Method, device, equipment and medium for constructing employee knowledge mastery map

Abstract

The present invention relates to the field of data processing technology, and discloses a method, device, equipment and medium for constructing an employee knowledge mastery map, including: obtaining training courseware data and job classification data, as well as employee knowledge mastery related data; preprocessing the training courseware data to obtain training feature data; vector construction processing of the training feature data to obtain a training feature vector sample set; clustering processing of the training feature vector sample set to obtain K training clustering results; matching the K training clustering results with the job classification data to generate a job resource map; generating an employee knowledge mastery map based on the job resource map and employee knowledge mastery related data. By implementing the method provided in this application, the display of employee knowledge mastery can be realized quickly and effectively, so as to understand the shortcomings of each employee in a targeted manner, which is conducive to each employee to carry out targeted learning.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device, equipment and medium for constructing an employee knowledge mastery map. Background Art

[0002] In the process of corporate employee training, in order to accurately represent the employees' mastery of job knowledge points, knowledge graph mastery spectrum is usually used for representation. Knowledge gaps are knowledge points that have not been learned or mastered during employee training and learning, and are a reflection of the lack of employee knowledge. Many employees have more or less knowledge gaps in training and learning, and it is difficult to discover them in time. At present, with the increasing application of knowledge graphs in the field of education and learning, it has become a trend to use knowledge graphs to visualize employee knowledge mastery. However, how to effectively combine knowledge graphs with employee knowledge mastery to achieve the purpose of quickly and effectively understanding the employee's knowledge mastery has become a problem that needs to be solved urgently. Summary of the invention

[0003] The present invention provides a method, device, equipment and medium for constructing an employee knowledge mastery map to solve the technical problem of effectively combining the knowledge map with the employee's knowledge mastery.

[0004] In the first aspect, a method for constructing an employee knowledge mastery map is provided, including:

[0005] Obtain training courseware data, job classification data, and employee knowledge mastery data;

[0006] Preprocess the training courseware data to obtain training feature data;

[0007] Perform vector construction processing on the training feature data to obtain a training feature vector sample set;

[0008] Cluster the training feature vector sample set to obtain K training clustering results;

[0009] Match the K training clustering results with the job classification data to generate a job resource map;

[0010] Generate an employee knowledge mastery map based on the job resource map and employee knowledge mastery related data.

[0011] In a second aspect, a device for constructing an employee knowledge mastery map is provided, comprising:

[0012] The acquisition module is used to obtain training courseware data, job classification data, and employee knowledge mastery data;

[0013] The first processing module is used to pre-process the training courseware data to obtain training feature data;

[0014] The second processing module is used to perform vector construction processing on the training feature data to obtain a training feature vector sample set;

[0015] The third processing module is used to perform clustering processing on the training feature vector sample set to obtain K training clustering results;

[0016] The fourth processing module is used to match the K training clustering results with the job classification data to generate a job resource map;

[0017] The fifth processing module is used to generate an employee knowledge mastery map based on the job resource map and employee knowledge mastery related data.

[0018] In a third aspect, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned employee knowledge mastery map construction method when executing the computer program.

[0019] In a fourth aspect, a computer-readable storage medium is provided, which stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned employee knowledge mastery map construction method are implemented.

[0020] In the scheme implemented by the above-mentioned employee knowledge mastery map construction method, device, equipment and medium, by obtaining training courseware data and job classification data, as well as employee knowledge mastery related data, the training courseware data can be pre-processed to obtain training feature data, and the training feature data can be vector constructed to obtain a training feature vector sample set, so that the training feature vector sample set can be clustered to obtain K training clustering results, and the K training clustering results can be further matched with the job classification data to generate a job resource map, and then the employee knowledge mastery map can be generated based on the job resource map and employee knowledge mastery related data, which can quickly and effectively realize the display of employee knowledge mastery, so as to understand the shortcomings of each employee in a targeted manner, which is conducive to each employee to carry out targeted learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0022] Figure 1 It is a flow chart of a method for constructing an employee knowledge mastery map in one embodiment of the present invention;

[0023] Figure 2 It is a structural schematic diagram of a device for constructing an employee knowledge mastery map in one embodiment of the present invention;

[0024] Figure 3 is a schematic diagram of a structure of a computer device in one embodiment of the present invention;

[0025] Figure 4 It is another structural schematic diagram of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] Currently, with the increasing application of knowledge graphs in the field of education and learning, it has become a trend to use knowledge graphs to visually present employees' knowledge mastery. Combining knowledge graphs to visualize and analyze employees' knowledge deficiencies after training and learning can facilitate the discovery of learning problems and make training more targeted. This application clusters the content of the training courseware and generates a resource map for the corresponding positions based on job classification to determine the knowledge points that need to be mastered for each position. By integrating the employees' mastery of each knowledge point in the course and constructing a knowledge mastery map, it can quickly and effectively display the knowledge points that employees have not mastered, allowing employees to conduct targeted learning.

[0028] See also Figure 1 As shown, Figure 1 A flowchart of a method for constructing an employee knowledge mastery map provided by an embodiment of the present invention includes the following steps:

[0029] S10: Obtain training courseware data and job classification data, as well as employee knowledge mastery data;

[0030] Among them, training courseware data may include but is not limited to internal training material data, such as relevant data in the company's internal training library and the company's online learning platform, and external training material data, such as training courseware data provided by external professional training institutions when cooperating with external professional training institutions, or general industry training courseware data provided by industry associations, or training resources publicly available on the Internet, etc. This application does not impose any restrictions on this.

[0031] Position classification data may include but is not limited to the position categories and hierarchical relationships of various departments of an enterprise. The above position classification data can be obtained through the enterprise's organizational structure, departmental function descriptions, etc. For example, through the organizational structure chart of a financial enterprise, it can be learned that the enterprise has different types of position classifications such as front-end business departments (such as account manager positions), middle-end risk control departments (such as risk assessment positions), and back-end operation support departments (such as financial accounting positions).

[0032] The employee knowledge mastery data can be used to indicate the employee's mastery of the knowledge in various knowledge areas required for the job. The employee knowledge mastery data can be obtained through knowledge tests and assessments, such as skill assessments, etc.; and / or through work performance assessments, such as project outcome assessments, etc.; and / or through employee self-assessments and training records, etc., and this application does not limit this.

[0033] S20: Preprocess the training courseware data to obtain training feature data.

[0034] Among them, preprocessing of training courseware data may include but is not limited to Chinese word segmentation of training courseware data, removing used words to create stop files, and constructing a bag-of-words space to count the word sets of all documents, etc. This application does not impose any restrictions on this.

[0035] For Chinese word segmentation, you can read the text content in the training courseware data and perform Chinese word segmentation operations on the extracted text one by one. You can use professional Chinese word segmentation tools, such as the Jieba word segmentation tool, to accurately segment sentences and generate data suitable for text analysis. For example, for the sentence "This is a training courseware", the word segmentation tool can segment it into: "This is", "A", and "Training Courseware".

[0036] For the construction of stop words, you can collect common stop word lists, or organize stop words according to the characteristics of the field involved in the training courseware data. For example, if the training courseware is in the field of computer technology, you can add common auxiliary words in technical terms, such as "based on" and "through", to the stop word file. The constructed stop words can be saved in a file, one stop word per line, to facilitate subsequent processing programs to read and filter.

[0037] By summarizing the words after segmentation and removal of stop words, we can form a set of different words that appear in all documents in the training courseware data, and this set is the bag-of-words space. For example, if the word sets obtained after processing two training courseware data are: {"training", "content", "method"} and {"training", "effect", "evaluation"}, then the bag-of-words space can be {"training", "content", "method", "training", "effect", "evaluation"}.

[0038] S30: Perform vector construction processing on the training feature data to obtain a training feature vector sample set.

[0039] First, we can count the word set to count the number of times the words in the bag-of-words space appear in the courseware, and use weight conversion methods, such as term frequency-inverse document frequency (TF-IDF), to convert the number of word occurrences into a weight that better reflects the importance of the word.

[0040] Based on the vectors (with weights) corresponding to each of the above documents, a training feature vector sample set can be constructed. The training feature vector sample set can be represented in the form of a matrix, where rows correspond to different training courses, columns correspond to words in the bag-of-words space, and the elements in the matrix can be the weights of the words. For example, for a data set containing 3 training courses and 5 words in the bag-of-words space, its training feature vector sample set may be a 3×5 matrix.

[0041] It can be understood that after constructing the feature model, each training courseware data can be represented as a training feature vector, which contains the weight information of the words in the document and reflects the semantic characteristics of the document, which is conducive to more accurate clustering processing based on the training feature vector in subsequent steps.

[0042] S40: clustering the training feature vector sample set to obtain K training clustering results.

[0043] In this application, clustering processing by K-means clustering (K-means) algorithm is used as an example for explanation, which does not constitute a limitation to this application. The K-means algorithm can divide the training feature vector sample set into K clusters so that the similarity of the data points within the cluster is as high as possible, while the similarity between clusters is as low as possible.

[0044] It should be understood that clustering the training feature vector sample set to obtain K training clustering results refers to the process of clustering. In step S40, clustering the training feature vector sample set to obtain K training clustering results includes the following steps:

[0045] S41: Select K samples from the n data objects in the training feature vector sample set as K initial cluster centers; wherein n is the number of training courseware data, and K is the number of job categories in the job classification data;

[0046] S42: Calculate the Euclidean distance between each training feature vector sample in the training feature vector sample set and the K initial cluster centers to obtain a Euclidean distance set;

[0047] S43: according to the Euclidean distance set, each training feature vector sample in the training feature vector sample set is divided into the initial cluster center with the closest Euclidean distance, so as to obtain a training feature cluster cluster set;

[0048] S44: According to the training feature clustering cluster set, the initial cluster center of each cluster is updated and optimized to obtain K training clustering results.

[0049] First, K cluster centers can be randomly initialized. Then, each data point (i.e., the training feature vector corresponding to each training courseware data) can be assigned to the cluster to which the nearest cluster center belongs. Then, the cluster center of each cluster can be recalculated, and the above process of assigning and updating cluster centers can be further repeated until the cluster center no longer changes or the preset number of iterations is reached, and K training clustering results can be obtained.

[0050] It should be understood that updating and optimizing the initial cluster center of each cluster according to the training feature clustering cluster set to obtain K training clustering results refers to the process of iteratively updating the cluster center. Among them, step S44, that is, updating and optimizing the initial cluster center of each cluster according to the training feature clustering cluster set to obtain K training clustering results, includes the following steps:

[0051] S441: clustering the cluster set based on the training features, calculating the mutual distance between the initial cluster centers of each cluster to obtain a mutual distance set;

[0052] S442: According to the training feature clustering cluster set, the surrounding element point density of the initial cluster center of each cluster is calculated to obtain a surrounding element point density set;

[0053] S443: According to the mutual distance set, the surrounding element point density set and the preset conditions, the initial cluster center of each cluster is updated and optimized to obtain K training clustering results.

[0054] Optionally, the process of clustering the cluster set based on the training features and calculating the mutual distances between the initial cluster centers of each cluster to obtain the mutual distance set can be performed using the following formula:

[0055]

[0056] Wherein, d represents the mutual distance between the initial cluster centers of each cluster, which is represented here by the mean of the sum of the distances of all initial cluster centers; K represents the number of training feature cluster clusters in the training feature cluster cluster set; i represents the index of the training feature cluster cluster in the training feature cluster cluster set; j represents the index of the training feature cluster cluster in the training feature cluster cluster set; c i represents the i-th cluster center, c j represents the jth cluster center.

[0057] Optionally, the process of calculating the surrounding element point density of the initial cluster center of each cluster according to the training feature clustering cluster set and obtaining the surrounding element point density set can be performed using the following formula:

[0058]

[0059] Among them, p i Represents the density set of surrounding element points, p i The larger the value, the greater the density; i represents the distance between each training feature vector in the training feature vector sample set; n represents the number of training feature vectors in the training feature vector sample set; z k Represents the Euclidean distance between each training feature vector sample in the training feature vector sample set and the initial cluster center.

[0060] The preset condition can be a pre-set condition for updating and optimizing the initial cluster center. Optionally, the present application can maximize the distance between the center points, that is, each distance between the center points in the distance set takes the maximum value; and make the center peripheral points dense, that is, each peripheral element point density in the peripheral element point density set takes the maximum value, as a preset condition, so as to realize the update optimization of the initial cluster center according to the preset condition, thereby obtaining K training clustering results.

[0061] According to the K training clustering results, it can be found that training courseware data with similar content can be clustered together, which is helpful for classification management, content analysis and further knowledge mining of training courseware. For example, it can be found that the training courseware data in the same cluster may all be about a specific skill training, while another class may be about management knowledge training, which is not limited in this application.

[0062] S50: Match the K training clustering results with the job classification data to generate a job resource map.

[0063] It is understandable that each of the K training clustering results can be represented by its cluster center, which is a representative point of the training courseware feature vector in the clustering result, and the training courseware data in each clustering result can have similar knowledge themes and content focuses. The present application can perform a detailed analysis of the courseware data corresponding to each training clustering result to extract the main knowledge and skills covered therein, thereby summarizing the knowledge themes represented by each training clustering result, and can form a list of knowledge themes for the training clustering result.

[0064] For job classification data, the job classification data corresponding to each job category can describe in detail the knowledge, skills and abilities required for the job. This application can obtain the knowledge areas and skills required for each job category by analyzing each job category, and organize and classify these knowledge requirements to form a knowledge requirement list for each job.

[0065] Furthermore, the knowledge topic list corresponding to the training clustering results can be compared and matched with the knowledge requirement list of each position, so as to check the K training clustering results one by one, and then associate them with the appropriate position category, and then obtain the position resource map. For example, if the theme of a training clustering result is "data analysis and visualization", and the data analyst position of the enterprise requires knowledge in data analysis, data mining and data presentation, then this training clustering result can be matched and associated with the data analyst position, and this application does not limit this.

[0066] S60: Generate an employee knowledge mastery map based on the job resource map and employee knowledge mastery related data.

[0067] The employee knowledge mastery map can be a tool that presents the employees' mastery of different knowledge fields in the form of a map (e.g., graphical form). By integrating and analyzing the mastery data of employees in various knowledge fields, the employees' knowledge strengths and weaknesses can be displayed in an intuitive visual form, providing a strong decision-making basis for the company's talent management, training planning, and knowledge inheritance.

[0068] It should be understood that generating an employee knowledge mastery map based on the job resource map and employee knowledge mastery related data refers to the process of updating and iterating the cluster center. Among them, step S60, that is, generating an employee knowledge mastery map based on the job resource map and employee knowledge mastery related data, includes the following steps:

[0069] S61: extracting information from the job resource map data to obtain a job resource entity information set;

[0070] S62: extracting information from employee knowledge-related data to obtain an employee knowledge entity information set;

[0071] S63: constructing triples according to the post resource entity information set and the employee knowledge entity information set to obtain a triple knowledge representation set;

[0072] S64: performing heterogeneous data disambiguation processing on each triple knowledge representation in the triple knowledge representation set by using an entity linking model to obtain a disambiguated knowledge representation set;

[0073] S65: performing knowledge fusion on each disambiguation knowledge representation in the disambiguation knowledge representation set to obtain fused knowledge mastering data;

[0074] S66: Construct a knowledge mastery model based on the integrated knowledge mastery data to obtain an employee knowledge mastery map.

[0075] The job resource map data after clustering is usually semi-structured and may exist in the form of tables, text with certain markup formats, etc. For example, the job resource map may list various job categories and their corresponding training courseware resources in a tree structure. The job resource map data may have a certain organizational form, but may not fully conform to the strict database structure.

[0076] Key information related to the position (which can also be understood as relevant entity information) can be extracted from the position resource map data, such as the position name, the subject of the corresponding training courseware, the type of courseware (such as theoretical courses, practical operation courses), etc., to obtain a set of position resource entity information. For example, similar information such as "R&D position-programming language training-theory + practice" can be extracted from the position resource map data of the R&D position, and this application does not limit this.

[0077] Employee knowledge mastery data is usually organized and stored in a structured manner. For example, the employee identity document (ID) is used as an index to store the assessment scores or grades of each employee in different knowledge fields (such as programming languages, project management, etc.). Information such as employee ID, knowledge field name, and mastery level (such as score or level) can be extracted from the structured employee knowledge mastery data to obtain an employee knowledge entity information set. For example, the mastery level of employee A in the field of programming languages ​​is extracted as 80 points.

[0078] The extracted information can be further converted into a triple of entities, relations, and attributes. For example, "employee A - master - programming language" is a triple, where "employee A" is an entity, "master" is a relation, and "programming language" is an attribute. Furthermore, the content of the triple can be further enriched. For example, in the triple of employee knowledge mastery, a specific value of the mastery degree can be added as a supplement to the attribute, such as "employee A - master (80 points) - programming language".

[0079] Optionally, the above information extraction process may include attribute extraction, relationship extraction, and entity extraction. Among them, entity extraction can be trained on the original corpus through machine learning methods, and the trained entity extraction model can be used to identify the courseware names and knowledge point entities in the job resource map data, and then the trained entity extraction model can be used to identify the employee names, mastered knowledge points, and unmastered knowledge points in the employee knowledge mastery related data.

[0080] Relation extraction can be done through supervised learning methods of machine learning relationship extraction, by training and learning in the labeled training set, and verifying with the same type of test set, and then using the trained relationship extraction model to identify the inclusion relationship, basic relationship, relationship between entity and knowledge point attribute value, instance relationship, parallel relationship, etc. of knowledge points. Attribute extraction can be done by customizing the name attribute (name), importance attribute (imp) and association attribute (on) of courseware chapter entity and knowledge point entity, and this application does not limit this.

[0081] Since the job resource map data and employee knowledge mastery-related data come from different systems and data sources, there may be heterogeneity in data format, semantics, etc. For example, the names of knowledge fields in the job resource map may use professional terms, while the names of knowledge fields in the employee knowledge mastery data may be more inclined to the popular names within the enterprise. This application can analyze and process these heterogeneous data through heterogeneous data disambiguation and knowledge fusion technology to eliminate ambiguity in the data and obtain fused knowledge mastery data after knowledge fusion.

[0082] It should be understood that performing knowledge fusion on each disambiguation knowledge representation in the disambiguation knowledge representation set to obtain fused knowledge mastering data refers to the process of performing knowledge fusion. In step S65, that is, performing knowledge fusion on each disambiguation knowledge representation in the disambiguation knowledge representation set to obtain fused knowledge mastering data, includes the following steps:

[0083] S651: Obtaining an entity representation subset and a context representation subset from a disambiguation knowledge representation set;

[0084] S652: Acquire candidate entities associated with each entity representation in the entity representation subset to obtain the candidate entity representation subset;

[0085] S653: calculating the correlation between each candidate entity representation in the candidate entity representation subset and each context courseware representation in the context courseware representation subset through the local entity linking model to obtain a local correlation score;

[0086] S654: Calculate the topic relevance between every two candidate entity representations in the candidate entity representation subset through the global entity linking model to obtain a global relevance score;

[0087] S655: Through the entity linking model, linking decisions are made according to the local relevance scores and the global relevance scores to obtain fused knowledge mastering data.

[0088] Among them, the training courseware data (i.e., courseware) may include multiple entity representations. For example, in a computer programming training courseware, the entity representation may be various concepts in the programming language, such as variables, functions, loop structures, etc. For each entity representation, there may be a corresponding candidate entity representation set, and the candidate entity representation may be other entity representations associated with the corresponding entity representation. For example, for the entity representation of "variable", its candidate entity representation set may include other related data type entity representations, scope entity representations, etc., and this application does not limit this.

[0089] The concept of upper and lower courses can be a relative one, usually referring to the previous course and the next course in a series of training courses. The upper and lower courses often have a certain degree of coherence and logical relationship with the current course in terms of content. This coherence can help us better understand the position and role of an entity (such as a knowledge point, concept, etc.) in the entire training content. It is understandable that when an entity has multiple possible explanations or application scenarios in the current course, the content of the upper and lower courses can provide additional information to eliminate ambiguity.

[0090] The local entity linking model can measure the correlation between candidate entities and contexts, disambiguate each entity individually, and generate a correlation score between each candidate entity and context using context information. Optionally, the local entity linking model is used to calculate the correlation between each candidate entity representation in the candidate entity representation subset and each context representation in the context representation subset to obtain a local correlation score, which can be obtained using the following formula:

[0091]

[0092] Among them, Ψ(e i , c i) represents the candidate entity representation in the local relevance score e i And the upper and lower courseware indicates c i The correlation score between i represents the i-th candidate entity representation in the candidate entity representation subset; c i represents the i-th context representation in the context representation subset; Represents the candidate entity representation e i Embedding Represents the candidate entity representation e i is the transpose of the embedding; B represents a learnable diagonal matrix; f(c i ) indicates that the upper and lower courses are c i The hard attention mechanism is applied to obtain the upper and lower courseware representation c i The feature representation of .

[0093] The global entity linking model encourages all entity representations in the courseware to have thematic consistency, and can usually be used to measure any pair of candidate entity representations, such as (e i , e j ), the topic relevance between them. Optionally, the process of calculating the topic relevance between every two candidate entity representations in the candidate entity representation subset through the global entity linking model to obtain the global relevance score can be adopted as follows:

[0094]

[0095] in, represents the candidate entity representation e in the global relevance score i and candidate entity representation e j The topic relevance score between i represents the i-th candidate entity representation in the candidate entity representation subset; e j represents the jth candidate entity representation in the candidate entity representation subset; D represents the training courseware data; n represents the number of entity representations in the entity representation subset; x ei Represents the candidate entity representation e i Embedding Represents the candidate entity representation e i The embedded transpose of ; C represents a diagonal matrix; x ej Represents the candidate entity representation e j Embedding.

[0096] Furthermore, an entity linking model can be introduced, and the input data of the entity linking model can contain multiple entity representations M = (m1, m2, ..., m n ), training courseware data D, and the candidate entity representation set E corresponding to each entity representation i =(ei1 , e i2 ,…,e in ). The output data of the entity linking model may be the linking decision with the highest probability, that is, the entity arrangement and combination with the highest probability. In the embodiment of the present application, the output data may be the data after knowledge fusion, that is, the fusion knowledge mastering data. Optionally, the entity linking model may be represented by the following formula:

[0097]

[0098] Where E* represents the output of the entity linking model, which indicates the linking decision with the highest probability, that is, the entity arrangement and combination that is most likely to be correct after model evaluation; E represents the linking decision; E1×…×E n represents the Cartesian product of candidate entity representation subsets corresponding to different entity representations; E1 represents the candidate entity representation subset corresponding to entity representation m1; E n Represents entity representation m n The corresponding candidate entity representation subset; n represents the number of entity representations in the entity representation subset; Ψ(e i , c i ) represents the output of the local entity linking model, which represents the candidate entity representation e in the local relevance score i And the upper and lower courseware indicates c i The correlation score between i represents the i-th candidate entity representation in the candidate entity representation subset; c i Represents entity representation m n Instant loading and unloading of courseware; represents the output of the global entity linking model, which represents the candidate entity representation e in the global relevance score i and candidate entity representation e j The topic relevance score between j represents the jth candidate entity representation in the candidate entity representation subset; D represents the training courseware data.

[0099] It can be understood that the solution of the entity linking model is divided into three parts. The first part is the search algorithm, which searches all possible linking decisions E1×E2×…×E n ; The second and third parts are search criteria, corresponding to the local entity linking model Ψ(e i , c i ) and the global entity linking model There are two search criteria, which are not limited in this application.

[0100] The entity linking model can accurately link and express fuzzy knowledge elements. For example, when multiple similar knowledge fields (such as different programming languages) are involved, the entity linking model can accurately map employees' mastery of specific programming languages ​​to avoid confusion. The model is used to optimize knowledge representation, ensuring that the entities, relationships, and attributes in each triple accurately reflect the actual situation, providing an accurate data foundation for subsequent knowledge fusion and graph construction.

[0101] Merge knowledge entities, relationships, and attributes extracted and represented from different data sources. For example, merge the knowledge attributes of "project management training courseware" in the job resource map with the employees' mastery of "project management" in the employee knowledge mastery to ensure the consistency and integrity of knowledge information. Resolve possible knowledge conflicts, such as when different data sources have different evaluations of the same employee's mastery of the same knowledge field, and use reasonable algorithms and rules to merge and adjust to obtain accurate knowledge information.

[0102] After acquiring the fused knowledge mastering data, the process of constructing the knowledge model may include: using the structured query language (MySQL) database to store instance data, using the direct mapping of relational databases to resource description framework (D2RQ) tool to realize the conversion to "entity-relationship-entity" or "entity-attribute-value" triple data, and then saving it in SQL, and then converting the SQL data to obtain RDF data, and associating it with the Web Ontology Language (OWL) file of the ontology to obtain the final knowledge graph, and then using a graph database management system, such as Nco4j graph database, to realize visualization, so as to store the structured data on the network.

[0103] It can be seen that in the above scheme, by obtaining training courseware data and job classification data, as well as employee knowledge mastery related data, the training courseware data can be preprocessed to obtain training feature data, and the training feature data can be vector constructed to obtain a training feature vector sample set, so that the training feature vector sample set can be clustered to obtain K training clustering results, and the K training clustering results can be further matched with the job classification data to generate a job resource map, and then the employee knowledge mastery map can be generated based on the job resource map and employee knowledge mastery related data, which can quickly and effectively realize the display of employee knowledge mastery, so as to understand the shortcomings of each employee in a targeted manner, which is conducive to each employee to carry out targeted learning.

[0104] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0105] In one embodiment, a device for constructing an employee knowledge mastery map is provided, and the device for constructing an employee knowledge mastery map corresponds one-to-one to the method for constructing an employee knowledge mastery map in the above embodiment. Figure 2 As shown, the employee knowledge mastery map construction device includes an acquisition module 101, a first processing module 102, a second processing module 103, a third processing module 104, a fourth processing module 105, and a fifth processing module 106. The detailed description of each functional module is as follows:

[0106] The acquisition module 101 is used for acquiring the training courseware data and job classification data, as well as employee knowledge mastery related data;

[0107] The first processing module 102 is used to pre-process the training courseware data to obtain training feature data;

[0108] The second processing module 103 is used to perform vector construction processing on the training feature data to obtain a training feature vector sample set;

[0109] The third processing module 104 is used to perform clustering processing on the training feature vector sample set to obtain K training clustering results;

[0110] The fourth processing module 105 is used to match the K training clustering results with the job classification data to generate a job resource map;

[0111] The fifth processing module 106 is used to generate an employee knowledge mastery map according to the job resource map and employee knowledge mastery related data.

[0112] In one embodiment, the third processing module 104 is used to perform clustering processing on the training feature vector sample set to obtain K training clustering results, specifically for:

[0113] Select K samples from the n data objects in the training feature vector sample set as K initial cluster centers; where n is the number of training courseware data, and K is the number of job categories in the job classification data;

[0114] Calculate the Euclidean distance between each training feature vector sample in the training feature vector sample set and the K initial cluster centers to obtain a Euclidean distance set;

[0115] According to the Euclidean distance set, each training feature vector sample in the training feature vector sample set is divided into the initial cluster center with the closest Euclidean distance, so as to obtain the training feature cluster cluster set;

[0116] According to the training feature clustering cluster set, the initial cluster center of each cluster is updated and optimized to obtain K training clustering results.

[0117] In one embodiment, the third processing module 104 is used to update and optimize the initial cluster center of each cluster according to the training feature clustering cluster set to obtain K training clustering results, specifically for:

[0118] According to the training feature clustering cluster set, the mutual distance between the initial cluster centers of each cluster is calculated to obtain the mutual distance set;

[0119] According to the training feature clustering cluster set, the surrounding element point density of the initial cluster center of each cluster is calculated to obtain the surrounding element point density set;

[0120] According to the mutual distance set, the surrounding element point density set and the preset conditions, the initial cluster center of each cluster is updated and optimized to obtain K training clustering results.

[0121] In one embodiment, the fifth processing module 106 is used to generate an employee knowledge mastery map according to the job resource map and employee knowledge mastery related data, specifically for:

[0122] Extract information from the job resource map data to obtain a set of job resource entity information;

[0123] Extract information from employee knowledge-related data to obtain employee knowledge entity information sets;

[0124] Construct triples based on the post resource entity information set and the employee knowledge entity information set to obtain a triple knowledge representation set;

[0125] Perform heterogeneous data disambiguation processing on each triple knowledge representation in the triple knowledge representation set through the entity linking model to obtain a disambiguated knowledge representation set;

[0126] Performing knowledge fusion on each disambiguation knowledge representation in the disambiguation knowledge representation set to obtain fused knowledge mastering data;

[0127] The knowledge mastery model is constructed based on the integrated knowledge mastery data to obtain the employee knowledge mastery map.

[0128] In one embodiment, the fifth processing module 106 is used to perform knowledge fusion on each disambiguation knowledge representation in the disambiguation knowledge representation set to obtain fused knowledge mastering data, specifically for:

[0129] Obtain entity representation subsets and context representation subsets from the disambiguation knowledge representation set;

[0130] Obtaining candidate entities associated with each entity representation in the entity representation subset to obtain a candidate entity representation subset;

[0131] By using the local entity linking model, the correlation between each candidate entity representation in the candidate entity representation subset and each context representation in the context representation subset is calculated to obtain a local correlation score;

[0132] Through the global entity linking model, the topic relevance between every two candidate entity representations in the candidate entity representation subset is calculated to obtain a global relevance score;

[0133] Through the entity linking model, linking decisions are made according to local relevance scores and global relevance scores to obtain fused knowledge mastering data.

[0134] The present invention provides a device for constructing an employee knowledge mastery map. By acquiring training courseware data and job classification data, as well as employee knowledge mastery-related data, the training courseware data can be preprocessed to obtain training feature data, and the training feature data can be vector constructed to obtain a training feature vector sample set, so that the training feature vector sample set can be clustered to obtain K training clustering results, and the K training clustering results can be further matched with the job classification data to generate a job resource map, and then an employee knowledge mastery map can be generated based on the job resource map and employee knowledge mastery-related data, which can quickly and effectively display the employee knowledge mastery level, so as to understand the shortcomings of each employee in a targeted manner, which is conducive to each employee to carry out targeted learning.

[0135] For the specific definition of the employee knowledge mastery map construction device, please refer to the definition of the employee knowledge mastery map construction method above, which will not be repeated here. Each module in the above-mentioned employee knowledge mastery map construction device can be implemented in whole or in part by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0136] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 3As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client through a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the server side of a method for constructing an employee knowledge mastery map.

[0137] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it realizes the functions or steps on the client side of a method for constructing an employee knowledge mastery map.

[0138] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0139] Obtain training courseware data, job classification data, and employee knowledge mastery data;

[0140] Preprocess the training courseware data to obtain training feature data;

[0141] Perform vector construction processing on the training feature data to obtain a training feature vector sample set;

[0142] Cluster the training feature vector sample set to obtain K training clustering results;

[0143] Match the K training clustering results with the job classification data to generate a job resource map;

[0144] Generate an employee knowledge mastery map based on the job resource map and employee knowledge mastery related data.

[0145] The present invention provides a computer device, which can pre-process the training courseware data to obtain training feature data by acquiring training courseware data and job classification data, as well as employee knowledge mastery-related data, and then perform vector construction processing on the training feature data to obtain a training feature vector sample set, so that the training feature vector sample set can be clustered to obtain K training clustering results, and the K training clustering results can be further matched with the job classification data to generate a job resource map, and then an employee knowledge mastery map can be generated according to the job resource map and employee knowledge mastery-related data, so as to quickly and effectively display the employee knowledge mastery level, so as to understand the shortcomings of each employee in a targeted manner, which is conducive to each employee to carry out targeted learning.

[0146] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0147] Obtain training courseware data, job classification data, and employee knowledge mastery data;

[0148] Preprocess the training courseware data to obtain training feature data;

[0149] Perform vector construction processing on the training feature data to obtain a training feature vector sample set;

[0150] Cluster the training feature vector sample set to obtain K training clustering results;

[0151] Match the K training clustering results with the job classification data to generate a job resource map;

[0152] Generate an employee knowledge mastery map based on the job resource map and employee knowledge mastery related data.

[0153] The present invention provides a computer-readable storage medium. By acquiring training courseware data and job classification data, as well as employee knowledge mastery-related data, the training courseware data can be preprocessed to obtain training feature data, and the training feature data can be vector constructed to obtain a training feature vector sample set, so that the training feature vector sample set can be clustered to obtain K training clustering results, and the K training clustering results can be further matched with the job classification data to generate a job resource map, and then an employee knowledge mastery map can be generated based on the job resource map and employee knowledge mastery-related data, which can quickly and effectively realize the display of employee knowledge mastery, so as to understand the shortcomings of each employee in a targeted manner, which is conducive to each employee to carry out targeted learning.

[0154] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0155] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0156] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0157] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for constructing an employee knowledge mastery map, characterized in that: The method comprises: Obtain training courseware data, job classification data, and employee knowledge mastery data; Preprocessing the training courseware data to obtain training feature data; Performing vector construction processing on the training feature data to obtain a training feature vector sample set; Performing clustering processing on the training feature vector sample set to obtain K training clustering results; Matching the K training clustering results with the job classification data to generate a job resource map; An employee knowledge mastery map is generated according to the job resource map and the employee knowledge mastery related data.

2. The method for constructing employee knowledge mastery map according to claim 1, characterized in that: The clustering process of the training feature vector sample set to obtain K training clustering results includes: Selecting K samples from the n data objects of the training feature vector sample set as K initial cluster centers; wherein n is the number of training courseware data, and K is the number of job categories in the job classification data; Calculating the Euclidean distance between each training feature vector sample in the training feature vector sample set and the K initial clustering centers to obtain a Euclidean distance set; According to the Euclidean distance set, each training feature vector sample in the training feature vector sample set is divided into the initial cluster center with the closest Euclidean distance to obtain a training feature cluster cluster set; According to the training feature clustering cluster set, the initial cluster center of each cluster is updated and optimized to obtain K training clustering results.

3. The method for constructing employee knowledge mastery map according to claim 2, characterized in that: The clustering set based on the training feature clustering is updated and optimized for the initial cluster center of each cluster to obtain K training clustering results, including: According to the training feature clustering cluster set, the mutual distance between the initial cluster centers of each cluster is calculated to obtain a mutual distance set; According to the training feature clustering cluster set, the surrounding element point density of the initial cluster center of each cluster is calculated to obtain a surrounding element point density set; According to the mutual distance set, the surrounding element point density set and preset conditions, the initial cluster center of each cluster is updated and optimized to obtain K training clustering results.

4. The method for constructing an employee knowledge mastery map according to claim 3, characterized in that: The step of generating an employee knowledge mastery map according to the position resource map and the employee knowledge mastery related data includes: Extract information from the job resource map data to obtain a set of job resource entity information; Extract information from employee knowledge-related data to obtain employee knowledge entity information sets; Constructing triples according to the post resource entity information set and the employee knowledge entity information set to obtain a triple knowledge representation set; Performing heterogeneous data disambiguation processing on each triple knowledge representation in the triple knowledge representation set by using an entity linking model to obtain a disambiguated knowledge representation set; Performing knowledge fusion on each disambiguation knowledge representation in the disambiguation knowledge representation set to obtain fused knowledge mastering data; A knowledge mastery model is constructed based on the fused knowledge mastery data to obtain an employee knowledge mastery map.

5. The method for constructing an employee knowledge mastery map according to any one of claims 1 to 4, characterized in that: The step of performing knowledge fusion on each disambiguation knowledge representation in the disambiguation knowledge representation set to obtain fused knowledge mastering data includes: Acquire entity representation subsets and context representation subsets from the disambiguation knowledge representation set; Acquire candidate entities associated with each entity representation in the entity representation subset to obtain a candidate entity representation subset; By using the local entity linking model, the correlation between each candidate entity representation in the candidate entity representation subset and each context representation in the context representation subset is calculated to obtain a local correlation score; Through the global entity linking model, the topic relevance between every two candidate entity representations in the candidate entity representation subset is calculated to obtain a global relevance score; Through the entity linking model, linking decisions are made according to the local relevance scores and the global relevance scores to obtain fused knowledge mastering data.

6. A device for constructing employee knowledge mastery graph, characterized in that: The employee knowledge mastery map construction device comprises: The acquisition module is used to obtain training courseware data, job classification data, and employee knowledge mastery data; A first processing module is used to pre-process the training courseware data to obtain training feature data; A second processing module is used to perform vector construction processing on the training feature data to obtain a training feature vector sample set; A third processing module is used to perform clustering processing on the training feature vector sample set to obtain K training clustering results; A fourth processing module is used to match the K training clustering results with the job classification data to generate a job resource map; The fifth processing module is used to generate an employee knowledge mastery map based on the job resource map and the employee knowledge mastery related data.

7. The employee knowledge mastery map construction device according to claim 6, characterized in that: The third processing module is used to perform clustering processing on the training feature vector sample set to obtain K training clustering results, specifically for: Selecting K samples from the n data objects of the training feature vector sample set as K initial cluster centers; wherein n is the number of training courseware data, and K is the number of job categories in the job classification data; Calculating the Euclidean distance between each training feature vector sample in the training feature vector sample set and the K initial clustering centers to obtain a Euclidean distance set; According to the Euclidean distance set, each training feature vector sample in the training feature vector sample set is divided into the initial cluster center with the closest Euclidean distance to obtain a training feature cluster cluster set; According to the training feature clustering cluster set, the initial cluster center of each cluster is updated and optimized to obtain K training clustering results.

8. The employee knowledge mastery map construction device according to claim 7, characterized in that: The third processing module is used to update and optimize the initial cluster center of each cluster according to the training feature clustering cluster set to obtain K training clustering results, specifically for: According to the training feature clustering cluster set, the mutual distance between the initial cluster centers of each cluster is calculated to obtain a mutual distance set; According to the training feature clustering cluster set, the surrounding element point density of the initial cluster center of each cluster is calculated to obtain a surrounding element point density set; According to the mutual distance set, the surrounding element point density set and preset conditions, the initial cluster center of each cluster is updated and optimized to obtain K training clustering results.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for constructing an employee knowledge mastery map as described in any one of claims 1 to 5 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for constructing an employee knowledge mastery map as described in any one of claims 1 to 5 is implemented.

Images