Talent management information retrieval system and method for chain enterprises

Abstract

The invention provides a talent management information retrieval system and method for a chain enterprise, and the method comprises the steps: receiving a retrieval request initiated by a retrieval terminal through a central talent database of the chain enterprise, carrying out the vectorization mapping of a flexible matching condition of the retrieval request, and obtaining a target skill set of the retrieval request; performing primary screening in the central talent database according to a rigid constraint condition of the retrieval request to generate a first candidate set of the retrieval request; performing weighted correction on the core matching score of the corresponding employee by using each regional flexibility additional score and each skill development potential additional score to generate a competency score of each employee in the first candidate set; and sorting the employees in the first candidate set according to all the comprehensive competency scores, and returning a sorted employee list as a retrieval result to a retrieval terminal of the chain enterprise. Based on the scheme, hierarchical competency assessment of chain enterprise talent information can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of talent information management, and more particularly to a talent management information retrieval system and method for a chain enterprise. BACKGROUND

[0002] A chain enterprise is a commercial organization with a headquarters for overall management and multiple standardized operation stores as subsidiaries. The business of the chain enterprise covers multiple regions and multiple formats, and the core features of the chain enterprise are unified brand, standardized process and centralized resources. The chain enterprise has a large-scale talent team, and due to the dispersion of the stores, the chain enterprise has special requirements for the regional adaptability, skill versatility and post coordination ability of talents. An efficient talent management system is needed to realize cross-store talent allocation and accurate post matching to ensure the unity of overall operation and the standardization of service.

[0003] In the traditional talent management information retrieval technology system of a chain enterprise, there is a technical shortcoming of missing flexible ability matching dimension in the talent screening link. Existing solutions often only use rigid indexes such as education, working years and title level as the core screening basis, rely on simple comparison of structured data to complete talent preliminary screening and matching, and do not establish an effective quantification and matching mechanism for flexible abilities such as professional skill proficiency, business literacy adaptability and post coordination awareness required by the actual post. Since flexible abilities are mostly unstructured and abstract ability elements, the existing technology cannot convert them into computable matching dimensions, resulting in that the screening process can only cover the basic job threshold, but it is difficult to meet the deep needs of the post. Ultimately, a large number of talents in the retrieval results meet the rigid indexes, but have low adaptability in actual business ability, which not only increases the labor cost of subsequent talent selection of the enterprise, but also easily causes the mismatch between the person and the post, and cannot support the precise talent allocation needs of the chain enterprise across stores and multiple posts. Therefore, how to realize the hierarchical competency evaluation of talent information of a chain enterprise and improve the adaptability efficiency of talent retrieval has become a difficult problem in the industry. SUMMARY

[0004] The present application provides a talent management information retrieval system and method for a chain enterprise, which can realize hierarchical competency evaluation of talent information of a chain enterprise and improve the adaptability efficiency of talent retrieval.

[0005] In a first aspect, the present application provides a talent management information retrieval method for a chain enterprise, comprising: A central talent database of a chain enterprise receives a retrieval request initiated by a retrieval terminal, and then extracts rigid constraint conditions and flexible matching conditions in the retrieval request; The flexible matching conditions are vectorized and mapped to obtain a target skill set of the retrieval request, and then a preliminary screening is performed in the central talent database according to the rigid constraint conditions to generate a first candidate set of the retrieval request; vector similarity between the actual skill vector of each employee in the first candidate set and the skill requirement vector in the target skill set, to obtain a core matching score of each employee, while determining a regional flexibility additional score and a skill development potential additional score of each employee in the first candidate set, and weighting and correcting the core matching score of the corresponding employee using the regional flexibility additional score and the skill development potential additional score, to generate a competency score of each employee in the first candidate set; According to all the comprehensive competency scores, the employees in the first candidate set are sorted, and the sorted employee list is returned to the search terminal of the chain enterprise as a search result.

[0006] In some embodiments, the flexible matching condition is vectorized and mapped to obtain the target skill set of the search request, which specifically includes: Based on the post competency standard library of the chain enterprise, the flexible matching condition is subjected to natural language keyword extraction to obtain the flexible keywords of the search request; The flexible keywords are subjected to semantic similarity matching with the standardized skill labels in the standard library, and then the non-standard descriptions in the flexible matching condition are mapped to the standardized skill labels to obtain a plurality of skill features; According to the preset weight of the standardized skill label in the standard library and the priority modifier word in the search request, an initial weight coefficient is assigned to each skill feature to obtain the target skill set of the search request.

[0007] In some embodiments, the primary screening is performed in the central talent database according to the rigid constraint condition to generate the first candidate set of the search request, which specifically includes: The rigid constraint condition is parsed to identify the constraint type and constraint value in the rigid constraint condition, the constraint type including: current store range, post sequence, job level range, and in-service state; Based on the index of the central talent database, the employee records satisfying all constraint types and constraint values are subjected to fast intersection query to obtain the first candidate set of the search request.

[0008] In some embodiments, the vector similarity between the actual skill vector of each employee in the first candidate set and the skill requirement vector in the target skill set is calculated to obtain a core matching score of each employee, which specifically includes: According to the skill label and weight coefficient in the target skill set, a skill requirement vector in the target skill set is constructed; The actual skill vector of each employee is constructed through the skill item score, skill authentication record and level recorded in the historical performance data of the employee, and the vector dimension is aligned with the skill requirement vector; a core matching score of each employee in the first candidate set is obtained by calculating the similarity between each actual skill vector and the skill requirement vector.

[0009] In some embodiments, determining the geographical flexibility additional score and the skill development potential additional score of each employee in the first candidate set specifically comprises: For each employee in the first candidate set, the number of cross-store transfers of the employee within a specified historical period is counted, and then the transfer activity of the employee is calculated, and the deployment willingness strength filled by the employee is obtained; The transfer activity and the deployment willingness strength are fused into the geographical flexibility additional score of the employee; The skill growth characteristics of the employee are extracted from the performance data of the employee related to the target skill set within the historical performance period, and the training performance score of the employee is extracted from the relevant training records participated by the employee; The skill growth characteristics and the training performance score are fused and mapped into the skill development potential additional score of the employee, and then the geographical flexibility additional score and the skill development potential additional score of each employee in the first candidate set are obtained.

[0010] In some embodiments, the core matching score of the corresponding employee is weighted and corrected using the geographical flexibility additional score and the skill development potential additional score, and the competency score of each employee in the first candidate set is generated specifically comprising: The scenario-based weight coefficients are set for the geographical flexibility additional score and the skill development potential additional score, and then the score correction values of the geographical flexibility and the skill development potential are determined respectively; The core matching score of the employee is corrected using the respective score correction values, and the competency score of each employee in the first candidate set is obtained.

[0011] In some embodiments, the employees in the first candidate set are sorted according to all the comprehensive competency scores, and the sorted employee list is returned to the search terminal of the chain enterprise as a search result specifically comprising: All employees in the first candidate set are arranged in descending order according to the comprehensive competency scores, and an initial employee list of the search request is obtained; When multiple employees come from the same store, the initial employee list is adjusted according to the preset rules to avoid the risk of excessive loss of talents from a single store, and an employee list of the search request is obtained; A recommended reason summary is generated for each employee in the employee list, and the summary content integrates the core matching advantages, geographical flexibility information and development potential highlights of the corresponding employee; The employee list with the recommended reason summary is packaged as structured data and returned to the search terminal initiating the request for visual display.

[0012] In a second aspect, the present application provides a talent management information retrieval system for a chain enterprise, comprising a retrieval unit, the retrieval unit comprising: An initialization module, configured to receive a retrieval request initiated by a retrieval terminal by a central talent database of the chain enterprise, and then extract rigid constraint conditions and flexible matching conditions in the retrieval request; A processing module, configured to vectorize map the flexible matching conditions to obtain a target skill set of the retrieval request, and then perform a primary screening in the central talent database according to the rigid constraint conditions to generate a first candidate set of the retrieval request; The processing module is further configured to calculate vector similarity between an actual skill vector of each employee in the first candidate set to the target skill set and a skill requirement vector in the target skill set to obtain a core matching score of each employee, determine a regional flexibility additional score and a skill development potential additional score of each employee in the first candidate set, and use the regional flexibility additional score and the skill development potential additional score to weight and correct the core matching score of the corresponding employee to generate an competency score of each employee in the first candidate set; An execution module, configured to sort the employees in the first candidate set according to all the comprehensive competency scores, and return an employee list sorted to the retrieval terminal of the chain enterprise as a retrieval result.

[0013] In a third aspect, the present application provides a computer device, comprising a memory and a processor, the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the talent management information retrieval method for a chain enterprise as described above.

[0014] In a fourth aspect, the present application provides a computer readable storage medium, the computer readable storage medium stores instructions or codes, when the instructions or codes are run on a computer, the computer executes the talent management information retrieval method for a chain enterprise as described above.

[0015] The technical scheme provided by the embodiments of the present application has the following beneficial effects: The talent management information retrieval system and method for a chain enterprise provided in the application, wherein a central talent database of a chain enterprise receives a retrieval request initiated by a retrieval terminal, and then extracts rigid constraint conditions and flexible matching conditions in the retrieval request; the flexible matching conditions are vectorized and mapped to obtain a target skill set of the retrieval request, and then a primary screening is performed in the central talent database according to the rigid constraint conditions to generate a first candidate set of the retrieval request; the vector similarity between the actual skill vector of each employee in the first candidate set to the target skill set and the skill requirement vector in the target skill set is calculated to obtain a core matching score of each employee, while the regional flexibility additional score and the skill development potential additional score of each employee in the first candidate set are determined, the core matching score of the corresponding employee is weighted and corrected using each regional flexibility additional score and each skill development potential additional score to generate the competency score of each employee in the first candidate set; the employees in the first candidate set are sorted according to all the comprehensive competency scores, and the sorted employee list is returned to the retrieval terminal of the chain enterprise as a retrieval result.

[0016] It can be seen that, in the present application, the employees in the first candidate set are ranked according to all the comprehensive competence scores, and the ranked employee list is returned to the search terminal of the chain enterprise as a search result; first, the first candidate set is determined, that is, a talent preliminary screening pool meeting the basic employment threshold is obtained, so that personnel who do not meet the core job requirements are quickly eliminated through the pre-filtering of rigid constraints, and the sample range of subsequent competence evaluation is greatly compressed, laying a solid foundation for hierarchical quantitative evaluation. From a technical point of view, the combination of precise extraction of rigid constraints and database screening logic can achieve the preliminary stratification of talent information, focusing the search range from full talent data to a subset that meets the hard standards, avoiding the waste of computing power and efficiency loss caused by indiscriminate quantitative evaluation; at the same time, this preliminary screening process ensures that all talents in the first candidate set have the basic qualifications required for the post, providing a qualified sample pool for the subsequent fine matching of flexible skills, making the starting point of hierarchical competence evaluation more targeted, and indirectly improving the response speed of the overall search process, laying a technical prerequisite for sample screening to improve adaptation efficiency; then, the competence score is determined, that is, a talent comprehensive adaptation quantitative value integrating multiple dimensions is obtained, so as to realize the hierarchical quantitative evaluation of talents, and directly promote the improvement of talent search adaptation efficiency; the core matching score realizes the precise benchmarking of skill dimension through vector similarity, and the weighted correction of regional flexibility and skill development potential bonus, which builds a multi-dimensional competence evaluation model, upgrading single skill matching to comprehensive ability stratification. This quantitative process can convert different dimensions of talent ability into a unified numerical standard, not only realizing the precise stratification of talent adaptation degree, but also providing an objective basis for the ranking of search results, avoiding the bias of subjective evaluation; at the same time, the standardized competence score can directly drive the intelligent ranking of search results, enabling the chain enterprise to quickly locate high-adaptation talents; in summary, based on the above scheme, hierarchical quantitative evaluation of talent information for chain enterprises can be realized, thereby improving the adaptation efficiency of talent search. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creating labor.

[0018] Figure 1 is an exemplary flowchart of a talent management information search method for a chain enterprise according to some embodiments of the present application; Figure 2 is a flowchart of determining a core matching score according to some embodiments of the present application; Figure 3is a structural schematic diagram of a retrieval unit according to some embodiments of the present application; Figure 4 is a structural schematic diagram of a computer device for implementing a talent management information retrieval method for a chain enterprise according to some embodiments of the present application. DETAILED DESCRIPTION

[0019] In order to better understand the technical solutions of the present application, the technical solutions of the present application will be described in detail below in combination with the drawings of the specification and specific embodiments.

[0020] Reference Figure 1 The figure is an exemplary flowchart of a talent management information retrieval method for a chain enterprise according to some embodiments of the present application, which mainly includes the following steps: In step 101, the central talent database of the chain enterprise receives a retrieval request initiated by a retrieval terminal, and then extracts rigid constraint conditions and flexible matching conditions in the retrieval request.

[0021] It should be noted that in the present application, the central talent database is a database for centrally storing standardized full-cycle career data of all chain employees; the retrieval terminal is a user interaction device or software interface for initiating talent retrieval instructions; the retrieval request is a composite input for expressing user talent retrieval requirements; and in specific implementation, the central talent database of the chain enterprise receives the retrieval request submitted in the form of a structured data packet from the retrieval terminal through a pre-set application program interface.

[0022] In some embodiments, the extraction of rigid constraint conditions and flexible matching conditions in the retrieval request can be implemented in the following manner, i.e., starting a natural language understanding process to analyze the composite retrieval request, which uses a trained named entity recognition model and a rule-based condition classifier; the named entity recognition model in the condition classifier is responsible for identifying all entities in the request text, such as job title, skill name, region name, number (e.g., years of experience), etc., and normalizing all entities; then, the condition classifier classifies all identified entities and condition phrases according to pre-set classification rules, which clearly stipulate that any entity and phrase involving exact matching and being a necessary condition (e.g., specified job code, specified job level range, mandatory in-service status, explicit minimum work experience requirement, specified current store range) is classified as a rigid constraint condition; any descriptive, classifiable or differentially existing preference phrase (e.g., "good communication skills", "baking experience", "leadership potential", "accept cross-zone relocation") involving skills, qualities, experience, etc., is classified as a flexible matching condition.

[0023] It should be noted that in the present application, the rigid constraint condition is an enumerated condition used for screening talent records, and the flexible matching condition is a preference condition used for similarity evaluation of talent records.

[0024] In step 102, the flexible matching condition is vectorized and mapped to obtain a target skill set of the search request, and then a primary screening is performed in the central talent database according to the rigid constraint condition to generate a first candidate set of the search request.

[0025] In some embodiments, the vectorization and mapping of the flexible matching condition to obtain the target skill set of the search request can be implemented by the following steps: Based on the post competency standard library of the chain enterprise, natural language keyword extraction is performed on the flexible matching condition to obtain flexible keywords of the search request; The flexible keywords are matched with standardized skill labels in the standard library in terms of semantic similarity, and then non-standard descriptions in the flexible matching condition are mapped to standardized skill labels to obtain a plurality of skill features; According to the preset weight of the standardized skill label in the standard library and the priority modifier in the search request, an initial weight coefficient is assigned to each skill feature to obtain a target skill set of the search request.

[0026] It should be noted that in the present application, the post competency standard library is a structured knowledge base used to define and store standardized skills, qualities and their weights required for each post of the chain enterprise; the flexible keywords are a set of vocabulary used for subsequent mapping and matching processing; the standardized skill label is a standardized name used to uniquely identify a specific skill in the post competency standard library; the non-standard description is a user input vocabulary used to express skill quality but not in the form of standardized skill label; the skill feature is a structured object representing a specific ability requirement, containing standardized skill label and preliminary weight information; the priority modifier is a descriptive word used to enhance or weaken the degree of skill requirement in the search request; and the target skill set is a requirement set used for talent matching degree calculation.

[0027] In a specific implementation, first, a Chinese word segmentation component is called to perform word segmentation processing on the text to cut continuous text into an independent word sequence; a business dictionary composed of high-frequency words in the business field of the chain enterprise is loaded from the post competency standard library of the chain enterprise, and the word sequence after word segmentation is filtered in combination with a general stop word table; the filtering rule is to retain words that belong to the business dictionary and are in the stop word table, and to eliminate virtual words such as “de” and “le” that have no actual meaning. The word sequence remaining after filtering is the extracted flexible keyword; for example, for the input “find employees good at communication and customer service conscious”, the flexible keywords “good at”, “communication”, “customer service”, and “conscious” can be obtained after word segmentation and filtering; the obtained flexible keywords are used as the flexible keywords of the retrieval request; then, all the flexible keywords and the preloaded post competency standard library are obtained, and the standard library contains full-quantity standardized skill labels; each flexible keyword is processed one by one: for each flexible keyword, the semantic similarity between the flexible keyword and each standardized skill label in the standard library is calculated; the calculation of the semantic similarity is based on a pre-trained word vector model, which maps words to vectors in a high-dimensional space, and the calculation process is to first convert the keyword and the skill label into corresponding word vectors, and then calculate the cosine value between the two vectors, which is used as the score of the semantic similarity; for each keyword, all the standardized skill labels with a semantic similarity score exceeding a preset threshold (default is 0.7) are found; if found, the label with the highest similarity is selected as the mapping result of the keyword; if not found, the keyword is regarded as a non-standard description that cannot be mapped, and can be manually intervened or temporarily ignored, and through the above-mentioned manner, the effective description in the flexible matching condition can be mapped to a specific standardized skill label; each successfully mapped standardized skill label is used as an independent skill feature; finally, for each skill feature corresponding to the standardized skill label, the pre-set basic weight value of the standardized skill label is queried from the post competency standard library, and the original flexible matching condition text is scanned to identify the priority modifier words existing therein, which are defined by a pre-defined rule library, for example, “master” and “skilled” usually correspond to weight addition, and “understand” and “have experience” usually correspond to a baseline or slightly lower weight; according to the type of the modifier word associated with the specified skill identified, a preset weight adjustment coefficient is called to multiply the basic weight of the skill label by the weight adjustment coefficient to obtain an initial weight coefficient for the skill feature; if a skill feature is not associated with any priority modifier word, the basic weight thereof is directly used as the initial weight coefficient; through the above-mentioned manner, each skill feature is associated with a quantified initial weight coefficient; all the skill features with initial weight coefficients are combined together to form a target skill set of the retrieval request.

[0028] In some embodiments, the primary screening in the central talent database according to the rigid constraint conditions can be implemented by the following steps: The rigid constraint conditions are parsed, and the constraint types and constraint values in the rigid constraint conditions are identified, the constraint types including: the current store range, the post sequence, the job level range, and the in-service state; Based on the index of the central talent database, the employee records satisfying all constraint types and constraint values are subjected to a quick intersection query, and a first candidate set of the search request is obtained.

[0029] It should be noted that in the present application, the first candidate set is a talent record set that preliminarily satisfies all rigid conditions and is used for fine matching calculation; the constraint type is a classification identifier used to define the specific dimension of the talent record limited by the rigid constraint condition; and the constraint value is a character segment used to define the specific filtering standard under the corresponding constraint type.

[0030] In specific implementation, first, the rigid constraint conditions exist in the form of a structured key-value pair set, for example, {“constraint type 1”: “constraint value 1”, “constraint type 2”: “constraint value 2”}, first, the received key-value pair is parsed according to the pre-defined rigid constraint condition metadata table, to verify its legality and clarify its semantics; the metadata table defines all allowed constraint types and their corresponding value domain formats, in the parsing process, each key in the input is mapped to an internally defined constraint type, and it is verified whether the corresponding constraint value conforms to the value domain format of the type, the identified constraint types at least include “current store range”, “post sequence”, “job level range”, and “in-service state”; for example, a constraint condition can be parsed as: the constraint type is “post sequence”, and the corresponding constraint value is “store manager sequence”; the constraint type is “in-service state”, and the corresponding constraint value is “in-service” or “on standby”; the entire constraint type and the corresponding constraint value that are parsed and verified are taken as the identification result; then, based on the identified rigid constraint type and constraint value, the corresponding database query sub-condition is generated, for each constraint type, the index of the central talent database is used for the corresponding field, for example, a B-tree index is established for the “current store code” field, and a hash index is established for the “in-service state” field; when querying, for each constraint sub-condition, the index is used to quickly locate the employee record primary key set satisfying the condition; the intersection operation is performed on the employee record primary key sets satisfying different conditions respectively, that is, the primary keys that appear in all primary key sets at the same time are found out, each primary key represents an employee who satisfies all rigid constraint conditions at the same time; the corresponding employee record details are completely retrieved from the central talent database according to these primary keys, for example, the skill information and the transfer history of the employee, to form a result set after preliminary screening; and the result set containing detailed employee records is taken as the first candidate set of the search request.

[0031] In step 103, vector similarity between the actual skill vector of each employee in the first candidate set and the skill requirement vector in the target skill set is calculated to obtain the core matching score of each employee, while the regional flexibility additional score and the skill development potential additional score of each employee in the first candidate set are determined, and the core matching score of the corresponding employee is weighted and corrected using the regional flexibility additional score and the skill development potential additional score to generate the competency score of each employee in the first candidate set.

[0032] In some embodiments, vector similarity between the actual skill vector of each employee in the first candidate set and the skill requirement vector in the target skill set is calculated to obtain the core matching score of each employee, and the core matching score is determined with reference to the graph. Figure 2 The graph is a flowchart for determining the core matching score in some embodiments of the present application. The core matching score in the present embodiment can be achieved by the following steps: In step 1031, the skill requirement vector in the target skill set is constructed according to the skill labels and weight coefficients in the target skill set. In step 1032, the actual skill vector of each employee is constructed by the skill item scores, skill authentication records and levels recorded in the historical performance data of the employee, and the vector dimension is aligned with the skill requirement vector. In step 1033, the similarity between each actual skill vector and the skill requirement vector is calculated to obtain the core matching score of each employee.

[0033] It should be noted that in the present application, the core matching score is a comprehensive score representing the matching degree of the skill vector of the employee and the skill requirement vector in the skill dimension; the skill requirement vector is a multi-dimensional numerical vector used to uniformly represent each skill and its weight in the target skill set in mathematics; and the actual skill vector is a multi-dimensional numerical vector used to uniformly represent each skill and its proficiency level mastered by the employee in mathematics.

[0034] In a specific implementation, first, the target skill set exists in the form of a list, each item in the list contains a standardized skill label and its corresponding weight coefficient, all skill labels in the target skill set are mapped to fixed vector positions according to a predefined global skill dimension table, which specifies the unique index number of each standardized skill in the vector; an initial value of zero is created, the length of the vector is equal to the total number of skills in the global skill dimension table, for each item in the target skill set, the value of the vector index position corresponding to its skill label is set to the weight coefficient of the item; if a certain skill is not included in the target skill set, the value of the skill at the corresponding position of the vector remains zero; the numerical vector filled with weight coefficients is used as the skill requirement vector; then, for each employee in the first candidate set, the employee's actual skill vector is constructed, the construction process begins with retrieving the employee's complete historical performance data and skill certification records from the central talent database, using the same global skill dimension table as the alignment reference as when constructing the skill requirement vector; a vector with the same dimension as the skill requirement vector is created, all elements have an initial value of zero, for each piece of historical performance data of the employee, extract the skill item name and corresponding skill item score recorded therein; according to the skill item name, find its corresponding vector index in the global skill dimension table, update the value at this index position to the skill item score; if there are multiple performance records for the same skill item, the highest score or the score of the latest period can be used for updating; process the employee's skill certification records: for each record, extract the certified skill name and level (e.g., primary, intermediate, advanced), convert the level to a standardized score according to the preset rules (e.g., 60 points for primary, 80 points for intermediate, and 100 points for advanced); Similarly, find the corresponding vector index according to the skill name, and update the value at this position to the certification conversion score; if there is a performance score at this position, take the higher of the performance score and the certification score as the final value; Through the above method, the proficiency of the employee in each skill dimension is quantified as a numerical value in the vector, the numerical vector generated for each employee is used as the employee's actual skill vector, and the actual skill vectors of each employee are obtained; finally, for each employee, obtain the employee's actual skill vector and the unified skill requirement vector, the specific calculation of similarity uses the cosine similarity method; first, calculate the dot product of the two vectors, that is, multiply the values of the two vectors in each corresponding dimension and sum them up, calculate the modulus length of the two vectors respectively, that is, the square sum of all dimension values of each vector is square rooted, divide the dot product by the product of the two modulus lengths, the result obtained is the cosine similarity value, its numerical range is between 0 and 1, the larger the value, the more consistent the direction of the two vectors, that is, the higher the matching degree;To get a more intuitive score, the cosine similarity value is usually multiplied by a full score base (default 100 points) for linear mapping, so as to convert it into a core matching score in the percentage or decimal system range as the core matching score of the employee. In this way, the core matching scores of various employees can be obtained.

[0035] In some embodiments, determining the regional flexibility additional score and the skill development potential additional score of each employee in the first candidate set can be achieved by the following steps: For each employee in the first candidate set, the number of cross-store transfers of the employee in a specified historical period is counted, and then the transfer activity of the employee is calculated, and the allocation willingness strength filled by the employee is obtained. The transfer activity and the allocation willingness strength are fused into the regional flexibility additional score of the employee. The skill growth characteristics of the employee are extracted from the performance data of the employee related to the target skill set in the historical performance period, and the training performance score of the employee is extracted from the relevant training records of the employee. The skill growth characteristics and the training performance score are fused and mapped into the skill development potential additional score of the employee, and then the regional flexibility additional score and the skill development potential additional score of each employee in the first candidate set are obtained.

[0036] It should be noted that in the present application, the regional flexibility additional score is an additional score value reflecting the adaptability and willingness of the employee to cross-regional allocation; the skill development potential additional score is an additional score value for predicting the future skill improvement speed and potential of the employee; the transfer activity is a quantitative index reflecting the active degree of the historical behavior of the employee in cross-store transfer; the allocation willingness strength is a quantitative numerical value representing the subjective willingness of the employee to accept future store allocation; the skill growth characteristics are numerical characteristics quantifying the ability progress trend and speed of the employee in the field related to the target skill; and the training performance score is a quantitative score measuring the participation, completion and learning effect of the employee in the training process.

[0037] In a specific implementation, first, for each employee in the first candidate set, all store transfer records in a specified historical period, such as "the past three years", are filtered from the employee's personnel change records, and the total number of valid transfer records is counted to obtain the number of cross-store transfers of the employee; the number is processed according to a preset transfer activity calculation model, which may perform nonlinear conversion on the number of transfers, for example, directly using the number as the activity value, or using a logarithmic function to avoid the disproportionate impact of too high a number; at the same time, the latest submitted deployment willingness survey result of the employee is read from the employee's file, which is usually a quantitative value (for example, on a scale of 1 to 5 points, 5 points represent "very willing"), and the value is the deployment willingness strength; second, according to a preset fusion rule, the transfer activity and the deployment willingness strength are combined, the fusion rule is to assign different weights to the two and then perform weighted summation, for example, the transfer activity weight is 0.7 and the deployment willingness strength weight is 0.3, and the weighted summation result is mapped into a standardized additional score interval, for example, 0 to 10 points, through a preset linear or nonlinear function, and the standardized score value obtained by the mapping is taken as the regional flexibility additional score of the employee; then, for each employee in the first candidate set, two feature extraction operations are performed in parallel, the first one is to extract skill growth features from the employee's historical performance data, according to the skill tags in the target skill set, the skill item score sequence related to these tags in the employee's historical performance data is filtered out; trend analysis is performed on the score sequence; for example, the linear regression slope of the skill score in the last three consecutive performance periods is calculated, and the slope value can reflect the speed of skill growth; or the growth rate of the score in the last period relative to the score in the first period is calculated; for multiple skills in the target skill set, the growth features of each skill can be calculated respectively and then averaged or weighted averaged to obtain a comprehensive skill growth feature value; the second one is to extract training performance scores from the employee's training records, and the related completed training records are filtered out, for each record, a single training performance score is calculated according to a preset rule by comprehensively considering the course completion rate, final examination score, classroom interaction participation (for example, if there is a record), and other indicators; the scores of all related training records are averaged to obtain the training performance score of the employee; finally, according to a preset potential evaluation model, the extracted skill growth feature value and the training performance score are fused and mapped. The model may assign different importance weights to the two and then perform weighted summation, and the summation result is mapped into a standardized potential score interval, for example, 0 to 10 points, through a predefined conversion function (for example, a piecewise linear function). The standardized score value obtained by the mapping is taken as the skill development potential additional score of the employee, and the regional flexibility additional score and the skill development potential additional score of each employee in the first candidate set can be obtained through the above method.

[0038] In some embodiments, the core matching score of the corresponding employee is weighted and corrected using the respective regional flexibility additional score and the respective skill development potential additional score, and the generation of the competency score of each employee in the first candidate set can be achieved by the following steps: The scenario-based weight coefficients are set for the regional flexibility additional score and the skill development potential additional score, and the scoring correction values of the regional flexibility and the skill development potential are determined respectively. The core matching score of the employee is corrected using the respective scoring correction value, and the competency score of each employee in the first candidate set is obtained.

[0039] It should be noted that in the present application, the competency score is a total score value used for final talent ranking; and the scenario-based weight coefficient is a configuration parameter used for dynamically adjusting the importance of different evaluation dimensions.

[0040] In a specific implementation, first, the current search request corresponding to the human resource business scenario is identified, and the basis for scenario identification includes: the scenario label (for example, "urgent job placement", "long-term reserve", "new store establishment") selected by the user in the search interface or the scenario category automatically inferred by analyzing the search request text; the system is preconfigured with a scenario weight configuration table, which specifies the scenario weight coefficients corresponding to the regional flexibility additional points and the skill development potential additional points for each defined business scenario; for example, for the "urgent job placement" scenario, the weight coefficient of regional flexibility (emphasizing rapid on-site) can be set to 0.8, and the weight coefficient of skill development potential (less important in the short term) can be set to 0.2; for the "long-term reserve" scenario, the weights of the two can be 0.3 and 0.7, respectively. The system retrieves the corresponding two sets of scenario weight coefficients from the configuration table according to the identified scenario; for each employee in the first candidate set, multiply the calculated regional flexibility additional points by the corresponding scenario weight coefficient to obtain the score correction value of the regional flexibility dimension; similarly, multiply the skill development potential additional points by the corresponding scenario weight coefficient to obtain the score correction value of the skill development potential dimension; then, through a pre-set comprehensive calculation model, the various score correction values are fused into a single competency score, which is a linear weighted aggregation model, specifically, for each employee, the core matching score is taken as the base score; the score correction value of the regional flexibility dimension and the score correction value of the skill development potential dimension are added to the base score in a weighted manner, and the total score after accumulation may need to be normalized, for example: through a linear scaling function, the scores of all employees are mapped to a unified, fixed score interval (for example: 0 to 100 points), to ensure that the scores generated under different search requests are comparable. The final value obtained after accumulation and possible normalization is the comprehensive competency score of the employee, and the comprehensive value generated for each employee in the first candidate set is taken as the competency score of the employee, and the competency score of each employee in the first candidate set can be obtained through the above method.

[0041] In step 104, the employees in the first candidate set are sorted according to all the comprehensive competency scores, and the sorted employee list is returned to the search terminal of the chain enterprise as the search result.

[0042] In some embodiments, the employees in the first candidate set are sorted according to all the comprehensive competency scores, and the sorted employee list is returned to the search terminal of the chain enterprise as the search result can be implemented by the following steps: The all employees in the first candidate set are arranged in descending order according to the comprehensive competency scores, and an initial employee list of the search request is obtained; When multiple employees from the same store appear, the initial employee list is risk-adjusted according to preset rules to avoid the risk of excessive loss of talents from a single store, and an employee list of a search request is obtained; A recommended reason summary is generated for each employee in the employee list, and the summary content integrates the core matching advantages, regional flexibility information and development potential highlights of the corresponding employee; The employee list with the recommended reason summary is encapsulated as structured data and returned to the search terminal initiating the request for visual display.

[0043] In a specific implementation, first, all employees are arranged in order of their comprehensive competence scores from high to low, with the employee with the highest score at the front of the sequence and the employee with the lowest score at the end of the sequence. If the comprehensive competence scores are the same, a secondary sorting key can be used to determine the order, such as ascending order according to the employee ID or name. After sorting, an employee identifier (e.g., employee ID) sequence is generated in this order, and the ordered employee identifier sequence is used as the initial employee list for the search request. Second, the initial employee list is received, and the store information to which each employee in the list currently belongs is loaded. The initial employee list is scanned from the beginning to the end, and a counter is maintained for each store to record how many employees from the store are included in the current recommended list. A risk threshold is defined in the preset rules, such as "no more than 2 employees from the same store" or "no more than 3 employees from the same store in the top 10 of the list". During the arrangement process, if adding the next employee in the sequence to the list causes the count of the store to which the employee belongs to exceed the preset risk threshold, a risk adjustment operation is triggered. The specific implementation process of risk adjustment is as follows: temporarily skip the employee, continue to check the subsequent positions in the list for employees from other stores and not exceeding the threshold of their own stores, and insert the one with the highest comprehensive competence score into the current position. The skipped employee is placed in a temporary queue, and in the subsequent list positions, when the count of the store to which the employee belongs is relatively reduced due to the insertion of employees from other stores, so that the addition of the employee no longer violates the threshold rule, the employee is inserted into the list. Through the above adjustment, the number of employees from any single store in the final list is controlled within the preset upper limit. The new ordered employee identifier sequence after the adjustment is used as the employee list for the search request. Then, a text form of recommendation reason summary is generated for each employee in the employee list. The generation process is templated: first, from the employee's skill matching calculation details, extract the top several skills with the highest matching degree (e.g., skill items with a matching degree of more than 90%), combine the names of these skills into a description, such as "mastering skill A and skill B", as the core matching advantage. Then, from the employee's regional flexibility additional score calculation basis, extract key information, such as "3 cross-city relocation records in the past two years and strong willingness to relocate", as regional flexibility information. Then, from the employee's skill development potential additional score calculation basis, extract positive signals, such as "relevant skill items have an average growth rate of 15% in the past three periods, and recent training assessment is excellent", as development potential highlights. Finally, a coherent text is composed according to a fixed sequence and conjunctions (e.g., "The employee is skilled in [core matching advantage]; in terms of regional flexibility, [regional flexibility information]; and shows good growth, [development potential highlights]").The text generated for each employee is used as the employee's recommendation reason summary, and the recommendation reason summary of each employee is obtained through the above method; finally, the employee list, the detailed profile data (for example: name, employee number, current position, and current store) of each employee in the list, and the corresponding recommendation reason summary are integrated and packaged into a structured data object, which is usually defined in a common data exchange format, for example, it can be a JSON object, which contains a field representing the success status, a field representing the total number of records, and an array field named resultList. Each element in the resultList array is an object representing a recommended employee, which contains the employee's identification field, basic information field, and a field specifically for storing the "recommendationReason" text. After packaging, the entire structured data object is returned to the search terminal that initially initiated the search request through the application programming interface. The search terminal parses the data object after receiving it and calls its user interface components to present the employee list, detailed information, and recommendation reasons in a clear and readable list or card form on the screen interface to the user, completing the visual display; the packaged and returned structured data object is the endpoint of the entire search process.

[0044] In addition, another aspect of the present application, in some embodiments, the present application provides a talent management information retrieval system for a chain enterprise, which comprises a retrieval unit, wherein Figure 3 The figure is a structural schematic diagram of a retrieval unit according to some embodiments of the present application, which comprises an initialization module 201, a processing module 202, and an execution module 203, which are described as follows: The initialization module 201 is mainly used for receiving a search request initiated by a search terminal from the central talent database of a chain enterprise, and then extracting rigid constraint conditions and flexible matching conditions in the search request; The processing module 202 is used for vectorizing the flexible matching conditions to obtain a target skill set of the search request, and then performing a preliminary screening in the central talent database according to the rigid constraint conditions to generate a first candidate set of the search request; It should be noted that the processing module 202 is also used to calculate the vector similarity between the actual skill vector of each employee in the first candidate set and the skill requirement vector in the target skill set, to obtain the core matching score of each employee, while determining the regional flexibility additional score and the skill development potential additional score of each employee in the first candidate set, and using the regional flexibility additional score and the skill development potential additional score to weight and correct the core matching score of the corresponding employee, to generate the competency score of each employee in the first candidate set; The execution module 203 is mainly used to sort the employees in the first candidate set according to all the comprehensive competency scores, and return the sorted employee list to the search terminal of the chain enterprise as a search result.

[0045] The above describes the examples of the talent management information search system and method for a chain enterprise provided by the embodiments of the present application in detail. It can be understood that the corresponding device includes the hardware structure and / or software module corresponding to the execution of each function in order to realize the above functions. Those skilled in the art should easily realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in the present application can be realized in the form of hardware or a combination of hardware and computer software. Whether a certain function is realized in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to realize the described functions for each specified application, but such implementation should not be considered beyond the scope of the present application.

[0046] In some embodiments, the present application also provides a computer device, which includes a memory and a processor, the memory is used to store a computer program, and the processor is used to call and run the computer program from the memory, so that the computer device executes the talent management information search method for a chain enterprise described above.

[0047] In some embodiments, with reference to Figure 4 The dashed line in the figure indicates that the unit or the module is optional, and the figure is a structural schematic diagram of a computer device for implementing the talent management information search method for a chain enterprise according to the embodiments of the present application. The talent management information search method for a chain enterprise described in the above embodiments can be realized by the computer device shown in the figure, which includes at least one processor 301, a memory 302 and at least one communication unit 305, and the computer device can be a terminal device or a server or a chip. Figure 4

[0048] ​The processor 301 can be a general processor or a special-purpose processor. For example, the processor 301 can be a central processing unit (CPU), which can be used to control a computer device, execute a software program, and process data of the software program. The computer device can further include a communication unit 305 to implement input (reception) and output (transmission) of signals.

[0049] For example, the computer device can be a chip, and the communication unit 305 can be an input and / or output circuit of the chip, or the communication unit 305 can be a communication interface of the chip, and the chip can be a component of a terminal device or a network device or other device.

[0050] For another example, the computer device can be a terminal device or a server, and the communication unit 305 can be a transceiver of the terminal device or the server, or the communication unit 305 can be a transceiver circuit of the terminal device or the server.

[0051] The computer device can include one or more memories 302, which store a program 304 that can be executed by the processor 301 to generate instructions 303, so that the processor 301 performs the method described in the above method embodiments according to the instructions 303. Optionally, the memory 302 can further store data (such as a target audit model). Optionally, the processor 301 can further read the data stored in the memory 302, and the data can be stored in the same storage address as the program 304, or the data can be stored in a different storage address from the program 304.

[0052] The processor 301 and the memory 302 can be separately arranged or integrated together, for example, integrated on a system on chip (SOC) of a terminal device.

[0053] It should be understood that each step of the above method embodiments can be completed by a logic circuit in the form of hardware or instructions in the form of software in the processor 301. The processor 301 can be a CPU, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, for example, discrete gates or transistor logic devices, or discrete hardware components.

[0054] Those skilled in the art will appreciate that embodiments of the present application can be readily used as a method, a system or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer-readable program code.

[0055] For example, in some embodiments, the present application also provides a computer-readable storage medium having stored therein instructions or codes, which, when executed on a computer, cause the computer to perform the above-mentioned method for retrieving talent management information for a chain store.

[0056] While the preferred embodiments of the application have been described, additional variations and modifications can be made to the embodiments by those skilled in the art once they learn of the basic inventive concepts. Therefore, the appended claims are intended to encompass within their scope all such variations and modifications as are included within the scope of the application.

[0057] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A human resource management information search method for a chain store, characterized by comprising: The method comprises the following steps: The central talent database of the chain enterprise receives a search request initiated by a search terminal, and then extracts rigid constraint conditions and flexible matching conditions in the search request; The flexible matching conditions are vectorized and mapped to obtain a target skill set of the search request, and then a primary screening is performed in the central talent database according to the rigid constraint conditions to generate a first candidate set of the search request; The vector similarity between the actual skill vector of each employee in the first candidate set and the skill requirement vector in the target skill set is calculated to obtain the core matching score of each employee, and the regional flexibility additional score and the skill development potential additional score of each employee in the first candidate set are determined, and the core matching score of the corresponding employee is weighted and corrected using the regional flexibility additional score and the skill development potential additional score to generate the competency score of each employee in the first candidate set; The employees in the first candidate set are sorted according to all the comprehensive competency scores, and the sorted employee list is returned to the search terminal of the chain enterprise as a search result.

2. The method of claim 1, wherein, The vectorization and mapping of the flexible matching conditions to obtain the target skill set of the search request specifically comprises: Based on the post competency standard library of the chain enterprise, the flexible matching conditions are subjected to natural language keyword extraction to obtain flexible keywords of the search request; The flexible keywords are subjected to semantic similarity matching with standardized skill labels in the standard library, and then non-standard descriptions in the flexible matching conditions are mapped to standardized skill labels to obtain a plurality of skill characteristics; According to the preset weight of the standard skill label in the standard library and the priority modifier word in the search request, an initial weight coefficient is assigned to each skill characteristic to obtain the target skill set of the search request.

3. The method of claim 1, wherein, The primary screening in the central talent database according to the rigid constraint conditions to generate the first candidate set of the search request specifically comprises: The rigid constraint conditions are analyzed to identify the constraint types and constraint values in the rigid constraint conditions, the constraint types including: the current store range, the post sequence, the job level range and the in-service state; Based on the index of the central talent database, the employee records satisfying all the constraint types and constraint values are subjected to a quick intersection query to obtain the first candidate set of the search request.

4. The method of claim 1, wherein, The calculation of the vector similarity between the actual skill vector of each employee in the first candidate set and the skill requirement vector in the target skill set to obtain the core matching score of each employee specifically comprises: According to the skill labels and weight coefficients in the target skill set, the skill requirement vector in the target skill set is constructed; The actual skill vector of each employee is constructed through the skill item scores, skill authentication records and levels recorded in the historical performance data of the employee, and the vector dimension is aligned with the skill requirement vector; The similarity between each actual skill vector and the skill requirement vector is calculated to obtain the core matching score of each employee.

5. The method of claim 1, wherein, The determination of the regional flexibility additional score and the skill development potential additional score of each employee in the first candidate set specifically comprises: For each employee in the first candidate set, the number of cross-store transfers of the employee in a specified historical period is counted, and then the transfer activity of the employee is calculated, and the transfer willingness strength filled by the employee is obtained; The transfer activity and the transfer willingness strength are fused into the regional flexibility additional score of the employee; The skill growth characteristics of the employee are extracted from the performance data of the employee related to the target skill set in the historical performance period, and the training performance score of the employee is extracted from the relevant training records participated by the employee; The skill growth characteristics and the training performance score are fused and mapped into the skill development potential additional score of the employee, and then the regional flexibility additional score and the skill development potential additional score of each employee in the first candidate set are obtained.

6. The method of claim 1, wherein, The core matching score of the corresponding employee is weighted and corrected using the regional flexibility additional score and the skill development potential additional score, and the competency score of each employee in the first candidate set is generated, which specifically includes: Scene weighting coefficients are set for the regional flexibility additional score and the skill development potential additional score, and then the score correction values of the regional flexibility and the skill development potential are determined respectively; The core matching score of the employee is corrected using the score correction values, and the competency score of each employee in the first candidate set is obtained.

7. The method of claim 1, wherein, The employees in the first candidate set are sorted according to all the comprehensive competency scores, and the sorted employee list is returned to the search terminal of the chain enterprise as a search result, which specifically includes: All employees in the first candidate set are arranged in descending order according to the comprehensive competency scores, and an initial employee list of the search request is obtained; When multiple employees come from the same store, the initial employee list is adjusted according to the preset rules to avoid the risk of excessive loss of talents from a single store, and an employee list of the search request is obtained; A recommended reason abstract is generated for each employee in the employee list, and the abstract content integrates the core matching advantages, regional flexibility information and development potential highlights of the corresponding employee; The employee list with the recommended reason abstract is packaged as structured data and returned to the search terminal initiating the request for visual display.

8. A personnel management information retrieval system for a chain store, the personnel management information retrieval system for a chain store comprising a retrieval unit, characterized by, The search unit includes: An initialization module is configured to receive a search request initiated by a search terminal from a central talent database of a chain enterprise, and then extract rigid constraint conditions and flexible matching conditions in the search request; A processing module is configured to vectorize and map the flexible matching conditions to obtain a target skill set of the search request, and then perform a preliminary screening in the central talent database according to the rigid constraint conditions to generate a first candidate set of the search request; The processing module is further configured to calculate the vector similarity between the actual skill vector of each employee in the first candidate set to the target skill set and the skill requirement vector in the target skill set, to obtain the core matching score of each employee, and determine the regional flexibility additional score and the skill development potential additional score of each employee in the first candidate set, and use the regional flexibility additional score and the skill development potential additional score to weight and correct the core matching score of the corresponding employee to generate the competency score of each employee in the first candidate set. The execution module is configured to sort the employees in the first candidate set according to all the comprehensive competence scores, and return a sorted employee list as a search result to a search terminal of the chain enterprise.

9. A computer device, comprising: The computer device comprises a memory and a processor, the memory is configured to store a computer program, and the processor is configured to call and run the computer program from the memory, so that the computer device executes the talent management information search method for a chain enterprise according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores instructions or codes, when the instructions or codes are run on a computer, so that the computer executes the talent management information search method for a chain enterprise according to any one of claims 1 to 7.

Images