Automatic standard knowledge extraction method

Through the standard automated knowledge extraction method, the knowledge extraction bias and error problems caused by the difference in diversified data sources are solved, the accuracy and reliability of knowledge extraction are improved, and data security is guaranteed through encrypted storage.

CN119025645AInactive Publication Date: 2024-11-26CHINA NAT INST OF STANDARDIZATION
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411087724.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-09
Publication Date
2024-11-26
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

When the existing knowledge extraction method captures key knowledge from diversified data sources, the knowledge extraction process is complicated, biased and errored, reducing utility and reliability due to the differences and inconsistencies between data sources.

Method used

Standard knowledge automation extraction methods are adopted, including data conversion, preprocessing, extraction and analysis, verification and storage of standard knowledge elements. By eliminating noise, filling in missing, adjusting coordinate system, verifying extraction accuracy and encrypting storage, the accuracy of knowledge elements is ensured and Integrity.

Benefits of technology

Improve the accuracy and reliability of knowledge extraction, identify potential errors, ensure data quality and integrity, and prevent unauthorized access through encrypted storage, ensuring the security of knowledge elements.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119025645A_ABST
    Figure CN119025645A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of electronic data extraction processing, and particularly discloses a standard knowledge automatic extraction method which comprises the steps of obtaining and preprocessing to-be-extracted data, extracting and analyzing standard knowledge elements, extracting and verifying the standard knowledge elements and storing the standard knowledge elements. According to the method, the accuracy and integrity of the data and the quality of the digital data of the standard literature database can be improved by preprocessing the preliminary digital data of the standard literature database, and the extraction of the standard knowledge elements is verified to ensure that the extracted standard knowledge elements are matched with the knowledge elements in the standard literature database; the accuracy and reliability of knowledge extraction are improved, the standard knowledge elements are backed up, encrypted and stored, unauthorized access and stealing of the standard knowledge elements can be effectively prevented, meanwhile, the standard knowledge elements are backed up, it is ensured that when the standard knowledge elements are lost or damaged, the standard knowledge elements can be rapidly recovered, and the integrity of the standard knowledge elements is guaranteed.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of electronic data extraction and processing, and in particular to a method for automatic extraction of standard knowledge. Background Art

[0002] Currently, knowledge extraction methods face the challenge of capturing key knowledge from diversified sources. The significant differences and inconsistencies between different data sources make the knowledge extraction process complicated. Furthermore, current data processing methods often show limitations, which may lead to significant deviations between the extracted knowledge and actual knowledge, or even errors, thereby reducing the effectiveness and reliability of knowledge extraction.

[0003] For example, the invention patent with announcement number CN111897781 B publicly provides a method and system for extracting knowledge graph data, the method includes: S1, obtaining the original data to be extracted; S2, checking the original data and determining the data file type; S3, using optical character recognition technology to identify the characters in the image type file; S4, extracting the character content in the file and generating document data containing a string list; S5, using natural language processing technology to process the string list and extracting entities and relationships in the document data; S6, extracting knowledge graph data based on the extracted entities and relationships. The system includes: a data input management component; a data structure identification component; an optical character recognition component; a document extraction component; a natural language processing component; and a knowledge graph construction component. The method and system described in the present disclosure can enrich the types of data supported by knowledge graph data extraction and improve the efficiency of data extraction. Expand the potential data sources of the knowledge graph.

[0004] For example, the invention patent with announcement number CN109189943B discloses a method for extracting capability knowledge and constructing a capability knowledge graph, including: data collection, obtaining document data from the network and objects connected to the network; data identification, processing the document data using natural language processing methods; data understanding, labeling the document data after natural language processing according to specific rules; data screening, extracting and screening capability knowledge from the labeled data through multiple models; data induction, identifying and summarizing the labels in the screened data, and then obtaining capability knowledge. It also discloses a method for constructing a capability knowledge graph, including: extracting capability ontology through data collection, identification, understanding, screening and induction processes; using machine learning methods to learn the extraction process of capability ontology; classifying and storing knowledge ontology, and then obtaining capability knowledge graph, which solves the problem of interoperability and sharing of heterogeneous resources in the Internet of Things and realizes the automation and intelligence of capability ontology extraction.

[0005] Combining the above technical solutions, it is found that when extracting knowledge from multiple data sources, due to the differences in quality and format between the data sources, these differences often lead to erroneous or inconsistent information in the data sources, which in turn affects the accuracy and reliability of knowledge extraction. At the same time, when processing complex data, due to the diversity and complexity of the content of the literature, existing methods are difficult to guarantee high accuracy of knowledge extraction. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a method for automatically extracting standard knowledge, which can effectively solve the problems involved in the above-mentioned background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for automatic extraction of standard knowledge, including: S1. Acquisition and preprocessing of data to be extracted: data conversion is performed on the literature database to obtain preliminary digitized data of the standard literature database, and preprocessing operations are performed on the preliminary digitized data of the standard literature database to obtain standard digitized literature data.

[0008] S2. Extraction and analysis of standard knowledge elements: Extract standard knowledge elements from standard digitized document data to obtain standard knowledge element extraction data, and analyze the standard knowledge element extraction data to obtain a standard knowledge element extraction index.

[0009] S3. Verification of standard knowledge element extraction: Compare the standard knowledge element extraction index with the standard knowledge extraction reference index preset in the standard knowledge database, and finally verify the extraction of standard knowledge elements.

[0010] S4. Standard knowledge element storage: store standard knowledge elements that are verified to comply with the information extraction specifications defined in the standard knowledge database.

[0011] As a further method, the preliminary digitized data of the literature database is obtained, and the specific process is as follows:

[0012] Identify the format of the document database, extract the pure document content from the document database, and encode the document content in a unified format, thereby obtaining preliminary digital data of the document database.

[0013] As a further method, the specific process of obtaining the standard digital document data is as follows:

[0014] By performing preprocessing operations on the preliminary standard digitized document data, including eliminating noise and outliers, filling in missing parts and adjusting the coordinate system to achieve data alignment, the standard digitized document data is finally obtained.

[0015] As a further method, the standard knowledge element extraction index, the specific analysis process is:

[0016] The standard knowledge element extraction equipment impact index and the standard knowledge content element extraction index are analyzed and processed comprehensively to obtain the standard knowledge element extraction index. The specific analysis formula is:

[0017]

[0018] Where θ is the standard knowledge element extraction index, e is a natural constant, KACI is the standard knowledge element extraction equipment impact index, SKCEI is the standard knowledge content element extraction index, k1 is the weight factor corresponding to the standard knowledge element extraction equipment impact index preset in the standard knowledge database, and s1 is the weight factor corresponding to the standard knowledge content element extraction index preset in the standard knowledge database.

[0019] As a further method, the extraction of the standard knowledge elements is verified, and the specific analysis process is as follows:

[0020] The standard knowledge element extraction index is compared with the standard knowledge extraction reference index preset in the standard knowledge database to obtain the standard knowledge element extraction accuracy value, and compared with the standard knowledge element extraction accuracy threshold. If the standard knowledge element extraction accuracy value is greater than or equal to the standard knowledge element extraction accuracy threshold, the extracted standard knowledge elements meet the information extraction specifications defined in the standard knowledge database. Conversely, if the standard knowledge element extraction accuracy value is less than the standard knowledge element extraction accuracy threshold, the extracted standard knowledge elements do not meet the information extraction specifications defined in the standard knowledge base, and it is necessary to extract standard knowledge elements from the standard digitized document data again until the standard knowledge element extraction accuracy value is greater than or equal to the standard knowledge element extraction accuracy threshold, and finally complete the extraction verification of the standard knowledge elements.

[0021] As a further method, the standard knowledge elements are stored, and the specific storage process is:

[0022] The standard knowledge elements that are verified to comply with the information extraction specifications defined in the standard knowledge base are classified and stored according to appearance elements and content elements. At the same time, the stored standard knowledge elements are backed up and encrypted using data confidentiality technology, ultimately completing the storage of the standard knowledge elements.

[0023] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0024] (1) The present invention can eliminate noise and outliers in the data and fill in missing parts by preprocessing the preliminary digitized data of the standard document database, thereby improving the accuracy and completeness of the data. At the same time, the coordinate alignment and data registration of the data are performed to ensure that the data scanned from different angles or positions can be accurately matched and fused, thereby significantly improving the quality of the standard digitized document data.

[0025] (2) By verifying the extraction of standard knowledge elements, the present invention can ensure that the extracted standard knowledge elements match the knowledge elements in the standard document database, thereby greatly improving the accuracy and reliability of knowledge extraction, and at the same time identifying potential errors or deficiencies in knowledge extraction.

[0026] (3) The present invention backs up and encrypts the standard knowledge elements for storage, which can effectively prevent unauthorized access and theft of the standard knowledge elements. At the same time, the standard knowledge elements are backed up to ensure that when the standard knowledge elements are lost or damaged, they can be quickly restored to protect the integrity of the standard knowledge elements. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0028] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0029] Figure 2 This is a simulation curve diagram for extracting chapter elements involved in the present invention. DETAILED DESCRIPTION

[0030] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0031] Reference Figure 1 As shown, the present invention provides a method for automatic extraction of standard knowledge, including: S1. Acquiring and preprocessing the data to be extracted: converting the document database to obtain preliminary digitized document database data, and preprocessing the preliminary document database data to obtain standard digitized document data.

[0032] S2. Extraction and analysis of standard knowledge elements: Extract standard knowledge elements from standard digitized document data to obtain standard knowledge element extraction data, and analyze the standard knowledge element extraction data to obtain a standard knowledge element extraction index.

[0033] S3. Verification of standard knowledge element extraction: Compare the standard knowledge element extraction index with the standard knowledge extraction reference index preset in the standard knowledge database, and finally verify the extraction of standard knowledge elements.

[0034] S4. Standard knowledge element storage: store standard knowledge elements that are verified to comply with the information extraction specifications defined in the standard knowledge database.

[0035] Specifically, the process of obtaining preliminary digital data of the literature database is as follows:

[0036] Identify the format of the document database, extract the pure document content from the document database, and encode the document content in a unified format, thereby obtaining preliminary digital data of the document database.

[0037] It should be explained that the above-mentioned identification of the format of the document database requires determining the specific storage format of the document database, such as a simple document file (such as TXT, CSV), a database file (such as an export file of SQLite, MySQL), or a document in a special format (such as PDF, Word, etc.). Appropriate parsing tools or libraries can be used to read these files. For example, for TXT or CSV files, a standard file reading and writing library can be used; for PDF or Word files, a special parsing library (such as PyPDF2 or python-docx in Python) is required. To extract pure document content from the document database, it is necessary to extract the document content according to the specific format of the file. For simple document files, this may only require reading line by line; for PDF or Word files, it is necessary to parse the structure of the file and extract the paragraphs, sentences or words therein. To encode the document content in a unified format, it is necessary to convert the document from the original encoding (such as UTF-8, GBK, I SO-8859-1, etc.) to a unified character encoding (usually using UTF-8) to ensure that special characters in the document (such as Chinese characters, emoticons, etc.) can be displayed and processed correctly.

[0038] Specifically, the process of obtaining the standard digital document data is as follows:

[0039] By preprocessing the preliminary digitized data of the literature database, including eliminating noise and outliers in the data, filling in missing parts in the data and adjusting the coordinate system to achieve data alignment, standard digitized literature data is finally obtained.

[0040] It should be explained that the above-mentioned elimination of noise and outliers in the data can remove the noise through a low-pass filter, and the outliers can be removed by data cleaning methods. The missing values ​​in the data can be filled by the same mean interpolation method to ensure the integrity and accuracy of the data. Due to factors such as the working mode of the extraction equipment, the preliminary digitized data of the literature database obtained may not be in a unified coordinate system. The 3D modeling software can be used to select key points or surfaces of the data to adjust the position and direction of the data for data alignment.

[0041] Furthermore, the standard knowledge element extraction data specifically includes extraction device data and standard knowledge extraction content data.

[0042] The extraction device data includes the extraction response time period of the extraction device, the total number of historical uses of the extraction device, the historical number of failures of the extraction device, and the historical number of extraction errors of the extraction device.

[0043] The standard knowledge extraction content data includes the number of chapters, the number of keywords, and the number of measurement units of the standard knowledge.

[0044] It needs to be explained that the extraction response time period of the above-mentioned extraction device can be obtained through the records of the built-in time monitoring function of the extraction device, the total historical usage times of the extraction device can be obtained through the built-in counter of the extraction device, the historical failure times of the extraction device can be obtained through the built-in fault diagnosis function of the extraction device, and the historical extraction error times of the extraction device can be obtained from the device extraction verification records.

[0045] It should be explained that the above keywords are words used to describe the core concepts and themes of literature, articles or research, and units of measurement are standard quantities used to express and compare similar quantities.

[0046] It should be explained that the number of chapters of the above-mentioned standard knowledge can be obtained by directly viewing the table of contents. Usually the table of contents will list all chapters and sub-chapters of the standard knowledge. The number of keywords can be automatically counted using literature analysis software, such as the word frequency counting function of Word. The number of measurement units can be identified and counted by using literature search tools to identify and count the number of different measurement units.

[0047] It should be explained that the number of chapters, keywords and measurement units of the above standard knowledge is not fixed, but is obtained through statistics and analysis based on the actual situation of the standard knowledge. The specific data is shown in Table 1: Statistics of Standard Knowledge Content:

[0048] Table 1 Statistics of standard knowledge content

[0049]

[0050] From the statistical table of standard knowledge content in Table 1, it can be seen that different standard knowledge has differences in the number of chapters, the number of keywords and the number of measurement units, which reflects the different characteristics of each standard knowledge in terms of content, structure and professional fields. For example, standard knowledge C has the largest number of chapters, which means that this standard knowledge covers a wider range of topics or more detailed content. At the same time, standard knowledge C also has the highest number of keywords, which shows that the content of standard knowledge C is richer and contains more important information and viewpoints. Standard knowledge D has a smaller number of measurement units, which means that the content discussed in the standard knowledge does not involve highly specialized or technical fields. In some professional fields, such as science, engineering, medicine, etc., a large number of professional measurement units are often used due to the need to accurately describe data, experimental results or treatment plans. If standard knowledge D has a smaller number of measurement units, then it may be more oriented to general readers, or involve more general or basic knowledge fields.

[0051] Specifically, the standard knowledge element extracts data for analysis, and the specific analysis process is as follows:

[0052] The failure rate of the extracted equipment is obtained by ratioing the historical failure times of the extracted equipment with the total historical usage times of the extracted equipment.

[0053] The historical number of extraction errors of the extraction device is ratioed with the total number of historical uses of the extraction device to obtain the extraction error rate of the extraction device.

[0054] Specifically, the standard knowledge element extraction device impact index, the specific analysis process is:

[0055] The extraction response time period, failure rate and extraction error rate of the extraction equipment are processed comprehensively to obtain the standard knowledge element extraction equipment impact index. The specific analysis formula is:

[0056]

[0057] In the formula, KACI is the standard knowledge element extraction equipment impact index. A larger extraction response time period means that the amount of data that the device can process per unit time is reduced, thereby reducing the overall data processing efficiency, which will delay the progress of knowledge element extraction and affect subsequent analysis and utilization. In order to complete the same amount of data extraction tasks, devices with long response times may require longer working hours, which will increase the operating cost and energy consumption of the equipment, and may also occupy more computing resources or storage space. Frequent failures will cause the equipment to shut down for maintenance, thereby interrupting the extraction process and significantly reducing the efficiency of data processing. Increased downtime not only reduces the effective working time, but may also cause data backlog and delayed processing, further affecting subsequent analysis and utilization. A high extraction error rate means that there is a lot of inaccurate or erroneous information in the knowledge elements extracted from the data source, which will directly It affects the accuracy and reliability of the extraction results. At the same time, the increased failure rate of the extraction equipment will lead to a longer extraction response time period. When the equipment fails, it takes longer to recover or bypass the fault point, thereby increasing the total time required to complete the extraction task. A high failure rate usually means that there are unstable factors in the equipment or system, which may also cause errors in the extraction process. For example, hardware failure may cause data transmission interruption or damage, and software defects may cause algorithm execution errors. Therefore, by analyzing the data in the standard knowledge element extraction equipment impact index, the performance bottlenecks and deficiencies of the extraction equipment can be revealed, and the hardware configuration, software algorithm or system architecture of the extraction equipment can be optimized in a targeted manner, thereby improving the overall performance and extraction efficiency of the extraction equipment, enhancing the accuracy and reliability of the extracted data, and providing strong support for subsequent extraction data analysis and application.

[0058] w1 is the correction factor corresponding to the extraction response time period preset in the standard knowledge database, which refers to the degree of influence of the change in extraction response time on the standard knowledge element extraction device impact index. The fitting curve corresponding to the extraction response time period can be obtained by fitting the corresponding relationship between the historical usage time of the extraction device and the extraction response time period. The real-time usage time of the extraction device is brought into the fitting curve corresponding to the extraction response time period, thereby obtaining the correction factor corresponding to the extraction response time period in this embodiment, and the value range is (0, 1).

[0059] T is the extraction response time period of the extraction device, which refers to the time period from when the system sends an extraction request to when the extraction device completes data extraction and returns the result.

[0060] The extraction reference response time period preset for the standard knowledge database refers to the maximum time limit from the issuance of a data extraction request to the system giving a response.

[0061] g is the failure rate of the extraction equipment, which refers to an indicator that measures the frequency of failures of the extraction equipment within a certain period of time.

[0062] w2 is the impact factor corresponding to the unit value of the failure rate preset in the standard knowledge database, which refers to the degree of influence of the change in the failure rate on the impact index of the standard knowledge element extraction device. The fitting curve corresponding to the failure rate can be fitted through the correspondence between the change in the number of historical failures and the failure rate. The real-time number of failures is brought into the fitting curve corresponding to the failure rate, thereby obtaining the impact factor corresponding to the failure rate in this embodiment, and the value range is (0, 1).

[0063] c is the extraction error rate of the extraction device, which refers to the proportion of extraction errors that occur during the extraction process of the extraction device.

[0064] w3 is the impact factor corresponding to the unit value of the extraction error rate preset in the standard knowledge database, which refers to the degree of influence of the change in the extraction error rate on the impact index of the standard knowledge element extraction device. The extraction error rate fitting curve can be obtained by fitting the corresponding relationship between the historical number of extraction errors and the extraction error rate. The real-time number of extraction errors is brought into the extraction error rate fitting curve to obtain the impact factor corresponding to the unit value of the extraction error rate in this embodiment, and the value range is (0, 1).

[0065] In this embodiment, by nonlinearly processing the extraction response time period, failure rate and extraction error rate of the extraction device, the complex and nonlinear relationship between these variables can be converted into a mathematical form that is easier to analyze and process, better reflecting the differences between the data, suppressing the influence of outliers, and facilitating mathematical operations and derivations, thereby simplifying the calculation process and improving calculation accuracy.

[0066] Furthermore, the specific analysis process of the standard knowledge content element extraction index is as follows:

[0067] The number of chapters, keywords and measurement units of standard knowledge are comprehensively analyzed to obtain the standard knowledge content element extraction index. The specific analysis formula is:

[0068]

[0069] Where SKCEI is the standard knowledge content element extraction index. When the number of chapters of standard knowledge is large, it may lead to information redundancy and dispersion. Each chapter may contain elements related to the standard knowledge content, but it may also contain a large amount of non-critical information, which increases the difficulty of extracting and identifying key information and may reduce the extraction efficiency. When the number of chapters of standard knowledge is small, it may mean that the standard knowledge content elements contained are small, which may lead to the extraction of insufficient information to accurately extract the required knowledge elements, thereby affecting the extraction effect. The number of relation words is small, and the complex relationship between entities may not be fully captured when extracting standard knowledge, resulting in the limitation of relation extraction. The small number of measurement units may limit the accurate expression of certain physical quantities or concepts, resulting in deviation or misleading of the extracted information. By analyzing the data of the standard knowledge content element extraction index, the weaknesses or deficiencies in the current knowledge extraction method can be identified, such as the problem of small number of relation words and measurement units. Based on the analysis results, the knowledge extraction method can be optimized in a targeted manner, such as increasing the recognition ability of relation words and measurement units, thereby improving the extraction efficiency and accuracy.

[0070] s c The number of chapters in the standard knowledge refers to the actual number of chapters in the standard knowledge.

[0071] Δs c The number of standard knowledge chapter references preset for the standard knowledge database refers to the chapter structure of the standard knowledge and the number of reference information.

[0072] f1 is the correction factor corresponding to the number of chapters of standard knowledge preset in the standard knowledge database, which refers to the degree of influence of the change in the number of chapters of standard knowledge on the extraction index of standard knowledge content elements. The fitting curve corresponding to the number of chapters can be obtained by fitting the correspondence between the changes in the historical chapter themes of standard knowledge and the changes in standard knowledge chapters. The real-time chapter themes are brought into the fitting curve corresponding to the number of chapters, thereby obtaining the correction factor corresponding to the number of chapters of standard knowledge in this embodiment, and the value range is (0, 1).

[0073] s k is the number of keywords, which refers to the number of keywords extracted from standard knowledge and marked as keywords.

[0074] Δs k The reference number of keywords preset for the standard knowledge database refers to the number of keywords pre-defined for the standard knowledge.

[0075] f2 is the correction factor corresponding to the number of keywords preset in the standard knowledge database, which refers to the degree of influence of the change in the number of keywords on the standard knowledge content element extraction index. The fitting curve corresponding to the number of keywords can be obtained by fitting the correspondence between the changes in the historical length and historical content complexity of the document and the changes in the number of keywords. The real-time document length and content complexity are brought into the fitting curve corresponding to the number of keywords, thereby obtaining the correction factor corresponding to the number of keywords in this embodiment, and the value range is (0, 1).

[0076] s u It is the number of measurement units, which refers to the actual number of measurement units appearing in standard knowledge.

[0077] Δs u The allowed number of measurement units preset for the standard knowledge database refers to the predefined limit number of measurement units.

[0078] f3 is the correction factor corresponding to the preset number of measurement units in the standard knowledge database, which refers to the degree of influence of the change in the number of measurement units on the extraction index of standard knowledge content elements. The corresponding relationship between the historical length of the document and the change in the number of measurement units is obtained by fitting the fitting curve corresponding to the measurement unit. The real-time document length is brought into the fitting curve corresponding to the measurement unit, thereby obtaining the correction factor corresponding to the number of measurement units in this embodiment, and the value range is (0, 1).

[0079] In this embodiment, the calculation and analysis process can be simplified by performing nonlinear operations on the number of chapters, the number of keywords, the number of measurement units of standard knowledge and the extraction index of knowledge content elements. For data with different orders of magnitude and distribution characteristics in the number of chapters, the number of keywords, and the number of measurement units of standard knowledge, nonlinear operations can help eliminate scale differences, making data from different sources or types more comparable.

[0080] In this embodiment, a simulation curve is constructed between the number of chapters of standard knowledge and the standard knowledge content element extraction index as shown in FIG. Figure 2 As shown in the chapter element extraction simulation curve, the horizontal axis in the figure is the number of chapters in the book, in units of pieces, and the vertical axis is the standard knowledge content element extraction index corresponding to the number of chapters, in units of %.

[0081] By observation Figure 2 The chapter element extraction simulation curve shows that when the number of chapters of standard knowledge is small (such as 1-5 chapters), the content element extraction index may be low, because fewer chapters may only contain part of the core content. As the number of chapters increases (such as 6-15 chapters), the content element extraction index may rise rapidly, because more chapters provide a wider range of content and information, which is conducive to extracting more content elements.

[0082] Specifically, the standard knowledge element extraction index has a specific analysis process as follows:

[0083] The standard knowledge element extraction equipment impact index and the standard knowledge content element extraction index are comprehensively analyzed to obtain the standard knowledge element extraction index. The specific analysis formula is:

[0084]

[0085] Where θ is the standard knowledge element extraction index. A higher standard knowledge element extraction device impact index means that there are more problems in the operation of the extraction device. These problems will reduce the efficiency and quality of knowledge extraction, increase additional correction and processing costs, and ultimately affect the reliability of the entire standard knowledge extraction. When the content element extraction index is low, it means that the key information in the standard knowledge (such as chapters, keywords, units of measurement, etc.) may not be completely extracted, resulting in a decline in the quality of standard knowledge and affecting users' understanding and application of standard knowledge elements. By analyzing the various data of the standard knowledge element extraction index and analyzing the various data of the extraction index in detail, the integrity of the standard knowledge element extraction can be ensured. Through data analysis, the possible errors or deviations in the extraction process can be revealed, so as to carry out targeted optimization and improvement and improve the accuracy of the extraction results.

[0086] In this embodiment, by using a nonlinear expression to calculate the standard knowledge element extraction index, it is possible to achieve nonlinear conversion of the standard knowledge element extraction device influence index and the standard knowledge content element extraction index, which can reveal nonlinear relationships and hidden patterns in the data. Nonlinear transformation can better capture subtle changes and dynamic characteristics in the data, thereby providing more accurate and comprehensive information. When the values ​​of the standard knowledge element extraction device influence index and the standard knowledge content element extraction index are very large or very small, the standard knowledge element extraction index may be too large or too small, which may make the calculation unstable or cause numerical overflow. At the same time, through nonlinear transformation, the data can be standardized to a range that is easier to process or compare, which can reduce the impact of extreme values ​​and make the numerical value more stable.

[0087] e is a natural constant, KACI is the standard knowledge element extraction equipment impact index, and SKCEI is the standard knowledge content element extraction index.

[0088] k1 is the weight factor corresponding to the standard knowledge element extraction device influence index preset in the standard knowledge database, and s1 is the weight factor corresponding to the standard knowledge content element extraction index preset in the standard knowledge database.

[0089] It should be explained that the weight factor corresponding to the standard knowledge element extraction device impact index preset in the above-mentioned standard knowledge database refers to the degree of influence of the change of the standard knowledge element extraction device impact index on the standard knowledge element extraction index. The fitting curve corresponding to the standard knowledge element extraction device impact index can be obtained by fitting the correspondence between the historical extraction response time, historical failure rate, historical extraction error rate of the extraction device and the standard knowledge element extraction device impact index. At the same time, the real-time extraction response time, real-time failure rate and real-time extraction error rate of the extraction device are brought into the fitting curve corresponding to the standard knowledge element extraction device impact index, thereby obtaining the weight factor corresponding to the standard knowledge element extraction device impact index in this embodiment, and the value range is (0, 1); the standard knowledge element extraction device impact index preset in the standard knowledge database The weight factor corresponding to the knowledge content element extraction index refers to the degree of influence of the change of the standard knowledge content element extraction index on the standard knowledge element extraction index. The fitting curve corresponding to the content element extraction can be fitted through the correspondence between the number of historical chapters, historical keywords, historical measurement units and the changes in content element extraction of the standard knowledge. At the same time, the real-time number of chapters, keywords and measurement units of the standard knowledge are brought into the fitting curve corresponding to the content element extraction to obtain the weight factor corresponding to the standard knowledge content element extraction index in this embodiment. In this embodiment, the size of the weight factor can be allocated according to the standard knowledge element extraction device influence index and the degree of influence of the standard knowledge content element extraction index on the standard knowledge element extraction index, where k1+s1=1.

[0090] In this embodiment, the weight factor k1 corresponding to the above-mentioned standard knowledge element extraction device influence index is set to 0.2, and the weight factor s1 corresponding to the standard knowledge content element extraction index is set to 0.8. The standard knowledge element extraction index is shown in Table 2:

[0091] Table 2 Standard knowledge element extraction table

[0092]

[0093] By observing the data in the standard knowledge element extraction table in Table 2 above, it is found that the standard knowledge element extraction index will change with the changes in the standard knowledge element extraction device influence index and the standard knowledge content element extraction index. When the standard knowledge element extraction device influence index and the standard knowledge content element extraction index are larger, the corresponding standard knowledge element extraction index will also be larger. For example, when the standard knowledge element extraction device influence index is 60 and the standard knowledge content element extraction index is 20, the standard knowledge element extraction index is 17.9; conversely, when the standard knowledge element extraction device influence index and the standard knowledge content element extraction index are smaller, the standard knowledge element extraction index will also be smaller. For example, when the standard knowledge element extraction device influence index is 20 and the standard knowledge content element extraction index is 5, the standard knowledge element extraction index is 15.9.

[0094] Specifically, the extraction of standard knowledge elements is verified, and the specific analysis process is as follows:

[0095] The standard knowledge element extraction index is compared with the standard knowledge extraction reference index preset in the standard knowledge database to obtain the standard knowledge element extraction accuracy value, and compared with the standard knowledge element extraction accuracy threshold preset in the standard knowledge database. If the standard knowledge element extraction accuracy value is greater than or equal to the standard knowledge element extraction accuracy threshold, the extracted standard knowledge elements meet the information extraction specifications defined in the standard knowledge database. Conversely, if the standard knowledge element extraction accuracy value is less than the standard knowledge element extraction accuracy threshold, the extracted standard knowledge elements do not meet the information extraction specifications defined in the standard knowledge database, and it is necessary to extract standard knowledge elements from the standard digitized document data again until the standard knowledge element extraction accuracy value is greater than or equal to the standard knowledge element extraction accuracy threshold, and finally complete the extraction verification of the standard knowledge elements.

[0096] It should be explained that the above-mentioned standard knowledge extraction reference indicators refer to the standards or indicators used to evaluate the performance of standard knowledge element extraction methods during the standard knowledge element extraction process, which can be obtained through the historical data of standard knowledge element extraction and the evaluation of extraction standards by standard knowledge element extraction experts; the standard knowledge element extraction accuracy threshold refers to the minimum standard for evaluating the accuracy of standard knowledge element extraction, which can be obtained by consulting experts in the standard knowledge field to evaluate the standard knowledge element extraction results, mark the correct extraction results and the wrong extraction results, and finally obtain the standard knowledge element extraction accuracy threshold; information extraction specifications refer to a series of standards and rules that need to be followed when extracting standard knowledge elements to ensure that the extracted standard knowledge elements meet the expected format, quality and accuracy. They can be obtained by searching for the latest research results and papers on information extraction through academic databases or library resources, and conducting a comprehensive analysis in combination with historical standard knowledge element extraction standards.

[0097] Furthermore, the standard knowledge elements are stored, and the specific storage process is:

[0098] The standard knowledge elements that are verified to meet the information extraction specifications defined in the standard knowledge base are classified and stored according to appearance elements and content elements. At the same time, the stored standard knowledge elements are backed up and encrypted using data confidentiality technology, and finally the standard knowledge elements are stored.

[0099] It needs to be explained that the above-mentioned standard knowledge elements are classified and stored according to appearance elements and content elements, which can quickly locate the required information and make the management of standard knowledge elements more orderly and efficient. Different management measures are specified for different categories of elements, such as backup, update, and deletion. At the same time, the main purpose of backing up the stored standard knowledge elements is to be able to quickly restore data when the data is lost or damaged. The stored standard knowledge elements can be backed up in an automated backup method to ensure that the standard knowledge elements can minimize data loss in the worst case. Encryption processing can ensure the security of data during storage and transmission, reduce the risk of data leakage, and even if the data is illegally obtained, the data content cannot be decrypted and read without the corresponding key, thereby reducing the possibility of data leakage.

[0100] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined in the specification, they should all fall within the protection scope of the present invention.

Claims

1. A method for automatic extraction of standard knowledge, characterized in that: include: S1. Acquisition and preprocessing of data to be extracted: converting the literature database to obtain preliminary digital data of the standard literature database, and preprocessing the preliminary digital data of the standard literature database to obtain standard digital literature data; S2. Extraction and analysis of standard knowledge elements: extract standard knowledge elements from standard digital document data to obtain standard knowledge element extraction data, and analyze the standard knowledge element extraction data to obtain a standard knowledge element extraction index; S3. Standard knowledge element extraction verification: compare the standard knowledge element extraction index with the standard knowledge extraction reference index preset in the standard knowledge database, and finally verify the extraction of standard knowledge elements; S4. Standard knowledge element storage: store standard knowledge elements that are verified to comply with the information extraction specifications defined in the standard knowledge database.

2. The method for automatic extraction of standard knowledge according to claim 1, characterized in that: The specific process of obtaining preliminary digital data of the literature database is as follows: Identify the format of the document database, extract the pure document content from the document database, and encode the document content in a unified format, thereby obtaining preliminary digital data of the document database.

3. The method for automatic extraction of standard knowledge according to claim 2, characterized in that: The specific process of obtaining the standard digital document data is as follows: By preprocessing the preliminary digitized data of the literature database, including eliminating noise and outliers, filling in missing parts and adjusting the coordinate system to achieve data alignment, standard digitized literature data is finally obtained.

4. The method for automatic extraction of standard knowledge according to claim 1, characterized in that: The specific analysis process of the standard knowledge element extraction index is as follows: The standard knowledge element extraction equipment impact index and the standard knowledge content element extraction index are analyzed, and the standard knowledge element extraction index is obtained by comprehensive processing. The specific analysis formula is: Where θ is the standard knowledge element extraction index, e is a natural constant, KACI is the standard knowledge element extraction equipment impact index, SKCEI is the standard knowledge content element extraction index, k1 is the weight factor corresponding to the standard knowledge element extraction equipment impact index preset in the standard knowledge database, and s1 is the weight factor corresponding to the standard knowledge content element extraction index preset in the standard knowledge database.

5. The method for automatic extraction of standard knowledge according to claim 1, characterized in that: The standard knowledge element extraction data specifically includes extraction device data and standard knowledge standard knowledge extraction content data; The extraction device data includes the extraction response time period of the extraction device, the total number of historical uses of the extraction device, the number of historical failures of the extraction device, and the number of historical extraction errors of the extraction device; The standard knowledge extraction content data includes the number of chapters, the number of keywords, and the number of measurement units of the standard knowledge.

6. The method for automatic extraction of standard knowledge according to claim 5, characterized in that: The analysis of the standard knowledge element extraction data is carried out, and the specific analysis process is as follows: The historical failure times of the extracted equipment are processed by ratio processing with the total historical usage times of the extracted equipment to obtain the failure rate of the extracted equipment; The historical number of extraction errors of the extraction device is ratioed with the total number of historical uses of the extraction device to obtain the extraction error rate of the extraction device.

7. The method for automatic extraction of standard knowledge according to claim 4, characterized in that: The standard knowledge element extraction equipment impact index, the specific analysis process is: The extraction response time period, failure rate and extraction error rate of the extraction device are comprehensively processed to obtain the standard knowledge element extraction device impact index.

8. The method for automatic extraction of standard knowledge according to claim 4, characterized in that: The specific analysis process of the standard knowledge content element extraction index is as follows: The number of chapters, keywords and measurement units of standard knowledge are comprehensively analyzed to obtain the standard knowledge content element extraction index.

9. The method for automatic extraction of standard knowledge according to claim 1, characterized in that: The extraction of standard knowledge elements is verified, and the specific analysis process is as follows: The standard knowledge element extraction index is compared with the standard knowledge extraction reference index preset in the standard knowledge database to obtain the standard knowledge element extraction accuracy value, and compared with the standard knowledge element extraction accuracy threshold. If the standard knowledge element extraction accuracy value is greater than or equal to the standard knowledge element extraction accuracy threshold, the extracted standard knowledge elements meet the information extraction specifications defined in the standard knowledge database. Conversely, if the standard knowledge element extraction accuracy value is less than the standard knowledge element extraction accuracy threshold, the extracted standard knowledge elements do not meet the information extraction specifications defined in the standard knowledge base, and it is necessary to extract standard knowledge elements from the standard digitized document data again until the standard knowledge element extraction accuracy value is greater than or equal to the standard knowledge element extraction accuracy threshold, and finally complete the extraction verification of the standard knowledge elements.

10. The method for automatic extraction of standard knowledge according to claim 1, characterized in that: The standard knowledge elements are stored, and the specific storage process is: The standard knowledge elements that are verified to comply with the information extraction specifications defined in the standard knowledge base are classified and stored according to appearance elements and content elements. At the same time, the stored standard knowledge elements are backed up and encrypted using data confidentiality technology, ultimately completing the storage of the standard knowledge elements.

Citation Information

Patent Citations

  • Knowledge graph quality evaluation method and device, computer equipment and storage medium

    CN113779261A

  • Video data processing method and device, computer equipment and storage medium

    CN114637846A