A retrieval system and method for knowledge governance based on enterprise classification criteria

Abstract

This invention provides a retrieval system and method based on enterprise classification standards and knowledge governance, relating to the field of data analysis technology. The retrieval system includes: an acquisition unit for acquiring newly added classification standard files within the enterprise; a conflict analysis unit for performing conflict analysis based on the newly added classification standard files and existing classification dimensions to obtain updated classification dimensions and their corresponding initial dimension filling weights; a knowledge governance unit for updating the initial dimension filling weights based on updated classification dimensions and dimension knowledge data when new knowledge is received, to obtain updated dimension filling weights; and an output analysis unit for responding to user query requests by performing hybrid retrieval and matching analysis on the repository based on dimension knowledge data and updated dimension filling weights to output retrieval results. The system and method provided by this invention solve the problem of significant retrieval result errors caused by a lack of dimensional knowledge.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, and in particular to a retrieval system and method for knowledge governance based on enterprise classification standards. Background Technology

[0002] In various large and medium-sized enterprises, especially in knowledge-intensive industries such as industrial engineering and equipment manufacturing, their core competitiveness increasingly depends on the efficient management and precise application of massive, multi-source, heterogeneous knowledge assets (such as technical specifications, design drawings, project documents, standard drawing sets, and failure cases).

[0003] In existing enterprise asset knowledge management systems, general, pre-set data models are often used, which cannot flexibly adapt to the unique and authoritative internal knowledge classification standards of different enterprises. Moreover, the data is often stored in an unstructured manner, and its intrinsic value (such as which product line it belongs to, what standards it follows, and which projects it is associated with) is not extracted and associated, becoming data silos. When searching, it is only used to directly search using the user's query keywords, which is not only noisy but also results in low search accuracy. At the same time, it is difficult to effectively determine the coverage of the knowledge corresponding to the search results in the knowledge base, which greatly reduces the reliability of the search results and makes the selection of search results too blind. Summary of the Invention

[0004] This invention provides a retrieval system and method for knowledge governance based on enterprise classification standards. This addresses the problems in enterprise asset knowledge management, where systems often employ generic, pre-defined data models that cannot flexibly adapt to the unique and authoritative internal knowledge classification standards of different enterprises. Furthermore, data is often stored in an unstructured manner, failing to extract and correlate its intrinsic value, resulting in data silos. Retrieval also relies solely on user-defined keywords, leading to high noise levels, low accuracy, and difficulty in effectively determining the coverage of knowledge dimensions corresponding to the retrieval results in the knowledge base, significantly reducing the reliability of the results and resulting in overly arbitrary selection of search results.

[0005] To achieve the above and other related objectives, this invention provides a retrieval system for knowledge governance based on enterprise classification standards, comprising: an acquisition unit for acquiring newly added classification standard files within the enterprise; a conflict analysis unit for performing conflict analysis based on the newly added classification standard files and existing classification dimensions to obtain updated classification dimensions and their corresponding initial dimension filling weights; a knowledge governance unit for governing new knowledge based on updated classification dimensions when new knowledge is received, obtaining dimension knowledge data for storage in a repository, and updating the initial dimension filling weights based on the dimension knowledge data to obtain updated dimension filling weights; and an output analysis unit for responding to user query retrieval requests by performing hybrid retrieval and matching analysis on the repository based on dimension knowledge data and updated dimension filling weights to output retrieval results, wherein the retrieval request includes retrieval elements and retrieval dimensions.

[0006] In one embodiment of the present invention, the conflict analysis unit includes: a file parsing subunit, used to parse the newly added classification standard file to obtain multiple dimensions to be classified and corresponding new dimension value ranges; a dimension comparison subunit, used to compare the dimensions to be classified and the existing dimension combinations corresponding to each existing classification standard file according to the new dimension value range, so as to calculate the dimension matching degree between the dimensions to be classified and each existing dimension combination; and a first dimension partitioning subunit, used to, when the dimension matching degree is greater than the matching degree threshold, use the newly added classification standard file as a correction file of the corresponding existing classification standard file, and partition the dimensions to be classified into consistent dimensions that correspond to and are consistent with the first existing classification dimension in the existing dimension combination, and with... The existing dimension combination includes a replacement dimension that corresponds to the second existing classification dimension and has dimension differences, and a new dimension that does not correspond to any of the existing classification dimensions in the existing dimension combination. The process involves deleting all existing classification dimensions in the existing dimension combination except for the first existing classification dimension, replacing the first existing classification dimension with the replacement dimension, and adding the new dimension to the existing dimension combination to obtain the updated classification dimension. A second dimension partitioning subunit is used to use all unclassified dimensions corresponding to the new classification standard file as the updated classification dimension when the dimension matching degree is less than the matching degree threshold. A data detection subunit is used to perform knowledge data detection on the repository based on the updated classification dimension to obtain the initial dimension filling weights corresponding to the updated classification dimension.

[0007] In one embodiment of the present invention, the dimension comparison subunit includes: a semantic comparison module, used to compare the semantic similarity of the dimension to be classified with each existing classification dimension in each existing dimension combination under the corresponding professional dimension to obtain the dimension similarity; a first dimension calculation module, used to take the dimension to be classified that has a dimension similarity greater than a first similarity threshold with the existing classification dimensions as consistent dimensions, obtain the first consistent dimension matching degree corresponding to each consistent dimension, compare the similarity of the first new dimension value range corresponding to the consistent dimension with the existing dimension value range corresponding to the corresponding existing classification dimension to obtain the second consistent dimension matching degree, and perform weighted fusion of the first consistent dimension matching degree and the second consistent dimension matching degree to calculate the first dimension matching degree between the dimension to be classified and each existing dimension combination; and a second dimension calculation module, used to take the dimension to be classified that has a dimension similarity less than the first similarity threshold and greater than the second similarity threshold with the existing classification dimensions as replacement dimensions, and calculate the replacement dimension based on the dimension similarity and the first matching conversion coefficient. The system obtains the first replacement dimension matching degree for each replacement dimension, compares the similarity between the second new dimension value range corresponding to the replacement dimension and the existing dimension value range corresponding to the corresponding existing classification dimension to obtain the second replacement dimension matching degree, and performs a weighted fusion of the first and second replacement dimension matching degrees to calculate the second dimension matching degree between the dimension to be classified and each combination of existing dimensions; the first matching conversion coefficient is obtained by averaging the first conversion coefficient corresponding to the existing classification dimension and the second conversion coefficient corresponding to the replacement dimension; the third dimension calculation module is used to treat the corresponding dimension to be classified as a new dimension when the dimension similarity between the existing classification dimension and the dimension to be classified is lower than the second similarity threshold, and obtains the matching degree reduction amount based on the third dimension weight of the new dimension in the dimension to be classified and the matching degree reduction coefficient; and the dimension fusion module is used to calculate the dimension matching degree between the dimension to be classified and each combination of existing dimensions based on the first dimension matching degree, the second dimension matching degree, and the matching degree reduction amount.

[0008] In one embodiment of the present invention, the formula for calculating the dimensional matching degree is: ;in, Indicates the degree of dimensional matching. This represents the first matching weight corresponding to the first dimension matching degree. This represents the second matching weight corresponding to the second dimension matching degree. This represents the third matching weight corresponding to the reduction in matching degree. This indicates the matching degree of the first consistency dimension. This indicates the matching degree of the second consistency dimension. This represents the weight of the first dimension when each dimension to be classified is a consistent dimension. This indicates the first-dimensional keyword corresponding to the replacement dimension. This indicates the second-dimensional keywords corresponding to the existing category dimensions. Indicates dimensional similarity. Indicates the first conversion factor. This represents the second conversion factor. Indicates the matching degree of the second replacement dimension. This represents the weight of the second dimension when each dimension to be classified is used as a replacement dimension. This represents the decreasing coefficient of matching degree. This represents the weight of the third dimension when each dimension to be classified is a newly added dimension.

[0009] In one embodiment of the present invention, the data detection subunit includes: a first weight calculation module, used to, when the updated classification dimension is a consistent dimension, use the existing dimension filling weight of the existing classification dimension corresponding to the consistent dimension as the initial dimension filling weight; a first knowledge output module, used to, when the updated classification dimension is a replacement dimension, perform semantic deviation analysis on the replacement dimension and the existing classification dimension to obtain the deleted semantic range and the added semantic range, control the deletion of the existing dimension knowledge of the existing classification dimension corresponding to the replacement dimension according to the deleted semantic range, and perform knowledge search on the repository according to the added semantic range to obtain the updated dimension knowledge; a second knowledge output module, used to, when the updated classification dimension is a new dimension, perform content search on the repository according to the new dimension to obtain the updated dimension knowledge; and a second weight calculation module, used to perform dimension filling degree analysis on the updated dimension knowledge to obtain the initial dimension filling weight corresponding to the updated classification dimension.

[0010] In one embodiment of the present invention, the first knowledge output module performs semantic deviation analysis on the replacement dimension and the existing classification dimension to obtain the deleted semantic range and the added semantic range. This process includes: a search submodule for searching for a first set of keywords corresponding to the semantics of the first dimension keywords of the replacement dimension, and a second set of keywords corresponding to the semantics of the second dimension keywords of the existing classification dimension; a difference comparison submodule for comparing the semantic differences between the first set of keywords and the second set of keywords to obtain first filter keywords that correspond only to the replacement dimension and second filter keywords that correspond only to the existing classification dimension; a deletion range output submodule for semantically fusing the first semantics of all the first filter keywords to obtain the deleted semantic range; and an added range output submodule for semantically fusing the second semantics corresponding to all the second filter keywords to obtain the added semantic range.

[0011] In one embodiment of the present invention, the second weight calculation module includes: a value range filtering submodule, used to filter the difference value ranges of the updated dimension knowledge to obtain the difference value ranges corresponding to the difference knowledge entries, and to obtain multiple coverage value ranges based on the difference value ranges; a quantity filling calculation submodule, used to obtain the quantity filling degree based on the range length corresponding to each coverage value range and the number of difference knowledge entries within the range length; a knowledge completeness calculation submodule, used to obtain the corresponding knowledge completeness based on the dimension value range and coverage value ranges corresponding to the updated classification dimension; and a weight output submodule, used to weight and fuse the quantity filling degree and knowledge completeness to obtain the initial dimension filling weight corresponding to the updated classification dimension; the calculation formula for the initial dimension filling weight is: ;in, This indicates that the initial dimension is filled with weights. This indicates the first fill weight corresponding to the quantity fill degree. This represents the second filling weight corresponding to the knowledge completeness. This indicates the number of knowledge entries representing differences within the specified range. Indicates the conversion factor. This indicates the range length corresponding to each covered value range. Indicates the number of covered value ranges. This indicates the total range length of the dimension value range corresponding to the updated category dimension.

[0012] In one embodiment of the present invention, the initial dimension filling weight includes quantity filling degree and knowledge completeness; the knowledge governance unit includes: a knowledge storage subunit, used to parse new knowledge and obtain the dimension knowledge data corresponding to each updated category dimension in the new knowledge for storage in the repository; an analysis and calculation subunit, used to update the quantity filling degree and knowledge completeness according to the dimension knowledge data to obtain updated quantity filling degree and updated knowledge completeness; and a weight update subunit, used to obtain updated dimension filling weights according to updated quantity filling degree and updated knowledge completeness.

[0013] In one embodiment of the present invention, the output analysis unit includes: a retrieval output subunit, configured to perform a mixed retrieval of the repository in response to a user's retrieval request to obtain output knowledge entries; a reliability calculation subunit, configured to obtain the output reliability of each output knowledge entry based on the update dimension filling weights of each update category dimension corresponding to the output knowledge entry and the importance coefficient of each update category dimension; and a sorting output subunit, configured to sort the output knowledge entries according to the output reliability to output the retrieval results; the formula for calculating the output reliability is: ;in, This indicates the output reliability of each output knowledge entry. This represents the update dimension fill weight for each output knowledge entry. This represents the importance coefficient of each output knowledge item.

[0014] To achieve the above and other related objectives, the present invention also provides a retrieval method for knowledge governance based on enterprise classification standards, comprising: acquiring new classification standard files within the enterprise through an acquisition unit; performing conflict analysis based on the new classification standard files and existing classification dimensions through a conflict analysis unit to obtain updated classification dimensions and their corresponding initial dimension filling weights; when new knowledge is received, the knowledge governance unit governs the new knowledge according to the updated classification dimensions to obtain dimension knowledge data for storage in the repository, and updates the initial dimension filling weights based on the dimension knowledge data to obtain updated dimension filling weights; and responding to a user's retrieval request through an output analysis unit, performing a hybrid retrieval and matching analysis on the repository based on dimension knowledge data and updated dimension filling weights to output retrieval results, wherein the retrieval request includes retrieval elements and retrieval dimensions.

[0015] The beneficial effects of this invention are as follows: This invention proposes a retrieval system and method based on enterprise classification standards for knowledge governance. By utilizing newly added classification standard files and combining them with existing classification dimensions for conflict analysis, updated classification dimensions are generated, and the initial dimension filling weights corresponding to the updated classification dimensions are output. Furthermore, when new knowledge is updated, the received new knowledge can be governed using updated classification dimensions to generate and store dimensional knowledge data corresponding to different updated classification dimensions. Moreover, the initial dimension filling weights are dynamically updated and adjusted based on the dimensional knowledge data, and the resulting updated dimension filling weights are used for user retrieval. During user retrieval, output knowledge entries are generated based on the search elements and search dimensions in the corresponding search request. By performing matching analysis on the output knowledge entries using updated dimension filling weights, the reliability of each output knowledge entry is determined. This reliability helps users better judge the accuracy of search results, avoiding significant cognitive errors caused by blindly trusting output search results based on a limited amount of dimensional knowledge when the amount of dimensional knowledge is insufficient. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0017] In the attached diagram: Figure 1A structural block diagram of a knowledge governance retrieval system based on enterprise classification standards provided in an embodiment of the present invention; Figure 2 The diagram shows a flowchart of a knowledge governance retrieval method based on enterprise classification standards provided in an embodiment of the present invention.

[0018] The attached figures are labeled as follows: Acquisition Unit 111; Conflict Analysis Unit 112; Knowledge Governance Unit 113; Output Analysis Unit 114. Detailed Implementation

[0019] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0020] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. The drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0021] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0022] Please see Figure 1This invention provides a knowledge governance retrieval system based on enterprise classification standards, comprising: an acquisition unit 111 for acquiring newly added classification standard files within the enterprise; a conflict analysis unit 112 for performing conflict analysis based on the newly added classification standard files and existing classification dimensions to obtain updated classification dimensions and their corresponding initial dimension filling weights; a knowledge governance unit 113 for governing new knowledge based on updated classification dimensions when new knowledge is received, obtaining dimension knowledge data for storage in the repository, and updating the initial dimension filling weights based on the dimension knowledge data to obtain updated dimension filling weights; and an output analysis unit 114 for responding to user query retrieval requests, performing hybrid retrieval and matching analysis on the repository based on dimension knowledge data and updated dimension filling weights to output retrieval results, wherein the retrieval request includes retrieval elements and retrieval dimensions.

[0023] As can be seen from the above, in the process of enterprise knowledge governance, the newly uploaded classification standard file can be obtained first through the acquisition unit 111. Prior to this, classification dimensions have been pre-defined based on other internal classification standards and conventional enterprise standards, serving as existing classification dimensions. Therefore, the conflict analysis unit 112 can utilize the newly uploaded classification standard file and combine it with existing classification dimensions to perform conflict analysis, thereby updating and generating new classification dimensions, and outputting the initial dimension filling weights corresponding to the updated classification dimensions. Furthermore, when new knowledge is updated, the knowledge governance unit 113 can also govern the received new knowledge using the updated classification dimensions to generate and store dimensional knowledge data corresponding to different updated classification dimensions. Moreover, the initial dimension filling weights will be dynamically updated and adjusted based on the dimensional knowledge data, and the resulting updated dimension filling weights will be used for user retrieval through the output analysis unit 114. When a user searches, output knowledge entries are generated based on the search elements and search dimensions in the corresponding search request. By matching and analyzing the output knowledge entries with updated dimension filling weights, the reliability of each output knowledge entry is determined. This reliability can then help users better judge the accuracy of search results, avoiding the huge cognitive error caused by blindly trusting output search results based on a small amount of dimensional knowledge when there is a lack of dimensional knowledge.

[0024] In the knowledge governance retrieval system based on enterprise classification standards of the present invention, the conflict analysis unit 112 may further include: a document parsing subunit, used to parse the newly added classification standard document to obtain multiple dimensions to be classified and corresponding new dimension value ranges; a dimension comparison subunit, used to compare the dimensions to be classified and the existing dimension combinations corresponding to each existing classification standard document according to the new dimension value range, so as to calculate the dimension matching degree between the dimension to be classified and each existing dimension combination; and a first dimension partitioning subunit, used to, when the dimension matching degree is greater than the matching degree threshold, use the newly added classification standard document as a correction file of the corresponding existing classification standard document, and partition the dimension to be classified into the first existing classification dimension pair in the existing dimension combination. The system comprises: a consistent dimension, a replacement dimension that corresponds to the second existing classification dimension in the existing dimension combination but has a dimension difference, and a new dimension that does not correspond to any of the existing classification dimensions in the existing dimension combination. The process involves deleting all existing classification dimensions in the existing dimension combination except for the first existing classification dimension, replacing the first existing classification dimension with the replacement dimension, and adding the new dimension to the existing dimension combination to obtain the updated classification dimension; a second dimension partitioning subunit, used to use all unclassified dimensions corresponding to the new classification standard file as the updated classification dimension when the dimension matching degree is less than the matching degree threshold; and a data detection subunit, used to perform knowledge data detection on the repository based on the updated classification dimension to obtain the initial dimension filling weights corresponding to the updated classification dimension.

[0025] During the initial dimension filling weight determination process, the pre-defined category dimensions in the pre-set dimension repository are used as a basis. The file parsing subunit performs a dimension content comparison query on the newly added category standard file to find multiple pre-defined category dimensions as the dimensions to be classified. The new dimension value range for each dimension to be classified is then output. This value range can represent the content filtering range or value range under the corresponding category dimension. After determining the new dimension value range, the dimension comparison subunit compares the dimensions to be classified with the existing dimension combinations corresponding to each existing category standard file using dimension keywords. This determines the dimension matching degree between each dimension to be classified and each existing dimension combination. By performing a threshold test on the dimension matching degree, when the dimension matching degree is greater than the threshold, the first dimension partitioning subunit uses the newly added category standard file as a correction file for the corresponding existing category standard file. This allows for updating the existing category dimensions through correction. Specifically, the unclassified dimensions can be divided into consistent dimensions that correspond to and are identical to the first existing classification dimension in the existing dimension combination, replacement dimensions that correspond to the second existing classification dimension in the existing dimension combination but have dimensional differences, and new dimensions that do not correspond to any of the existing classification dimensions in the existing dimension combination. Furthermore, by deleting all existing classification dimensions except the first existing classification dimension from the existing dimension combination, replacing the first existing classification dimension with replacement dimensions, and adding new dimensions to the existing dimension combination, the existing classification dimensions can be quickly updated to obtain the updated classification dimensions. When the dimension matching degree is less than the matching degree threshold, all unclassified dimensions corresponding to the new classification standard file can be directly used as updated classification dimensions through the second dimension division subunit. Finally, the data detection subunit uses the determined updated classification dimensions to perform knowledge data detection on the repository, efficiently and accurately querying and calculating the initial dimension filling weight corresponding to each updated classification dimension, and dynamically adjusting the initial dimension filling weight based on new knowledge, thereby updating the asset knowledge filling status corresponding to each classification dimension in real time, ensuring the knowledge coverage of the search results, and improving the reliability of the search result output.

[0026] In the conflict analysis unit 112, the dimension comparison subunit may further include: a semantic comparison module, used to compare the semantic similarity of the dimension to be classified with each existing classification dimension in each existing dimension combination under the corresponding professional dimension to obtain the dimension similarity; a first dimension calculation module, used to take the dimension to be classified that has a dimension similarity greater than the first similarity threshold with the existing classification dimensions as consistent dimensions, obtain the first consistent dimension matching degree corresponding to each consistent dimension, compare the similarity of the first new dimension value range corresponding to the consistent dimension with the existing dimension value range corresponding to the corresponding existing classification dimension to obtain the second consistent dimension matching degree, and perform weighted fusion of the first consistent dimension matching degree and the second consistent dimension matching degree to calculate the first dimension matching degree between the dimension to be classified and each existing dimension combination; and a second dimension calculation module, used to take the dimension to be classified that has a dimension similarity less than the first similarity threshold and greater than the second similarity threshold with the existing classification dimensions as replacement dimensions, and transform according to the dimension similarity and the first matching degree. The system employs a coefficient to obtain the first replacement dimension matching degree for each replacement dimension. It then compares the similarity of the second new dimension value range corresponding to the replacement dimension with the existing dimension value range corresponding to the corresponding existing classification dimension to obtain the second replacement dimension matching degree. The first and second replacement dimension matching degrees are weighted and fused to calculate the second dimension matching degree between the dimension to be classified and each combination of existing dimensions. The first matching conversion coefficient is obtained by averaging the first conversion coefficient corresponding to the existing classification dimension and the second conversion coefficient corresponding to the replacement dimension. A third dimension calculation module is used when the dimensional similarity between the existing classification dimension and the dimension to be classified is lower than the second similarity threshold. In this case, the corresponding dimension to be classified is treated as a new dimension, and the matching degree reduction is obtained based on the third dimension weight of the new dimension in the dimension to be classified and the matching degree reduction coefficient. Finally, a dimension fusion module is used to calculate the dimensional matching degree between the dimension to be classified and each combination of existing dimensions based on the first dimension matching degree, the second dimension matching degree, and the matching degree reduction.

[0027] Preferably, the formula for calculating the dimensional matching degree can be expressed as: ; in, Indicates the degree of dimensional matching. This represents the first matching weight corresponding to the first dimension matching degree. This represents the second matching weight corresponding to the second dimension matching degree. This represents the third matching weight corresponding to the reduction in matching degree. This indicates the matching degree of the first consistency dimension. This indicates the matching degree of the second consistency dimension. This represents the weight of the first dimension when each dimension to be classified is a consistent dimension. This indicates the first-dimensional keyword corresponding to the replacement dimension. This indicates the second-dimensional keywords corresponding to the existing category dimensions. Indicates dimensional similarity. Indicates the first conversion factor. This represents the second conversion factor. Indicates the matching degree of the second replacement dimension. This represents the weight of the second dimension when each dimension to be classified is used as a replacement dimension. This represents the decreasing coefficient of matching degree. This represents the weight of the third dimension when each dimension to be classified is a newly added dimension.

[0028] In passing When determining dimensional similarity, cosine similarity can be used. and The similarity value between them can be used as the corresponding dimensional similarity. Of course, it can also be the dimensional similarity calculated in other ways.

[0029] When calculating dimensional matching scores, the semantic comparison module can be used to compare the semantic similarity of each dimension to be classified with the existing classification dimensions in each combination of existing dimensions under the corresponding professional dimension to obtain the corresponding dimensional similarity. This professional dimension can be determined by the professional attributes of the newly added classification standard file corresponding to the dimension to be classified. After obtaining the dimensional similarity, the first dimensional calculation module can be used to divide the dimensions to be classified into consistent dimensions with existing classification dimensions whose dimensional similarity is greater than the first similarity threshold, and the corresponding first consistent dimension matching score can be calculated based on each consistent dimension. Simultaneously, the similarity between the first new dimension value range corresponding to the consistency dimension and the existing dimension value range corresponding to the corresponding existing classification dimension is compared to obtain the second consistency dimension matching degree. The matching scores of the first and second consistent dimensions are combined with the weight of the first dimension when each dimension to be classified is a consistent dimension. We perform weighted fusion to calculate the first-dimensional matching degree between all consistent dimensions and each existing dimension combination. The second-dimensional calculation module can achieve dimensional similarity based on the replacement dimension by using the dimension to be classified that has a similarity to the existing classification dimensions that is less than the first similarity threshold but greater than the second similarity threshold. Combined with the corresponding first matching transformation coefficient To calculate the matching degree of the first replacement dimension for each replacement dimension. The similarity score of the second replacement dimension is calculated by comparing the value range of the second new dimension corresponding to the replacement dimension with the value range of the existing dimension corresponding to the existing classification dimension. Finally, the matching scores of the first and second replacement dimensions are combined with the weight of the second dimension when each unclassified dimension is a replacement dimension. We perform weighted fusion to calculate the second dimension matching degree between all replacement dimensions and each existing dimension combination. When the dimensional similarity between the existing classification dimension and the dimension to be classified is lower than the second similarity threshold, the corresponding dimension to be classified can be directly added as a new dimension through the third dimension calculation module, and the new dimension can be added based on its third-dimensional weight within the dimension to be classified. and matching degree decreasing coefficient Result in the reduction in matching degree for all newly added dimensions. Finally, the dimension fusion module utilizes the first matching weight corresponding to the first dimension matching degree. The second matching weight corresponding to the second dimension matching degree and the third matching weight corresponding to the matching degree reduction Matching degree of the first dimension Second dimension matching degree and matching degree reduction Weighted fusion is performed to calculate the dimensionality matching degree between the dimension to be classified and each existing dimension combination, i.e., the formula is: Among them, the first matching weight Second matching weight Third matching weight All are pre-calibrated based on empirical values, with the first dimension weighting. Second dimension weight and the third dimension weight The first conversion coefficient is also obtained by pre-calibrating empirical values. Second conversion coefficient Similarly, this is based on empirical values. From the above, it can be seen that when the dimensional similarity between the existing classification dimensions and the dimension to be classified is below the similarity threshold, the dimension to be classified can be considered a new dimension; conversely, when all dimensions to be classified are new dimensions relative to the existing classification dimensions, all new dimensions are considered as updated classification dimensions.

[0030] In the conflict analysis unit 112, the data detection subunit may further include: a first weight calculation module, used to use the existing dimension filling weight of the existing classification dimension corresponding to the consistent dimension as the initial dimension filling weight when the updated classification dimension is a consistent dimension; a first knowledge output module, used to perform semantic deviation analysis on the replacement dimension and the existing classification dimension when the updated classification dimension is a replacement dimension, to obtain the deleted semantic range and the added semantic range, to control the deletion of the existing dimension knowledge of the existing classification dimension corresponding to the replacement dimension according to the deleted semantic range, and to perform knowledge search on the repository according to the added semantic range to obtain the updated dimension knowledge; a second knowledge output module, used to perform content search on the repository according to the added dimension when the updated classification dimension is a new dimension to obtain the updated dimension knowledge; and a second weight calculation module, used to perform dimension filling degree analysis on the updated dimension knowledge to obtain the initial dimension filling weight corresponding to the updated classification dimension.

[0031] When performing data inspection on the repository, since the consistency dimension is the same as the existing classification dimension, the existing dimension fill weights of the existing classification dimension corresponding to the consistency dimension can be directly used as the initial dimension fill weights through the first weight calculation module. For replacement dimensions, the first knowledge output module can first perform semantic deviation analysis on the replacement dimension and the existing classification dimension to separate the deleted semantic range and the added semantic range. Then, based on the added semantic range, knowledge is searched in the repository to obtain the updated dimension knowledge. For new dimensions, the second knowledge output module can directly perform content search in the repository to obtain the updated dimension knowledge. Finally, the second weight calculation module performs dimension fill degree analysis on the updated dimension knowledge to adjust the initial dimension fill weights corresponding to the updated classification dimension, so as to realize the dynamic adjustment of the dimension fill weights based on the classification dimension adjustment, which serves as the basis for calculating the updated dimension fill weights.

[0032] In the data detection subunit, the first knowledge output module performs semantic deviation analysis on the replacement dimension and the existing classification dimension to obtain the deleted semantic range and the added semantic range. This process may further include: a search submodule, used to search for the set of first keywords corresponding to the semantics of the first dimension keywords of the replacement dimension, and the set of second keywords corresponding to the semantics of the second dimension keywords of the existing classification dimension; a difference comparison submodule, used to compare the semantic differences between the first keyword set and the second keyword set to obtain the first filter keywords that correspond only to the replacement dimension and the second filter keywords that correspond only to the existing classification dimension; a deletion range output submodule, used to semantically fuse the first semantics of all the first filter keywords to obtain the deleted semantic range; and an added range output submodule, used to semantically fuse the second semantics corresponding to all the second filter keywords to obtain the added semantic range.

[0033] In determining the semantic scope for deletion and addition, the search submodule first finds the set of first keywords corresponding to the semantics of the first dimension keywords of the replacement dimension, and the set of second keywords corresponding to the semantics of the second dimension keywords of the existing classification dimension. Then, the difference comparison submodule compares the semantic differences between the first and second keyword sets, obtaining the first keywords that only correspond to the replacement dimension as the first filter keywords, and the second keywords that only correspond to the existing classification dimension as the second filter keywords. In other words, the first filter keywords do not correspond to the existing classification dimension, and the second filter keywords do not correspond to the replacement dimension. Subsequently, the deletion scope output submodule semantically fuses all the first semantics corresponding to all the first filter keywords to obtain the deleted semantic scope, and the addition scope output submodule semantically fuses all the second semantics corresponding to all the second filter keywords to obtain the added semantic scope.

[0034] For example, when the second-dimensional keyword corresponding to an existing classification dimension is "architecture," and the first-dimensional keyword corresponding to the replacement dimension is "intelligent building," we can find second-dimensional keywords such as "architectural design," "architectural construction," "construction drawings," "building codes," "energy-saving design," and "site planning" in the second-dimensional keyword set, and first-dimensional keywords such as "BIM (Building Information Modeling)," "Internet of Things," "building automation," "intelligent security," "energy-saving optimization," and "digital twin" in the first-dimensional keyword set. When comparing the semantic differences between the first-dimensional keyword set and the second-dimensional keyword set, the semantics of "energy-saving design" and "energy-saving optimization" can be directly treated as the same. In addition, we can also find that the first-dimensional filtering keywords that correspond only to the replacement dimension "intelligent building" and do not correspond to the existing classification dimension "architecture" can be "BIM (Building Information Modeling)," "Internet of Things," "building automation," "intelligent security," and "digital twin." Therefore, we can determine the corresponding deletion semantic range based on the corresponding semantics of these first-dimensional filtering keywords. The second filter keywords, which correspond only to the existing classification dimension "Architecture" and do not correspond to the replacement dimension "Intelligent Building," can be "Architectural Design," "Architectural Construction," "Construction Drawings," and "Building Codes." Therefore, the corresponding new semantic scope can be determined based on the semantic meaning of these second filter keywords. Alternatively, it can also be a division of the semantic scope for deletion and addition of other first-dimensional and second-dimensional keywords.

[0035] As can be seen from the above, by utilizing different dimensional content filling strategies, the accuracy of the filled content data can be effectively guaranteed, while also improving filling efficiency. This enables distributed and efficient filling of repository data based on new classification standards, in order to calculate dimensional filling weights.

[0036] In the data detection subunit, the second weight calculation module may further include: a value range filtering submodule, used to filter the difference value ranges of the updated dimension knowledge to obtain the difference value ranges corresponding to the difference knowledge items, and based on the difference value ranges, obtain multiple coverage value range ranges; a quantity filling calculation submodule, used to obtain the quantity filling degree according to the range length corresponding to each coverage value range and the number of difference knowledge items within the range length; a knowledge completeness calculation submodule, used to obtain the corresponding knowledge completeness according to the dimension value range and coverage value range corresponding to the updated classification dimension; and a weight output submodule, used to weight and fuse the quantity filling degree and knowledge completeness to obtain the initial dimension filling weight corresponding to the updated classification dimension.

[0037] The formula for calculating the initial dimension fill weights can be expressed as: ; in, This indicates that the initial dimension is filled with weights. This indicates the first fill weight corresponding to the quantity fill degree. This represents the second filling weight corresponding to the knowledge completeness. This indicates the number of knowledge entries representing differences within the specified range. Indicates the conversion factor. This indicates the range length corresponding to each covered value range. Indicates the number of values ​​covered. This indicates the total range length of the dimension value range corresponding to the updated category dimension.

[0038] During the initial dimension filling weight calculation process, the value range filtering submodule can filter the difference value ranges of the updated dimension knowledge to find the difference value ranges corresponding to the difference knowledge entries. Based on the coverage of the difference value ranges, multiple different coverage value ranges can be obtained. Subsequently, the quantity filling calculation submodule uses the range length corresponding to each coverage value range. The number of knowledge entries with differences within the range length The quantity conversion coefficient is set based on empirical values, and the quantity is calculated based on the coverage range. The average value is used to calculate the quantity fill rate. Subsequently, the knowledge integrity calculation submodule utilizes the sum of the range lengths of the dimensional value ranges corresponding to the updated classification dimensions. and the total range length of the covered value range To calculate the corresponding knowledge completeness. Finally, the quantity fill rate and knowledge completeness can be combined with the first fill weight preset by experience values ​​through the weight output submodule. Second filling weight We perform weighted fusion to obtain the initial dimension filling weights corresponding to the updated classification dimensions, as expressed by the formula: Using the above method, the initial dimension filling weights corresponding to the updated classification dimension can be accurately determined from two main aspects: the number of knowledge entries covered by each covered value range and the remaining length of the covered value range.

[0039] As can be seen from the above, the initial dimension filling weights include quantity filling degree and knowledge completeness. In the knowledge governance retrieval system based on enterprise classification standards of the present invention, the knowledge governance unit 113 may further include: a knowledge storage subunit, used to parse new knowledge and obtain the dimension knowledge data corresponding to each updated classification dimension in the new knowledge for storage in the repository; an analysis and calculation subunit, used to update the quantity filling degree and knowledge completeness according to the dimension knowledge data to obtain the updated quantity filling degree and updated knowledge completeness; and a weight update subunit, used to obtain the updated dimension filling weights according to the updated quantity filling degree and updated knowledge completeness.

[0040] When calculating the updated dimension filling weights, the knowledge storage subunit can parse the new knowledge and retrieve the dimensional knowledge data corresponding to each updated category dimension for storage in the repository. Simultaneously, the analysis and calculation subunit updates the quantity filling degree and knowledge completeness based on the dimensional knowledge data to obtain the updated quantity filling degree and updated knowledge completeness. Specifically, for the quantity filling degree, the length of the coverage range can be determined by identifying the coverage range of the dimensional threshold of the corresponding category dimension in the new knowledge. The increase in the number of knowledge entries and the number of entries with differences within the range length. Adding cases allows for a more flexible approach, which can then be based on the updated range length. and number of entries By updating the weights of the sub-units, the weight calculation formula can be filled in based on the initial dimensions mentioned above. The dimension filling weights are dynamically adjusted to update the dimension filling weights. .

[0041] In the knowledge governance retrieval system based on enterprise classification standards of the present invention, the output analysis unit 114 includes: a retrieval output subunit, used to perform a mixed retrieval of the repository in response to a user's retrieval request to obtain output knowledge entries; a reliability calculation subunit, used to obtain the output reliability of each output knowledge entry based on the update dimension filling weight of each update classification dimension corresponding to the output knowledge entry and the importance coefficient of each update classification dimension; and a sorting output subunit, used to sort the output knowledge entries according to the output reliability to output retrieval results.

[0042] The formula for calculating output reliability is: ; in, This indicates the output reliability of each output knowledge entry. This represents the update dimension fill weight for each output knowledge entry. This represents the importance coefficient of each output knowledge item.

[0043] During the output of search results, the retrieval output subunit can receive or actively collect user query requests. It can then perform a direct mixed search using the search elements corresponding to the request and the user-selected search dimensions to retrieve the output knowledge entries. To ensure that each output knowledge entry accurately reflects the reliability of the data stored in the current memory, the reliability calculation subunit can fill in the weights based on the update dimensions of each updated classification dimension corresponding to the output knowledge entry. And the importance coefficient of each updated classification dimension After multiplication and summation, the output reliability of each output knowledge item is obtained, i.e. Finally, by sorting the output sub-units, all output knowledge items are sequentially sorted in descending order of output reliability to serve as search results. This helps users filter out output knowledge items with higher reliability as trustworthy knowledge items, avoiding blindly choosing and trusting output search results based on a small amount of dimensional knowledge due to a lack of dimensional knowledge, thus improving the reliability of search results.

[0044] Please see Figure 2 The present invention also provides a retrieval method for knowledge governance based on enterprise classification standards, including: Step S10: Obtain the newly added classification standard file within the enterprise through acquisition unit 111; Step S20: The conflict analysis unit 112 performs conflict analysis based on the newly added classification standard file and the existing classification dimensions to obtain the updated classification dimensions and the initial dimension filling weights corresponding to the updated classification dimensions. Step S30: When new knowledge is received by the knowledge governance unit 113, the new knowledge is governed according to the updated classification dimension to obtain dimension knowledge data for storage in the repository, and the initial dimension filling weight is updated according to the dimension knowledge data to obtain the updated dimension filling weight. Step S40: In response to the user's query retrieval request, the output analysis unit 114 performs a hybrid retrieval and matching analysis on the repository based on dimensional knowledge data and updated dimension filling weights to output retrieval results. The retrieval request includes retrieval elements and retrieval dimensions.

[0045] In summary, this invention discloses a knowledge governance retrieval system and method based on enterprise classification standards. By utilizing newly added classification standard files and combining them with existing classification dimensions for conflict analysis, updated classification dimensions are generated, and the initial dimension filling weights corresponding to the updated classification dimensions are output. Furthermore, when new knowledge is updated, the received new knowledge can be governed using updated classification dimensions to generate and store dimensional knowledge data corresponding to different updated classification dimensions. The initial dimension filling weights are also dynamically updated and adjusted based on the dimensional knowledge data, and the resulting updated dimension filling weights are used for user retrieval. During user retrieval, output knowledge entries are generated based on the search elements and search dimensions in the corresponding search request. By performing matching analysis on the output knowledge entries using updated dimension filling weights, the reliability of each output knowledge entry is determined. This reliability helps users better judge the accuracy of search results, avoiding blindly trusting output search results based on limited knowledge when dimensional knowledge is insufficient, thus preventing significant cognitive errors. Therefore, this invention effectively overcomes the various shortcomings of existing technologies and has high industrial application value.

[0046] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A search system for knowledge governance based on enterprise taxonomy criteria, characterized in that, include: The acquisition unit is used to acquire new classification standard files within the enterprise. The conflict analysis unit is used to perform conflict analysis based on the newly added classification standard file and the existing classification dimensions to obtain the updated classification dimensions and the initial dimension filling weights corresponding to the updated classification dimensions. The knowledge governance unit is used to govern the new knowledge according to the updated classification dimension when new knowledge is received, so as to obtain dimension knowledge data for storage in the repository, and update the initial dimension filling weight according to the dimension knowledge data to obtain the updated dimension filling weight. as well as The output analysis unit is used to respond to the user's retrieval request by performing a hybrid retrieval and matching analysis on the repository based on the dimensional knowledge data and the updated dimension fill weights, so as to output retrieval results. The retrieval request includes retrieval elements and retrieval dimensions.

2. The retrieval system for knowledge governance based on enterprise classification standards according to claim 1, characterized in that, The conflict analysis unit includes: The file parsing subunit is used to parse the newly added classification standard file to obtain the corresponding multiple dimensions to be classified and the corresponding new dimension value range; The dimension comparison subunit is used to compare the dimension to be classified and the existing dimension combinations corresponding to each existing classification standard file according to the new dimension value range, so as to calculate the dimension matching degree between the dimension to be classified and each existing dimension combination. The first dimension partitioning subunit is used to, when the dimension matching degree is greater than the matching degree threshold, use the newly added classification standard file as the correction file of the corresponding existing classification standard file, divide the dimension to be classified into a consistent dimension that corresponds to and is consistent with the first existing classification dimension in the existing dimension combination, a replacement dimension that corresponds to the second existing classification dimension in the existing dimension combination but has a dimension difference, and a new dimension that does not correspond to any of the existing classification dimensions in the existing dimension combination, delete all existing classification dimensions in the existing dimension combination except for the first existing classification dimension, replace the first existing classification dimension with the replacement dimension, and add the new dimension to the existing dimension combination to obtain the updated classification dimension; The second dimension partitioning subunit is used to, when the dimension matching degree is less than the matching degree threshold, use all the dimensions to be classified corresponding to the newly added classification standard file as the updated classification dimensions; and The data detection subunit is used to perform knowledge data detection on the repository according to the updated classification dimension in order to obtain the initial dimension filling weight corresponding to the updated classification dimension.

3. The retrieval system for knowledge governance based on enterprise classification standards according to claim 2, characterized in that, The dimension comparison subunit includes: The semantic comparison module is used to compare the semantic similarity of the dimension keywords under the corresponding professional dimension with each existing classification dimension in each combination of existing dimensions to obtain the dimension similarity. The first dimension calculation module is used to take the dimension to be classified that has a dimension similarity greater than a first similarity threshold with the existing classification dimensions as a consistent dimension, obtain a first consistent dimension matching degree corresponding to each consistent dimension, compare the value range of the first new dimension corresponding to the consistent dimension with the value range of the existing dimension corresponding to the corresponding existing classification dimension to obtain a second consistent dimension matching degree, and perform a weighted fusion of the first consistent dimension matching degree and the second consistent dimension matching degree to calculate the first dimension matching degree between the dimension to be classified and each combination of existing dimensions; The second dimension calculation module is used to select the unclassified dimension whose similarity to the existing classification dimension is less than the first similarity threshold and greater than the second similarity threshold as the replacement dimension; obtain the first replacement dimension matching degree corresponding to each replacement dimension based on the dimension similarity and the first matching conversion coefficient; compare the similarity between the second new dimension value range corresponding to the replacement dimension and the existing dimension value range corresponding to the corresponding existing classification dimension to obtain the second replacement dimension matching degree; and perform a weighted fusion of the first replacement dimension matching degree and the second replacement dimension matching degree to calculate the second dimension matching degree between the unclassified dimension and each combination of existing dimensions; the first matching conversion coefficient is obtained by averaging the first conversion coefficient corresponding to the existing classification dimension and the second conversion coefficient corresponding to the replacement dimension. The third-dimensional calculation module is used to, when the dimensional similarity between the existing classification dimension and the dimension to be classified is lower than the second similarity threshold, treat the corresponding dimension to be classified as a new dimension, and obtain the matching degree reduction based on the third-dimensional weight of the new dimension in the dimension to be classified and the matching degree reduction coefficient; and The dimension fusion module is used to calculate the dimension matching degree between the dimension to be classified and each of the existing dimension combinations based on the first dimension matching degree, the second dimension matching degree, and the matching degree reduction amount.

4. The knowledge governance retrieval system based on enterprise classification standards according to claim 3, characterized in that, The formula for calculating the dimensional matching degree is: ； in, Indicates the degree of dimensional matching. This represents the first matching weight corresponding to the first dimension matching degree. This represents the second matching weight corresponding to the second dimension matching degree. This represents the third matching weight corresponding to the reduction in matching degree. This indicates the matching degree of the first consistency dimension. This indicates the matching degree of the second consistency dimension. This represents the weight of the first dimension when each dimension to be classified is a consistent dimension. This indicates the first-dimensional keyword corresponding to the replacement dimension. This indicates the second-dimensional keywords corresponding to the existing category dimensions. Indicates dimensional similarity. Indicates the first conversion factor. This represents the second conversion factor. Indicates the matching degree of the second replacement dimension. This represents the weight of the second dimension when each dimension to be classified is used as a replacement dimension. This represents the decreasing coefficient of matching degree. This represents the weight of the third dimension when each dimension to be classified is a newly added dimension.

5. The retrieval system for knowledge governance based on enterprise classification standards according to claim 2, characterized in that, The data detection subunit includes: The first weight calculation module is used to use the existing dimension filling weight of the existing classification dimension corresponding to the consistent dimension as the initial dimension filling weight when the updated classification dimension is the consistent dimension. The first knowledge output module is used to perform semantic deviation analysis between the replacement dimension and the existing classification dimension when the updated classification dimension is the replacement dimension, to obtain the deleted semantic range and the added semantic range, to control the deletion of the existing dimension knowledge of the existing classification dimension corresponding to the replacement dimension according to the deleted semantic range, and to perform knowledge search on the repository according to the added semantic range to obtain the updated dimension knowledge. The second knowledge output module is used to perform a content search on the repository based on the new dimension when the updated classification dimension is the new dimension, so as to obtain the updated dimension knowledge; and The second weight calculation module is used to perform dimension filling degree analysis on the updated dimension knowledge to obtain the initial dimension filling weight corresponding to the updated classification dimension.

6. The retrieval system for knowledge governance based on enterprise classification standards according to claim 5, characterized in that, The first knowledge output module performs semantic deviation analysis on the replacement dimension and the existing classification dimension to obtain the semantic range for deletion and addition, including: The search submodule is used to search for the first set of keywords that corresponds to the semantic meaning of the first dimension keywords of the replacement dimension, and the second set of keywords that corresponds to the semantic meaning of the second dimension keywords of the existing classification dimension. The difference comparison submodule is used to perform a semantic difference comparison between the first keyword set and the second keyword set to obtain a first filter keyword that corresponds only to the replacement dimension and a second filter keyword that corresponds only to the existing classification dimension. The deletion range output submodule is used to semantically fuse the first semantics of all the first filtered keywords to obtain the deletion semantic range; and A new range output submodule is added to perform semantic fusion on the second semantics corresponding to all the second filtering keywords to obtain the new semantic range.

7. The knowledge governance retrieval system based on enterprise classification standards according to claim 5, characterized in that, The second weight calculation module includes: The value range filtering submodule is used to filter the difference value range of the updated dimension knowledge, obtain the difference value range corresponding to the difference knowledge item, and obtain multiple coverage value ranges based on the difference value range. The quantity filling calculation submodule is used to obtain the quantity filling degree based on the range length corresponding to each of the coverage value ranges and the number of the difference knowledge entries within the range length; The knowledge completeness calculation submodule is used to obtain the corresponding knowledge completeness based on the dimension value range corresponding to the updated classification dimension and the coverage value range; and The weight output submodule is used to weight and fuse the quantity filling degree and the knowledge completeness to obtain the initial dimension filling weight corresponding to the updated classification dimension. The formula for calculating the initial dimension padding weight is: ； in, This indicates that the initial dimension is filled with weights. This indicates the first fill weight corresponding to the quantity fill degree. This represents the second filling weight corresponding to the knowledge completeness. This indicates the number of knowledge entries representing differences within the specified range. Indicates the conversion factor. This indicates the range length corresponding to each covered value range. Indicates the number of covered value ranges. This indicates the total range length of the dimension value range corresponding to the updated category dimension.

8. The retrieval system for knowledge governance based on enterprise classification standards according to claim 1, characterized in that, The initial dimension filling weights include quantity filling degree and knowledge completeness; The knowledge governance unit includes: The knowledge storage subunit is used to parse the new knowledge and obtain the dimension knowledge data corresponding to each of the updated classification dimensions in the new knowledge for storage in the repository. An analysis and calculation subunit is used to update the quantity fill rate and the knowledge completeness based on the dimensional knowledge data to obtain updated quantity fill rate and updated knowledge completeness; and The weight update subunit is used to obtain the update dimension filling weight based on the update quantity fill degree and the update knowledge completeness.

9. The knowledge governance retrieval system based on enterprise classification standards according to claim 1, characterized in that, The output analysis unit includes: The retrieval output subunit is used to perform a mixed retrieval on the repository in response to a user's retrieval request, and to obtain output knowledge entries. A reliability calculation subunit is used to obtain the output reliability of each output knowledge entry based on the update dimension fill weights of each update classification dimension corresponding to the output knowledge entry and the importance coefficient of each update classification dimension; and The sorting output subunit is used to sort the output knowledge items according to the output reliability in order to output the retrieval results; The formula for calculating the output reliability is as follows: ； in, This indicates the output reliability of each output knowledge entry. This represents the update dimension fill weight for each output knowledge entry. This represents the importance coefficient of each output knowledge item.

10. A retrieval method for knowledge governance based on enterprise classification standards, characterized in that, include: The newly added classification standard documents within the enterprise are obtained through the acquisition unit; The conflict analysis unit performs conflict analysis based on the newly added classification standard file and the existing classification dimensions to obtain the updated classification dimensions and the initial dimension filling weights corresponding to the updated classification dimensions. When new knowledge is received by the knowledge governance unit, the new knowledge is governed according to the updated classification dimension to obtain dimensional knowledge data for storage in the repository, and the initial dimension filling weight is updated according to the dimensional knowledge data to obtain the updated dimension filling weight. In response to a user's query request, the output analysis unit performs a hybrid retrieval and matching analysis on the repository based on the dimensional knowledge data and the updated dimension fill weights to output retrieval results. The retrieval request includes retrieval elements and retrieval dimensions.

Images