extracting relevant information of staff academic and relevant information of staff research and development staff duty ratio from the preprocessed human resource management report includes: a. extracting the related information of the academic: the NLP technology is used for entity identification, and entities related to the academic in the report, such as the academic degree, the school, the specialty and the like are identified. Sentences or paragraphs related to the academy are identified and extracted using keyword matching or rule matching methods. For example, sentences including keywords (e.g., "bachelor's degree", "scholar's study", "doctor's degree", etc.) may be used for extraction. The extracted academic information may include the type of academic degree, school name, specialty, etc. b. Extracting relevant information of the occupancy ratio of the research personnel: entity recognition is performed using NLP technology to identify entities in the report that are related to employee positions and departments, such as "research and development department", "technical team", and so forth. Sentences or paragraphs related to the developer's duty ratio are identified and extracted using keyword matching or rule matching methods. For example, sentences containing keywords (e.g., "developer", "technical team duty", etc.) may be used for extraction. The extracted developer duty cycle information may include the number of developers, the total number of employees, and the calculated developer duty cycle.
The related information of the staff academic and the related information of the staff research and development staff duty ratio of the staff are text information, which needs to be converted into a feature vector which can be processed by a machine, and related features of the staff structure of the enterprise are extracted from the feature vector. And therefore, inputting the related information of the staff academic and the related information of the staff research and development staff duty ratio of the enterprise staff into a human information text encoder to obtain the staff structure feature vector. The structure of the human-information text encoder is a converter-based context encoder that includes a word embedding layer. The word embedding layer is part of a human information text encoder that maps each word into a continuous vector space. This has the advantage that the text information can be converted into a numerical representation of the character, so that the computer can better understand and process the information. The Word embedding layer may obtain a vector representation of the Word through a pre-trained Word vector model (e.g., word2Vec, gloVe, etc.). The converter is a powerful deep learning model and is widely applied to natural language processing tasks. It is capable of processing sequence data, such as text, and it implements feature extraction and encoding through self-attention mechanisms and multi-layer perceptrons.
FIG. 3 is a block diagram of a talent structural feature coding module in an intelligent visual management system for enterprise reporting in accordance with an embodiment of the present application. As shown in fig. 3, the talent structural feature coding module 150 includes: a human information word segmentation unit 151, configured to perform word segmentation processing on the related information of the staff learning of the enterprise and the related information of the duty ratio of the staff research and development of the enterprise to obtain a plurality of talent information related words; a human information word embedding unit 152, configured to pass the plurality of talent information related words through a word embedding layer to transform each of the plurality of talent information related words into a talent information related word embedding vector to obtain a sequence of talent information related word embedding vectors, where the embedding layer uses a learnable embedding matrix to perform embedded encoding on each of the talent information related words; a human information context semantic coding unit 153, configured to input a sequence of the talent information related words embedded vectors into a converter module of the human information text encoder to obtain a plurality of human resource structural feature vectors; and the human information cascading unit 154 is configured to cascade the plurality of human resource structural feature vectors to obtain the talent structural feature vector of the enterprise.
Assets are resources and rights owned by an enterprise, including cash, equipment and real estate, and the size, composition and value of the assets of the enterprise can be known by extracting the related information of the assets of the enterprise from the financial report, so that the financial strength and potential value of the enterprise can be evaluated; liability is an economic obligation borne by an enterprise for external creditors, including borrowing, accounts payable, interest payable and the like, and the liability scale, the liability repayment capacity and the financial risk of the enterprise can be known by extracting liability related information, so that the financial stability of the enterprise is evaluated; business income is income obtained by enterprises from business activities, including sales income, service income and the like, and the income source, scale and growth trend of the enterprises can be known by extracting the relevant information of the business income, thereby helping to evaluate the profitability and market competitiveness of the enterprises; the research and development investment is the resources and funds invested by enterprises in research and development activities, including research and development expenditure, the number of research and development personnel and the like, and the innovation capability and the technical strength of the enterprises can be known by extracting the related information of the research and development investment, so that the long-term competitiveness of the enterprises is evaluated. Therefore, in the technical scheme of the application, the enterprise asset related information, the liability related information, the business income related information and the research and development investment related information are extracted from the preprocessed financial report. By extracting the information, the financial condition of the enterprise can be quantified and analyzed, and an important basis is provided for enterprise valuation.
Financial information plays an important role in enterprise valuation, and through analysis of financial data, conditions in aspects such as financial condition, profitability and financial stability of enterprises can be known. The purpose of the feature analysis of financial information is to extract key features from it that relate to the value and potential of the business. Thus, the business asset-related information, the liability-related information, the business revenue-related information, and the research and development investment-related information are input into a financial information text encoder to obtain a business financial analysis feature vector. The structure of the financial information text encoder is here a converter-based text encoder comprising a word embedding layer. By carrying out feature analysis on the financial information, complex financial data can be converted into feature vectors with more representativeness and interpretability, and subsequent data processing is facilitated.
FIG. 4 is a block diagram of a financial information encoding module in an intelligent visualization management system for enterprise reporting, in accordance with an embodiment of the present application. As shown in fig. 4, the financial information encoding module 170 includes: a financial information word segmentation unit 171 for performing word segmentation processing on the enterprise asset-related information, the liability-related information, the business income-related information, and the research and development investment-related information to obtain a plurality of financial information-related words; a financial information word embedding unit 172, configured to pass the plurality of financial information related words through a word embedding layer to convert each of the plurality of financial information related words into a financial information related word embedding vector to obtain a sequence of financial information related word embedding vectors, where the embedding layer uses a learnable embedding matrix to perform embedding encoding on each of the financial information related words; a financial information context semantic coding unit 173 for inputting a sequence of the financial information related words embedded vectors into a converter module of the financial information text encoder to obtain a plurality of financial analysis feature vectors; and a financial information cascading unit 174, configured to cascade the plurality of financial analysis feature vectors to obtain the enterprise financial analysis feature vector.
In the evaluation of small and medium-sized enterprises in the scientific and technological type, the technical innovation capability is an important consideration. By extracting features of intellectual property related information, the strength, potential and competitive advantage of enterprises in technical innovation can be evaluated. That is, the business intellectual property related information is input into a technical innovation information text encoder to obtain a business technical innovation feature vector. The structure of the technical innovation text encoder is a converter-based context encoder that includes an embedded layer. Through the technical innovation information text encoder, the characteristic extraction can be carried out on the intellectual property related information of enterprises. These characteristics may include the number of patents in the technical field, the quality and innovation of the patents, the trend of technology development, the strength of the development team, etc. By analyzing the characteristics, the capability and achievement of enterprises in technical innovation can be evaluated, and the advantages and potential of the enterprises in market competition can be judged. This allows the technological innovation capabilities of the enterprise to be quantified and measured. By extracting key features and converting them into feature vectors, complex intellectual property information can be converted into more representative and interpretable data forms.
FIG. 5 is a block diagram of a technical innovation coding module in an intelligent visual management system for enterprise reporting in accordance with an embodiment of the present application. As shown in fig. 5, the technical innovation information encoding module 180 includes: an application creation information word segmentation unit 181, configured to perform word segmentation processing on the enterprise intellectual property related information to obtain a plurality of application creation information related words; an application creation information word embedding unit 182, configured to pass the plurality of application creation information related words through a word embedding layer to convert each of the plurality of application creation information related words into an application creation information related word embedding vector to obtain a sequence of application creation information related word embedding vectors, where the embedding layer uses a learnable embedding matrix to perform embedding encoding on each of the application creation information related words; an application creation information context semantic coding unit 183 for inputting a sequence of the application creation information related words embedded vectors into a converter module of the technical innovation information text encoder to obtain a plurality of application creation analysis feature vectors; the application creation information concatenating unit 184 is configured to concatenate the plurality of application creation analysis feature vectors to obtain the enterprise technical innovation feature vector.
The value and the potential of an enterprise are determined by factors of various aspects, including talent structure, financial condition, technical innovation capability and the like, so that in the technical scheme of the application, the enterprise talent structure feature vector, the enterprise financial analysis feature vector and the enterprise technical innovation feature vector are constructed into an enterprise comprehensive information feature matrix. In this way, key features of different aspects can be integrated together to form a comprehensive data structure describing enterprise value and potential.
In this embodiment, constructing the talent structure feature vector of the enterprise, the financial analysis feature vector of the enterprise, and the technical innovation feature vector of the enterprise as an enterprise comprehensive information feature matrix may be: collecting data comprising enterprise talent structure feature vectors, enterprise financial analysis feature vectors and enterprise technical innovation feature vectors, ensuring the accuracy and the integrity of the data, and aligning according to the same time period (such as the year or the quarter); carrying out normalization processing on each feature vector to eliminate dimension differences among different features, wherein a common normalization method comprises minimum-maximum normalization or normalization; combining the normalized talent structure feature vector of the enterprise, the enterprise financial analysis feature vector and the enterprise technical innovation feature vector according to the same time period to ensure that the sequence of each feature vector is consistent with the time label; checking whether missing values exist in the combined feature matrix, and if so, selecting a proper method for processing, such as interpolation or deleting rows with more missing values; according to actual demands and business knowledge, weight is distributed to each feature vector, the weight can be determined according to the importance of the feature and the influence on the performance of enterprises, and the sum of the weights is ensured to be equal to 1; multiplying each feature vector by a corresponding weight, and weighting and summing the feature vectors to obtain a comprehensive feature vector, wherein the feature vector represents comprehensive information features of enterprises; and arranging the comprehensive feature vectors in time sequence to construct an enterprise comprehensive information feature matrix, wherein each row represents a time point and each column represents a feature. The dimension of the feature matrix is ensured to be consistent with the time period of the data.
The enterprise comprehensive information feature matrix contains a comprehensive representation of a plurality of features, but the features therein may have different importance and interrelationships. Thus, the enterprise comprehensive information feature matrix is input to a bi-directional attention mechanism based feature extractor to obtain an enterprise valuation feature matrix. The feature extractor based on the bidirectional attention mechanism can automatically identify and emphasize the features with important influence on enterprise estimation according to the input enterprise comprehensive information feature matrix. This helps reduce redundant information and noise and extract features that are most relevant and meaningful to the estimation decision.
FIG. 6 is a block diagram of an enterprise valuation feature generation module in an intelligent visualization management system for enterprise reporting, in accordance with an embodiment of the present application. As shown in fig. 6, the enterprise valuation feature generation module 200 includes: an enterprise comprehensive information feature matrix pooling unit 201, configured to pool the enterprise comprehensive information feature matrix along a horizontal direction and a vertical direction respectively to obtain a first enterprise comprehensive information feature pooling vector and a second enterprise comprehensive information feature pooling vector; a bidirectional feature association encoding unit 202, configured to perform association encoding on the first enterprise comprehensive information feature pooling vector and the second enterprise comprehensive information feature pooling vector to obtain a bidirectional association matrix; a bi-directional correlation weight matrix generating unit 203, configured to input the bi-directional correlation matrix into a Sigmoid activation function to obtain a bi-directional correlation weight matrix; and the matrix weight calculation unit 204 is configured to calculate the enterprise estimation feature matrix by multiplying the bi-directional association weight matrix and the enterprise comprehensive information feature matrix according to location points.
In particular, in the technical solution of the present application, it is considered that the feature values in the enterprise valuation feature matrix may come from different types of features, such as human resource management, financial reporting, and intellectual property related information. The range and distribution of values of these features may have discontinuities. For example, asset values in an enterprise's financial reports may vary from millions to billions, while employee learning information may include discrete learning categories. Such discontinuities may lead to a lack of constraint on the probability distribution of the feature values. At the same time, different types of features may have different dimensions, such as monetary amount in financial reports and number of employees in human resources management reports. Such dimensional differences may lead to a lack of constraints on the probability distribution of feature values, because the range of values and the magnitude of the variation of different features are different, and it is difficult to unify them into one probability distribution. Furthermore, different types of features may have different distribution shapes, for example, asset values in a financial report may exhibit a biased distribution, while employee learning information may exhibit a multimodal distribution. Such a difference in distribution shape causes a lack of constraint on the probability distribution of feature values, because the distribution shapes of different features may not be uniformly constrained to a particular probability distribution form.
That is, the probability distribution of feature values in the enterprise valuation feature matrix may lack constraints due to feature discontinuities. This lack of constraints can lead to difficulties in the model in making enterprise estimates, as the probability distribution of feature values does not provide explicit information, which can affect the accuracy and stability of the model to enterprise value. In order to solve the problem, in the technical scheme of the application, the rigidity consistency of the parameterized geometric relation transition priori features is carried out on the enterprise estimation feature matrix so as to carry out bidirectional constraint on the information entropy dimension on each feature value sample in the enterprise estimation feature matrix.
FIG. 7 is a block diagram of an optimization module in an intelligent visualization management system for enterprise reporting, in accordance with an embodiment of the present application. As shown in fig. 7, the optimizing module 210 includes: a probabilistic feature generating unit 211, configured to input the enterprise estimation feature matrix into a Sigmoid function to obtain a probabilistic enterprise estimation feature matrix; and the optimized enterprise estimation feature generation unit 212 is configured to perform rigid equalization on the parameterized geometric relationship transition prior feature on the probabilistic enterprise estimation feature matrix to obtain the optimized enterprise estimation feature matrix.
Specifically, the optimized enterprise valuation feature generation unit 212 includes: carrying out rigid consistency of parameterized geometric relation transition priori features on the probabilistic enterprise estimation feature matrix by using the following optimization formula to obtain the optimized enterprise estimation feature matrix; wherein, the optimization formula is:
Wherein, Representing the first of the probabilistic enterprise valuation feature matricesThe characteristic value of the location is used to determine,Which represents a predetermined super-parameter that is to be used,The logarithmic function value is represented with a base of 2,Representing the first of the optimized enterprise valuation feature matricesCharacteristic values of the location.
It should be appreciated that a transitional prior feature is first created by parameterizing the geometric relationships of the eigenvalue samples at each location in the enterprise valuation feature matrix, which feature may reflect the distribution and variation of the eigenvalue samples among the different categories. A rigid reconciliation mechanism may then be used to adjust and optimize the transitional prior feature to better conform to the underlying structure and logic of the data. Therefore, the two-way constraint on the information entropy dimension can be carried out on the characteristic value samples, namely, the high entropy of the characteristic value samples in the categories is maintained, and the low entropy of the characteristic value samples in the categories is maintained, so that the distinguishing capability and the expression capability of the characteristic value samples are improved.
And finally, the optimized enterprise estimation feature matrix is passed through a generator to obtain a generation result, wherein the generation result is an enterprise estimation visualization table. By generating enterprise valuation visualization tables, abstract data can be converted into a visual chart or table format, making it easier for investors to understand and interpret the enterprise situation. The visual table can provide visual display to help quickly acquire key information, so that decision making and service management are supported.
In summary, an intelligent visual management system 100 for enterprise report according to the embodiment of the present application is illustrated, which adopts a text analysis technology based on artificial intelligence, and performs feature extraction on human resource management report of an enterprise, financial report of an enterprise and intellectual property related information of an enterprise to obtain talent structure feature vectors of the enterprise, financial analysis feature vectors of the enterprise and technical innovation feature vectors of the enterprise, then performs feature encoding on feature matrices formed by the feature vectors to obtain enterprise valuation features, and finally inputs the enterprise valuation features into a generator to obtain a visual enterprise valuation table.
As described above, an intelligent visual management system 100 for an enterprise report according to an embodiment of the present application may be implemented in various terminal devices, for example, a server for intelligent visual management of an enterprise report, etc. In one example, an intelligent visualization management system 100 for enterprise reporting in accordance with embodiments of the present application can be integrated into a terminal device as a software module and/or hardware module. For example, the intelligent visualization management system 100 of an enterprise report may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the intelligent visualization management system 100 of the enterprise report can also be one of a plurality of hardware modules of the terminal device.
Alternatively, in another example, the one type of enterprise reporting intelligent visual management system 100 and the terminal device may be separate devices, and the one type of enterprise reporting intelligent visual management system 100 may be connected to the terminal device through a wired and/or wireless network and transmit the interactive information in a agreed data format.
FIG. 8 is a flow chart of a method for intelligent visual management of enterprise reports in accordance with an embodiment of the application. As shown in fig. 8, in an intelligent visual management method for enterprise report, the method includes: s110, acquiring a human resource management report, a financial report and intellectual property related information of an enterprise to be evaluated; s120, preprocessing the human resource management report of the enterprise to be evaluated to obtain a preprocessed human resource management report; s130, preprocessing the financial report of the enterprise to be evaluated to obtain a preprocessed financial report; s140, extracting relevant information of staff academic and relevant information of staff research and development staff duty ratio of the enterprise from the preprocessed human resource management report; s150, inputting the related information of the staff academic and the related information of the staff ratio of the staff research and development staff of the enterprise into a human information text encoder to obtain a talent structure feature vector of the enterprise; s160, extracting enterprise asset related information, liability related information, business income related information and research and development investment related information from the preprocessed financial report; s170, inputting the enterprise asset-related information, the liability-related information, the business income-related information and the research and development investment-related information into a financial information text encoder to obtain an enterprise financial analysis feature vector; s180, inputting the related information of the enterprise intellectual property into a technical innovation information text encoder to obtain an enterprise technical innovation feature vector; s190, constructing the talent structure feature vector of the enterprise, the financial analysis feature vector of the enterprise and the technical innovation feature vector of the enterprise into an enterprise comprehensive information feature matrix; s200, inputting the enterprise comprehensive information feature matrix into a feature extractor based on a bidirectional attention mechanism to obtain an enterprise estimation feature matrix; s210, optimizing the enterprise estimation feature matrix to obtain an optimized enterprise estimation feature matrix; and S220, the optimized enterprise estimation feature matrix is passed through a generator to obtain a generation result, wherein the generation result is an enterprise estimation visualization table.
Here, it will be understood by those skilled in the art that the specific operations of the respective steps in the above-described method for intelligent visual management of an enterprise report have been described in detail in the above description of an intelligent visual management system of an enterprise report with reference to fig. 1 to 7, and thus, repetitive descriptions thereof will be omitted.
In summary, an intelligent visual management method for enterprise report according to the embodiments of the present application is illustrated, which adopts a text analysis technology based on artificial intelligence, and performs feature extraction on human resource management report of an enterprise, financial report of an enterprise, and intellectual property related information of an enterprise to obtain talent structure feature vectors of the enterprise, financial analysis feature vectors of the enterprise, and technical innovation feature vectors of the enterprise, then performs feature encoding on feature matrices formed by the feature vectors to obtain enterprise valuation features, and finally inputs the enterprise valuation features into a generator to obtain a visual enterprise valuation table.
It will be evident to those skilled in the art that the application is not limited to the details of the foregoing illustrative embodiments, and that the present application may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. The scope of the application is indicated by the appended claims rather than by the foregoing description, and all changes that come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Finally, it should be noted that the above-mentioned embodiments are merely for illustrating the technical solution of the present application and not for limiting the same, and although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made to the technical solution of the present application without departing from the spirit of the technical solution of the present application.

Claims (6)

1. An intelligent visualization management system for enterprise reporting, comprising:
The enterprise report data acquisition module is used for acquiring human resource management reports, financial reports and intellectual property related information of the enterprise to be evaluated;
The human resource report preprocessing module is used for preprocessing the human resource management report of the enterprise to be evaluated to obtain a preprocessed human resource management report;
The financial report preprocessing module is used for preprocessing the financial report of the enterprise to be evaluated to obtain a preprocessed financial report;
the talent structure information extraction module is used for extracting relevant information of staff academic and relevant information of staff research and development staff duty ratio of the enterprise from the preprocessed human resource management report;
The talent structure feature coding module is used for inputting the related information of the staff academic and the related information of the staff research and development staff duty ratio of the staff to a human information text coder to obtain a talent structure feature vector of the enterprise;
The financial related information extraction module is used for extracting enterprise asset related information, liability related information, business income related information and research and development investment related information from the preprocessed financial report;
the financial information encoding module is used for inputting the enterprise asset related information, the liability related information, the business income related information and the research and development investment related information into a financial information text encoder to obtain enterprise financial analysis feature vectors;
The technical innovation information coding module is used for inputting the related information of the intellectual property rights of the enterprises into the technical innovation information text coder to obtain the technical innovation feature vectors of the enterprises;
the comprehensive information feature construction module is used for constructing the enterprise talent structure feature vector, the enterprise financial analysis feature vector and the enterprise technical innovation feature vector into an enterprise comprehensive information feature matrix;
The enterprise estimation feature generation module is used for inputting the enterprise comprehensive information feature matrix into a feature extractor based on a bidirectional attention mechanism to obtain an enterprise estimation feature matrix;
The optimization module is used for optimizing the enterprise estimation characteristic matrix to obtain an optimized enterprise estimation characteristic matrix;
The enterprise valuation visualization result generation module is used for enabling the optimized enterprise valuation feature matrix to pass through a generator to obtain a generation result, wherein the generation result is an enterprise valuation visualization table;
wherein, talent structural feature coding module includes:
the human information word segmentation unit is used for carrying out word segmentation processing on the related information of the staff academic and research staff and the related information of the staff research and development staff of the enterprise so as to obtain a plurality of talent information related words;
The human information word embedding unit is used for converting each talent information related word in the plurality of talent information related words into a talent information related word embedding vector through a word embedding layer to obtain a sequence of talent information related word embedding vectors, wherein the embedding layer uses a learnable embedding matrix to carry out embedded coding on each talent information related word;
The human information context semantic coding unit is used for inputting the sequence of the talent information related words embedded vector into a converter module of the human information text encoder so as to obtain a plurality of human resource structural feature vectors;
the human information cascading unit is used for cascading the plurality of human resource structural feature vectors to obtain the enterprise talent structural feature vector;
Wherein, the optimization module includes:
The probabilistic characteristic generating unit is used for inputting the enterprise estimated characteristic matrix into a Sigmoid function to obtain a probabilistic enterprise estimated characteristic matrix;
the optimized enterprise estimation feature generation unit is used for carrying out rigidity consistency on parameterized geometric relationship transition priori features on the probabilistic enterprise estimation feature matrix to obtain the optimized enterprise estimation feature matrix;
Wherein, the optimizing enterprise valuation characteristic generating unit includes: carrying out rigid consistency of parameterized geometric relation transition priori features on the probabilistic enterprise estimation feature matrix by using the following optimization formula to obtain the optimized enterprise estimation feature matrix;
Wherein, the optimization formula is:
Wherein/> Representing the/>, in the probabilistic enterprise valuation feature matrixCharacteristic value of location,/>Representing a predetermined hyper-parameter,/>Represents a logarithmic function value based on 2,/>Representing the/>, in the optimized enterprise valuation feature matrixA characteristic value of the location;
In the optimization formula, a transition priori feature is established by parameterizing the geometric relationship of the eigenvalue samples at each position in the enterprise valuation feature matrix, the feature can reflect the distribution and change of the eigenvalue samples among different categories, and then a rigid unification mechanism can be used for adjusting and optimizing the transition priori feature to enable the transition priori feature to conform to the internal structure and logic of data better, so that bidirectional constraint on the information entropy dimension can be carried out on the eigenvalue samples, namely, high entropy of the eigenvalue samples in the categories is maintained, and low entropy of the eigenvalue samples among the categories is maintained, thereby improving the distinguishing capability and the expression capability of the eigenvalue samples.
2. The intelligent visualization management system for enterprise reporting of claim 1, wherein the financial information encoding module comprises:
The financial information word segmentation unit is used for word segmentation processing of the enterprise asset related information, the liability related information, the business income related information and the research and development investment related information to obtain a plurality of financial information related words;
a financial information word embedding unit, configured to pass the plurality of financial information related words through a word embedding layer to convert each financial information related word in the plurality of financial information related words into a financial information related word embedding vector to obtain a sequence of financial information related word embedding vectors, where the embedding layer uses a learnable embedding matrix to perform embedding encoding on each financial information related word;
A financial information context semantic coding unit for inputting a sequence of the financial information related word embedded vectors into a converter module of the financial information text encoder to obtain a plurality of financial analysis feature vectors;
And the financial information cascading unit is used for cascading the plurality of financial analysis feature vectors to obtain the enterprise financial analysis feature vectors.
3. The intelligent visualization management system for enterprise reporting of claim 2, wherein the technical innovation information encoding module comprises:
the invention creation information word segmentation unit is used for carrying out word segmentation processing on the enterprise intellectual property related information to obtain a plurality of invention creation information related words;
An invention creation information word embedding unit, configured to pass the plurality of invention creation information related words through a word embedding layer to convert each of the plurality of invention creation information related words into an invention creation information related word embedding vector to obtain a sequence of the invention creation information related word embedding vector, where the embedding layer uses a learnable embedding matrix to perform embedded encoding on each of the invention creation information related words;
The invention creation information context semantic coding unit is used for inputting a sequence of the invention creation information related word embedded vectors into a converter module of the technical innovation information text encoder so as to obtain a plurality of invention creation analysis feature vectors;
The invention creation information cascading unit is used for cascading the plurality of invention creation analysis feature vectors to obtain the enterprise technical innovation feature vector.
4. The intelligent visualization management system for enterprise reporting of claim 3, wherein the enterprise valuation characteristics generation module comprises:
The enterprise comprehensive information feature matrix pooling unit is used for pooling the enterprise comprehensive information feature matrix along the horizontal direction and the vertical direction respectively to obtain a first enterprise comprehensive information feature pooling vector and a second enterprise comprehensive information feature pooling vector;
The bidirectional feature association coding unit is used for carrying out association coding on the first enterprise comprehensive information feature pooling vector and the second enterprise comprehensive information feature pooling vector to obtain a bidirectional association matrix;
the bidirectional association weight matrix generation unit is used for inputting the bidirectional association matrix into a Sigmoid activation function to obtain a bidirectional association weight matrix;
And the matrix weighting calculation unit is used for calculating the point-by-point multiplication between the bidirectional association weight matrix and the enterprise comprehensive information feature matrix to obtain the enterprise estimation feature matrix.
5. An intelligent visual management method for enterprise report, comprising:
Acquiring a human resource management report, a financial report and intellectual property related information of an enterprise to be evaluated;
Preprocessing the human resource management report of the enterprise to be evaluated to obtain a preprocessed human resource management report;
preprocessing the financial report of the enterprise to be evaluated to obtain a preprocessed financial report;
Extracting relevant information of staff academic and staff research and development staff duty ratio of the enterprise from the preprocessed human resource management report;
Inputting the related information of the staff academic and the related information of the staff occupancy rate of the staff research and development staff of the enterprise into a human information text encoder to obtain a staff structure feature vector of the enterprise;
Extracting enterprise asset-related information, liability-related information, business income-related information and research and development investment-related information from the preprocessed financial report;
Inputting the enterprise asset-related information, the liability-related information, the business income-related information and the research and development investment-related information into a financial information text encoder to obtain an enterprise financial analysis feature vector;
inputting the related information of the intellectual property of the enterprise into a technical innovation information text encoder to obtain technical innovation feature vectors of the enterprise;
constructing the talent structure feature vector of the enterprise, the financial analysis feature vector of the enterprise and the technical innovation feature vector of the enterprise into an enterprise comprehensive information feature matrix;
Inputting the enterprise comprehensive information feature matrix to a feature extractor based on a bidirectional attention mechanism to obtain an enterprise estimation feature matrix;
Optimizing the enterprise estimation feature matrix to obtain an optimized enterprise estimation feature matrix;
the optimized enterprise estimation feature matrix is passed through a generator to obtain a generation result, wherein the generation result is an enterprise estimation visual table;
The method for inputting the related information of the staff academic and the related information of the staff duty ratio of the staff research and development staff of the enterprise into the human information text encoder to obtain the staff structure feature vector of the enterprise comprises the following steps:
Word segmentation processing is carried out on the related information of the staff academic and the related information of the staff research and development staff duty ratio of the enterprise so as to obtain a plurality of talent information related words;
The talent information related words are converted into talent information related word embedding vectors by a word embedding layer to obtain a sequence of talent information related word embedding vectors, wherein the embedding layer uses a learnable embedding matrix to carry out embedding coding on each talent information related word;
Inputting the sequence of the talent information related words embedded vectors into a converter module of the human information text encoder to obtain a plurality of human resource structure feature vectors;
Cascading the plurality of human resource structural feature vectors to obtain the enterprise talent structural feature vector;
Wherein optimizing the enterprise valuation feature matrix to obtain an optimized enterprise valuation feature matrix comprises:
Inputting the enterprise estimation feature matrix into a Sigmoid function to obtain a probabilistic enterprise estimation feature matrix;
Carrying out rigid consistency of parameterized geometric relation transition priori features on the probabilistic enterprise estimation feature matrix to obtain the optimized enterprise estimation feature matrix;
The method for performing rigid consistency of parameterized geometric relation transition prior features on the probabilistic enterprise estimation feature matrix to obtain the optimized enterprise estimation feature matrix comprises the following steps: carrying out rigid consistency of parameterized geometric relation transition priori features on the probabilistic enterprise estimation feature matrix by using the following optimization formula to obtain the optimized enterprise estimation feature matrix;
Wherein, the optimization formula is:
Wherein/> Representing the/>, in the probabilistic enterprise valuation feature matrixCharacteristic value of location,/>Representing a predetermined hyper-parameter,/>Represents a logarithmic function value based on 2,/>Representing the/>, in the optimized enterprise valuation feature matrixA characteristic value of the location;
In the optimization formula, a transition priori feature is established by parameterizing the geometric relationship of the eigenvalue samples at each position in the enterprise valuation feature matrix, the feature can reflect the distribution and change of the eigenvalue samples among different categories, and then a rigid unification mechanism can be used for adjusting and optimizing the transition priori feature to enable the transition priori feature to conform to the internal structure and logic of data better, so that bidirectional constraint on the information entropy dimension can be carried out on the eigenvalue samples, namely, high entropy of the eigenvalue samples in the categories is maintained, and low entropy of the eigenvalue samples among the categories is maintained, thereby improving the distinguishing capability and the expression capability of the eigenvalue samples.
6. The method of claim 5, wherein inputting the enterprise complex information feature matrix into a feature extractor based on a bi-directional attention mechanism to obtain an enterprise valuation feature matrix, comprises:
Pooling the enterprise comprehensive information feature matrix along the horizontal direction and the vertical direction respectively to obtain a first enterprise comprehensive information feature pooling vector and a second enterprise comprehensive information feature pooling vector;
Performing association coding on the first enterprise comprehensive information feature pooling vector and the second enterprise comprehensive information feature pooling vector to obtain a bidirectional association matrix;
Inputting the bidirectional association matrix into a Sigmoid activation function to obtain a bidirectional association weight matrix;
and calculating the point-by-point multiplication between the bidirectional association weight matrix and the enterprise comprehensive information feature matrix to obtain the enterprise estimation feature matrix.
CN202410058497.6A 2024-01-16 2024-01-16 Intelligent visual management system and method for enterprise report Active CN117575827B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410058497.6A CN117575827B (en) 2024-01-16 2024-01-16 Intelligent visual management system and method for enterprise report

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410058497.6A CN117575827B (en) 2024-01-16 2024-01-16 Intelligent visual management system and method for enterprise report

Publications (2)

Publication Number Publication Date
CN117575827A CN117575827A (en) 2024-02-20
CN117575827B true CN117575827B (en) 2024-05-03

Family

ID=89862834

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410058497.6A Active CN117575827B (en) 2024-01-16 2024-01-16 Intelligent visual management system and method for enterprise report

Country Status (1)

Country Link
CN (1) CN117575827B (en)

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109933703A (en) * 2019-03-14 2019-06-25 鹿寨知航科技信息服务有限公司 A kind of construction method of Intellectual Property Right of Enterprises appraisal Model
CN109934441A (en) * 2017-12-18 2019-06-25 北京心流慧估科技有限公司 Growing/adult form enterprise the estimation method of one kind and valuation system
CN110880051A (en) * 2019-11-12 2020-03-13 上海融贷通金融信息服务有限公司 Enterprise value evaluation cloud service system
CN112836939A (en) * 2021-01-13 2021-05-25 温州大学 On-line valuation method and system for entrepreneurship type enterprise
CN116049243A (en) * 2023-01-09 2023-05-02 中知数通(北京)信息技术有限公司 Enterprise intellectual property big data information analysis system, method and storage medium
CN116090915A (en) * 2023-03-03 2023-05-09 上海华鑫股份有限公司 Visual analysis method and system for enterprise market value analysis
CN117010697A (en) * 2023-09-25 2023-11-07 山东财经大学 Visual enterprise risk assessment method based on artificial intelligence
CN117057942A (en) * 2023-10-12 2023-11-14 之江实验室科技控股有限公司 Intelligent financial decision big data analysis system
CN117078054A (en) * 2023-06-07 2023-11-17 科学技术部火炬高技术产业开发中心 Scientific and technological enterprise innovation ability quantitative assessment method and system
CN117235630A (en) * 2023-11-15 2023-12-15 吉林大学 Intelligent disease area visual management system and method thereof
CN117333076A (en) * 2023-11-02 2024-01-02 北京心流慧估科技有限公司 Evaluation method, device, equipment and medium based on mixed expert model

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160162821A1 (en) * 2014-12-04 2016-06-09 International Business Machines Corporation Comparative peer analysis for business intelligence
EP3834166A4 (en) * 2018-07-27 2022-06-29 Rocky Mountain Innovations Insights LLC Cloud-based, data-driven artificial intelligence and machine learning financial planning and analysis visualization platform

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109934441A (en) * 2017-12-18 2019-06-25 北京心流慧估科技有限公司 Growing/adult form enterprise the estimation method of one kind and valuation system
CN109933703A (en) * 2019-03-14 2019-06-25 鹿寨知航科技信息服务有限公司 A kind of construction method of Intellectual Property Right of Enterprises appraisal Model
CN110880051A (en) * 2019-11-12 2020-03-13 上海融贷通金融信息服务有限公司 Enterprise value evaluation cloud service system
CN112836939A (en) * 2021-01-13 2021-05-25 温州大学 On-line valuation method and system for entrepreneurship type enterprise
CN116049243A (en) * 2023-01-09 2023-05-02 中知数通(北京)信息技术有限公司 Enterprise intellectual property big data information analysis system, method and storage medium
CN116090915A (en) * 2023-03-03 2023-05-09 上海华鑫股份有限公司 Visual analysis method and system for enterprise market value analysis
CN117078054A (en) * 2023-06-07 2023-11-17 科学技术部火炬高技术产业开发中心 Scientific and technological enterprise innovation ability quantitative assessment method and system
CN117010697A (en) * 2023-09-25 2023-11-07 山东财经大学 Visual enterprise risk assessment method based on artificial intelligence
CN117057942A (en) * 2023-10-12 2023-11-14 之江实验室科技控股有限公司 Intelligent financial decision big data analysis system
CN117333076A (en) * 2023-11-02 2024-01-02 北京心流慧估科技有限公司 Evaluation method, device, equipment and medium based on mixed expert model
CN117235630A (en) * 2023-11-15 2023-12-15 吉林大学 Intelligent disease area visual management system and method thereof

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Visual Display Method of Enterprise Internal Control Information Based on Cluster Analysis Algorithm;Ranhong Cui et.al;《WIRELESS COMMUNICATIONS & MOBILE COMPUTING》;20220918;第1-11页 *
基于直觉模糊网络分析法和二元语义的文化创意企业价值评估方法;周晓光;肖喻;何欣;;运筹与管理;20200125(01);第152-160页 *

Also Published As

Publication number Publication date
CN117575827A (en) 2024-02-20

Similar Documents

Publication Publication Date Title
Minh et al. Deep learning approach for short-term stock trends prediction based on two-stream gated recurrent unit network
JP4398916B2 (en) Probabilistic model generation apparatus and program
Ko et al. Prediction of corporate financial distress: An application of the composite rule induction system
CN108153729B (en) Knowledge extraction method for financial field
US11494850B1 (en) Applied artificial intelligence technology for detecting anomalies in payroll data
Popa et al. Composite financial performance index prediction–a neural networks approach
CN113255321A (en) Financial field chapter-level event extraction method based on article entity word dependency relationship
CN114637827A (en) Carbon transaction text event extraction method based on graph neural network
CN112308230A (en) Construction and application method of asset management full-life-cycle knowledge base
Wang et al. Is image encoding beneficial for deep learning in finance?
Muslim et al. An Ensemble Stacking Algorithm to Improve Model Accuracy in Bankruptcy Prediction
US11461616B2 (en) Method and system for analyzing documents
CN117575827B (en) Intelligent visual management system and method for enterprise report
Jeyaraman et al. Practical Machine Learning with R: Define, build, and evaluate machine learning models for real-world applications
Brabazon et al. Diagnosing corporate stability using grammatical evolution
CN111626331B (en) Automatic industry classification device and working method thereof
CN113159419A (en) Group feature portrait analysis method, device and equipment and readable storage medium
Li Enterprise Financing Risk Control of Machine Learning Combined with Blockchain Technology
Allouche et al. Structured dictionary learning of rating migration matrices for credit risk modeling
Li et al. Prediction and Sensitivity Analysis of Companies’ Return on Equity Based on Deep Neural Network
CN117172232B (en) Audit report generation method, audit report generation device, audit report generation equipment and audit report storage medium
Trigueiros et al. Discovering the optimal set of ratios to use in accounting-based models
CN117708759B (en) Method and device for positioning industry link of enterprise
CN116484230B (en) Method for identifying abnormal business data and training method of AI digital person
Porta Vallés A supervised learning framework for news-based sentiment scores conditioned to financial returns

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant