Talent portrait construction method and device

Abstract

The invention provides a talent portrait construction method and device. The talent portrait construction method comprises the steps of collecting multi-dimensional talent data of target scientific and technological talents; performing feature extraction on the multi-dimensional talent data to obtain talent tags; for each piece of multi-dimensional talent data, obtaining a plurality of secondary index scores; for each data dimension, multiplying the secondary index score corresponding to each field in the data dimension by a weight corresponding to the data dimension, and then summing to obtain a talent score corresponding to the data dimension; and generating a talent portrait based on the talent tag, the talent score corresponding to each data dimension and an ECharts chart library technology. In the scheme, multi-dimensional talent data is collected and integrated and analyzed by using a unified data specification, a personalized talent tag is generated by using a feature extraction technology, scores of scientific and technological talents in multiple dimensions are automatically calculated, and finally the talent portrait is generated by combining the scores of the talent tag in multiple dimensions. Therefore, the purpose of improving the comprehensiveness and objectivity of talent portraits is achieved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and device for constructing a talent portrait. Background Art

[0002] As companies' requirements for human resource management continue to increase, especially for core resources such as scientific and technological talents, how to accurately evaluate and effectively motivate them to promote their development has become an important issue facing companies.

[0003] Talent portrait is a comprehensive, objective and visual description of talents. It helps enterprises to understand the characteristics, advantages and development potential of talents more accurately, and provides a scientific basis for talent selection, training and management.

[0004] In the process of constructing existing talent portraits, the dimension of scientific and technological talent data is single, resulting in a lack of comprehensiveness and objectivity in the talent portraits. On the other hand, there is a lack of unified standards and rules in the process of data collection and processing, which may lead to inconsistencies between data from different sources, making it difficult to conduct effective integration and analysis. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides a method and device for constructing a talent portrait, so as to achieve the effective integration and analysis of multi-dimensional talent data and improve the comprehensiveness and objectivity of the talent portrait.

[0006] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0007] A first aspect of an embodiment of the present invention discloses a method for constructing a talent portrait, the method comprising:

[0008] Use multi-dimensional data templates to collect multi-dimensional talent data of target scientific and technological talents;

[0009] Extracting features from each of the multi-dimensional talent data to obtain a talent label;

[0010] For each of the multi-dimensional talent data, an initial score input by an expert is obtained, and the initial score is multiplied by the field coefficient of the corresponding field to obtain multiple secondary indicator scores; each of the fields belongs to one of the multiple data dimensions that need to be included in the talent portrait; a corresponding weight is preset for each data dimension;

[0011] For each of the data dimensions, the scores of the secondary indicators corresponding to each field in the data dimension are multiplied by the weight corresponding to the data dimension and then summed to obtain the talent score corresponding to the data dimension;

[0012] A talent portrait is generated based on the talent label, the talent score corresponding to each of the data dimensions and the ECharts chart library technology.

[0013] Preferably, before generating talent labels based on the preset label generation rules and the multi-dimensional talent data, the method further includes:

[0014] Based on the preset verification rules, verify whether the multi-dimensional talent data meets the requirements;

[0015] If not, return to the step of using the multi-dimensional data template to collect the multi-dimensional talent data input by the target scientific and technological talents until it is verified that the multi-dimensional talent data meets the requirements.

[0016] Preferably, extracting features from each of the multi-dimensional talent data to obtain talent labels includes:

[0017] Preprocess each of the multi-dimensional talent data to remove interference information;

[0018] Using natural language processing technology to extract features from each of the pre-processed multi-dimensional talent data to obtain multiple talent features;

[0019] Classification processing is performed based on each of the talent characteristics to obtain the talent label corresponding to the target scientific and technological talent.

[0020] Preferably, the method further comprises:

[0021] Input each of the multi-dimensional talent data into a pre-trained recurrent neural network model to obtain talent training recommendations corresponding to the target scientific and technological talents;

[0022] The talent portrait and the talent training suggestions are visualized and displayed at the same time.

[0023] Preferably, the training process of the recurrent neural network model includes:

[0024] Obtain multi-dimensional talent data corresponding to multiple scientific and technological talents and perform pre-processing;

[0025] Extracting features from the preprocessed multi-dimensional talent data corresponding to the multiple scientific and technological talents to obtain multiple target features, and adding labels to the multiple target features;

[0026] Dividing each of the target features after adding the labels to obtain a training set and a validation set;

[0027] Build the recurrent neural network model to be trained;

[0028] Using the training set, training the recurrent neural network model to be trained to obtain a trained recurrent neural network model;

[0029] Using the validation set to evaluate the trained recurrent neural network model;

[0030] If the evaluation passes, the trained recurrent neural network model is determined to be obtained.

[0031] A second aspect of an embodiment of the present invention discloses a device for constructing a talent portrait, the device comprising:

[0032] A collection unit, used to collect multi-dimensional talent data of target scientific and technological talents using a multi-dimensional data template;

[0033] An extraction unit, configured to extract features from each of the multi-dimensional talent data to obtain a talent label;

[0034] The first scoring unit is used to obtain an initial score input by an expert for each of the multi-dimensional talent data, and multiply the initial score by a field coefficient of a corresponding field to obtain multiple secondary indicator scores; each of the fields belongs to one of the multiple data dimensions that need to be included in the talent profile; and a corresponding weight is preset for each data dimension;

[0035] A second scoring unit is used to, for each of the data dimensions, multiply the secondary indicator scores corresponding to each field in the data dimension by the weight corresponding to the data dimension and then sum them up to obtain the talent score corresponding to the data dimension;

[0036] The portrait generation unit is used to generate a talent portrait based on the talent label, the talent score corresponding to each of the data dimensions and the ECharts chart library technology.

[0037] Preferably, the device further comprises:

[0038] The verification unit is used to verify whether the multi-dimensional talent data meets the requirements based on preset verification rules before generating talent labels based on preset label generation rules and the multi-dimensional talent data; if not, return to execute the step of using the multi-dimensional data template to collect the multi-dimensional talent data input by the target scientific and technological talents until it is verified that the multi-dimensional talent data meets the requirements.

[0039] Preferably, the extraction unit is specifically used for:

[0040] Preprocess each of the multi-dimensional talent data to remove interference information;

[0041] Using natural language processing technology to extract features from each of the pre-processed multi-dimensional talent data to obtain multiple talent features;

[0042] Classification processing is performed based on each of the talent characteristics to obtain the talent label corresponding to the target scientific and technological talent.

[0043] Preferably, the device further comprises:

[0044] A suggestion generating unit, used for inputting each of the multi-dimensional talent data into a pre-trained recurrent neural network model to obtain talent training suggestions corresponding to the target scientific and technological talents;

[0045] The talent portrait and the talent training suggestions are visualized and displayed at the same time.

[0046] Preferably, the device further comprises:

[0047] A training unit, used to obtain multi-dimensional talent data corresponding to multiple scientific and technological talents and perform pre-processing;

[0048] Extracting features from the preprocessed multi-dimensional talent data corresponding to the multiple scientific and technological talents to obtain multiple target features, and adding labels to the multiple target features;

[0049] Dividing each of the target features after adding the labels to obtain a training set and a validation set;

[0050] Build the recurrent neural network model to be trained;

[0051] Using the training set, training the recurrent neural network model to be trained to obtain a trained recurrent neural network model;

[0052] Using the validation set to evaluate the trained recurrent neural network model;

[0053] If the evaluation passes, the trained recurrent neural network model is determined to be obtained.

[0054] Based on the method and device for constructing a talent portrait provided by the above embodiment of the present invention, a multi-dimensional data template is used to collect multi-dimensional talent data of target scientific and technological talents; feature extraction is performed on each of the multi-dimensional talent data to obtain a talent label; for each of the multi-dimensional talent data, an initial score input by an expert is obtained, and the initial score is multiplied by the field coefficient of the corresponding field to obtain multiple secondary indicator scores; each of the fields belongs to one of the multiple data dimensions that need to be included in the talent portrait; a corresponding weight is preset for each data dimension; for each of the data dimensions, the secondary indicator scores corresponding to each field in the data dimension are multiplied by the weight corresponding to the data dimension and then summed to obtain the talent score corresponding to the data dimension; based on the talent label, the talent score corresponding to each of the data dimensions and the ECharts chart library technology, a talent portrait is generated. In this solution, by collecting multi-dimensional talent data and integrating and analyzing them using unified data specifications, using feature extraction technology to generate personalized talent labels, and automatically calculating the scores of scientific and technological talents in multiple dimensions, and finally combining the scores of multiple dimensions of the talent label to generate a talent portrait, the purpose of improving the comprehensiveness and objectivity of the talent portrait is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0056] Figure 1 A flowchart of a method for constructing a talent portrait disclosed in an embodiment of the present invention;

[0057] Figure 2 A flowchart of a method for generating a talent label disclosed in an embodiment of the present invention;

[0058] Figure 3 A flowchart of training a recurrent neural network model disclosed in an embodiment of the present invention;

[0059] Figure 4 This is a structural diagram of a talent portrait generation device disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0060] 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 only 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.

[0061] In this application, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0062] From the background technology, we can see that in the existing process of constructing talent portraits, the dimension of scientific and technological talent data is single, resulting in the lack of comprehensiveness and objectivity in the talent portraits. On the other hand, there is a lack of unified standards and rules in the process of data collection and processing, which may lead to inconsistencies between data from different sources, making it difficult to effectively integrate and analyze.

[0063] Therefore, an embodiment of the present invention discloses a method and device for constructing a talent portrait. In this solution, multi-dimensional talent data is collected and integrated and analyzed using a unified data specification, feature extraction technology is used to generate personalized talent labels, and the scores of scientific and technological talents in multiple dimensions are automatically calculated. Finally, a talent portrait is generated by combining the scores of the talent labels in multiple dimensions, thereby achieving the purpose of improving the comprehensiveness and objectivity of the talent portrait.

[0064] like Figure 1 FIG. 1 is a flowchart of a method for constructing a talent portrait disclosed in an embodiment of the present invention, which mainly includes the following steps:

[0065] Step S101: using a multi-dimensional data template to collect multi-dimensional talent data of target scientific and technological talents.

[0066] In step S101, the data collection method can be to integrate the scientific and technological talent data from different institutions and systems, covering multiple dimensions such as basic information, educational experience, project resume, etc., and / or the data collection method adopts the method of centralized filling in of scientific and technological talents by downloading templates, and setting multiple verification rules in the input interface to ensure the integrity, accuracy and consistency of the data.

[0067] In one embodiment, based on a preset verification rule, it is verified whether the multi-dimensional talent data meets the requirements; if so, step S102 is executed; if not, the process returns to step S101 until the multi-dimensional talent data is verified to meet the requirements.

[0068] Exemplarily, the verification rules are as follows:

[0069] 1. Data format verification:

[0070] (1) Ensure that the formats of all date, number and other fields are unified. For example, dates are unified in the "YYYY-MM-DD" format, and ID card numbers are unified in a specific number of digits.

[0071] (2) Automatically standardize text fields, such as removing spaces and unifying upper and lower case.

[0072] 2. Data consistency check:

[0073] (1) Verification of the relationship between fields: Check whether there are logical contradictions between fields, such as the date of employment being earlier than the date of graduation, whether there is any overlap in working hours, etc.

[0074] (2) Multi-source data comparison: If the data comes from multiple systems or channels, a comparison is performed to ensure data consistency. For example, if the academic degree reported manually is a bachelor's degree, but the academic degree automatically captured by the system is a master's degree, the system will automatically feedback the error information for the person filling out the form to modify and confirm.

[0075] 3. Data rule verification:

[0076] (1) Set data validation rules based on business needs. For example, age should be within a reasonable range (e.g., 18-40 years old) and educational background should be a valid educational level.

[0077] In case of errors in filling out the form, error information will be automatically fed back for the person filling out the form to modify; after submission, the administrator will manually review the data after preliminary screening and rule verification to ensure the accuracy and completeness of the data.

[0078] Step S102: extract features from each multi-dimensional talent data to obtain a talent label.

[0079] In step S102, relying on multi-dimensional talent data, natural language processing (NLP) technology is introduced to perform text analysis on the collected unstructured data (such as project descriptions, etc.) through NLP technology to extract key information. Subsequently, feature extraction is performed on this information to intelligently create personalized labels for scientific and technological talents. The labels cover multiple dimensions such as basic capabilities, R&D capabilities, professional qualities, and academic information. Each label corresponds to a list of scientific and technological talents with similar characteristics or capabilities. It is convenient for enterprises to formulate talent management strategies more accurately based on the outstanding capabilities of scientific and technological talents.

[0080] See also Figure 2 , is a flow chart of a method for generating a talent label disclosed in an embodiment of the present invention, and the specific implementation method includes the following steps:

[0081] Step S201: pre-process each multi-dimensional talent data to remove interference information.

[0082] In step S201, the unstructured data (such as project description, work experience, research results, etc.) in each multi-dimensional talent data is cleaned and preprocessed, including steps such as removing noise (such as irrelevant characters, special symbols, etc.), laying the foundation for subsequent analysis.

[0083] Step S202: Using natural language processing technology to extract features from each pre-processed multi-dimensional talent data to obtain multiple talent features.

[0084] In step S202, the bag-of-words model and word embedding technology in NLP are used to extract key features such as skills, project types, and research fields from the pre-processed multi-dimensional talent data text to obtain multiple talent features.

[0085] Step S203: Perform classification processing based on the characteristics of each talent to obtain a talent label corresponding to the target scientific and technological talent.

[0086] In step S203, target scientific and technological talents are classified based on multiple talent characteristics, for example, by identifying technical terminology characteristics and achievement description characteristics in project descriptions, talents are classified into different professional fields or skill levels, and personalized talent labels are generated.

[0087] Step S103: For each multi-dimensional talent data, obtain the initial score input by the expert, multiply the initial score by the field coefficient of the corresponding field, and obtain multiple secondary indicator scores.

[0088] Among them, each field belongs to one of the multiple data dimensions that need to be included in the talent portrait; a corresponding weight is preset for each data dimension.

[0089] In step S103, since there is a lot of multi-dimensional talent data and they belong to different fields, it is very difficult for experts to score directly based on the importance of the field. Therefore, the initial score range entered by the expert is fixed, such as 0-10 points, or 0-100 points. After multiplying the initial score by the field coefficient, the secondary indicator score reflecting the importance of the field can be obtained.

[0090] It should be noted that in the process of designing the points system, the key elements that constitute the talent profile are determined based on the talent selection goals and training plans, covering four dimensions: basic abilities, R&D abilities, professional qualities, and comprehensive qualities.

[0091] The basic abilities include writing, speech, multimedia, PPT, style and innovation; R&D abilities include development abilities in seven areas: front-end, back-end, architecture, UI, APP / applets, testing and database; comprehensive qualities include experience in six areas: project experience, training experience, secondment, honorary certificates, and participation in competitions and activities; professional qualities include abilities in eight areas: development, data, demand, security, network, operation and maintenance, system and project management. For each of the above areas, a corresponding field coefficient is set based on expert experience.

[0092] Finally, according to the importance of each data dimension, a reasonable weight distribution is set to ensure the fairness and scientificity of the points.

[0093] Step S104: For each data dimension, multiply the secondary indicator scores corresponding to each field in the data dimension by the weight corresponding to the data dimension and sum them up to obtain the talent score corresponding to the data dimension.

[0094] For example, under data dimension 1, there are domain 1, domain 2 and domain 3, and the corresponding initial scores are score 1, score 2 and score 3. The weight corresponding to the data dimension is a. The calculation formula is: score 1*a+score 2*a+score 3*a=the talent score corresponding to data dimension 1.

[0095] Step S105: Generate a talent portrait based on the talent label, the talent score corresponding to each data dimension and the ECharts chart library technology.

[0096] In step S105, a radar chart of scientific and technological talents (i.e., talent portrait) is automatically formed through the ECharts chart library technology. With its unique graphic structure, it clearly shows the scores of scientific and technological talents in various dimensions. At the same time, it is supplemented by personalized label display, covering the key characteristics and outstanding performance of scientific and technological talents, forming a vivid and three-dimensional portrait of scientific and technological talents, and realizing a visual evaluation of the performance of talents in various dimensions.

[0097] In one embodiment, various multi-dimensional talent data are input into a pre-trained recurrent neural network model to obtain talent training recommendations corresponding to target scientific and technological talents; the talent portraits and talent training recommendations are visualized and displayed at the same time.

[0098] In an embodiment of the present invention, with the help of AI artificial intelligence technology, a recurrent neural network (RNN) model in deep learning is adopted to intelligently analyze the advantages and disadvantages of scientific and technological talents in different dimensions. The multi-dimensional talent data of scientific and technological talents is used as input, and feature extraction and pattern recognition are performed through a multi-layer neural network. Personalized talent training suggestions are further given, including career development path planning, etc., to help enterprises better manage talents and improve the scientific and technological capabilities of employees.

[0099] like Figure 3 FIG. 1 is a flowchart of a training method of a recurrent neural network model disclosed in an embodiment of the present invention, comprising the following steps:

[0100] Step S301: Obtain multi-dimensional talent data corresponding to multiple scientific and technological talents and perform pre-processing.

[0101] In step S301, the multi-dimensional talent data corresponding to the collected multiple scientific and technological talents are cleaned, denoised and standardized to ensure the quality and consistency of the data.

[0102] Step S302: extracting features from the preprocessed multi-dimensional talent data corresponding to the multiple scientific and technological talents to obtain multiple target features, and adding labels to the multiple target features.

[0103] In step S302, it is feature engineering, that is, according to business needs and data characteristics, extract and construct target features that have an important impact on talent training recommendations, such as skill proficiency, project experience, training participation, etc.; use feature selection technology to screen out target features that have an important impact on talent training recommendations, reduce the complexity of the model, and improve prediction accuracy. Add labels (such as academic information, project experience, etc.) to the target features.

[0104] Step S303: Divide each target feature after adding labels to obtain a training set and a validation set.

[0105] Step S304: construct a recurrent neural network model to be trained.

[0106] Step S305: using the training set to train the recurrent neural network model to be trained, to obtain a trained recurrent neural network model.

[0107] In step S305, the training set is input into the RNN model. During the training process, early stopping method and learning rate decay technology are used to prevent the model from overfitting and ensure the generalization ability of the model so that the model can accurately predict and generate personalized talent training suggestions.

[0108] Step S306: Evaluate the trained recurrent neural network model using the validation set.

[0109] In step S306, the performance of the model is evaluated using indicators such as accuracy, recall, and F1 score to ensure the performance and reliability of the model.

[0110] If the evaluation fails, the training parameters are continuously adjusted according to the performance on the validation set, and finally the learning rate is set to 0.001, the batch size is 64, and the training rounds are 55. The weights and bias parameters of the model are adjusted through the back propagation algorithm and the gradient descent method to minimize the loss function.

[0111] Step S307: If the evaluation passes, it is determined that a trained recurrent neural network model is obtained.

[0112] Based on the method for constructing a talent portrait disclosed in the above embodiment of the present invention, a multi-dimensional data template is used to collect multi-dimensional talent data of target scientific and technological talents; feature extraction is performed on each multi-dimensional talent data to obtain a talent label; for each multi-dimensional talent data, an initial score input by an expert is obtained, and the initial score is multiplied by the field coefficient of the corresponding field to obtain multiple secondary indicator scores; each field belongs to one of the multiple data dimensions that need to be included in the talent portrait; a corresponding weight is preset for each data dimension; for each data dimension, the secondary indicator scores corresponding to each field in the data dimension are multiplied by the weight corresponding to the data dimension and then summed to obtain the talent score corresponding to the data dimension; based on the talent label, the talent score corresponding to each data dimension and the ECharts chart library technology, a talent portrait is generated. In this solution, by collecting multi-dimensional talent data and integrating and analyzing it using a unified data specification, using feature extraction technology to generate personalized talent labels, and automatically calculating the scores of scientific and technological talents in multiple dimensions, and finally combining the scores of multiple dimensions of the talent label to generate a talent portrait, the purpose of improving the comprehensiveness and objectivity of the talent portrait is achieved.

[0113] Corresponding to the method for generating a talent portrait disclosed in the above embodiment of the present invention, Figure 4 As shown, it is a structural diagram of a talent portrait generation device disclosed in an embodiment of the present invention, including: a collection unit 401, an extraction unit 402, a first scoring unit 403, a second scoring unit 404 and a portrait generation unit 405.

[0114] The collection unit 401 is used to collect multi-dimensional talent data of target scientific and technological talents using a multi-dimensional data template.

[0115] In one embodiment, the device further comprises:

[0116] The verification unit is used to verify whether the multi-dimensional talent data meets the requirements based on preset verification rules before generating talent tags based on preset tag generation rules and multi-dimensional talent data; if not, it returns to execute the step of using the multi-dimensional data template to collect the multi-dimensional talent data input by the target scientific and technological talents until the multi-dimensional talent data is verified to meet the requirements.

[0117] The extraction unit 402 is used to extract features from each multi-dimensional talent data to obtain a talent label.

[0118] In one embodiment, the extraction unit 402 is specifically configured to:

[0119] Pre-process each multi-dimensional talent data to remove interference information;

[0120] Use natural language processing technology to extract features from each multi-dimensional talent data after preprocessing to obtain multiple talent features;

[0121] Classification is performed based on the characteristics of each talent to obtain the talent labels corresponding to the target scientific and technological talents.

[0122] The first scoring unit 403 is used to obtain the initial score input by the expert for each multi-dimensional talent data, multiply the initial score by the field coefficient of the corresponding field, and obtain multiple secondary indicator scores; each field belongs to one of the multiple data dimensions that need to be included in the talent portrait; and a corresponding weight is preset for each data dimension.

[0123] The second scoring unit 404 is used to multiply the secondary indicator scores corresponding to each field in the data dimension by the weight corresponding to the data dimension and then sum them up to obtain the talent score corresponding to the data dimension.

[0124] The portrait generation unit 405 is used to generate a talent portrait based on the talent label, the talent score corresponding to each data dimension and the ECharts chart library technology.

[0125] In one embodiment, the device further comprises:

[0126] A suggestion generation unit is used to input various multi-dimensional talent data into a pre-trained recurrent neural network model to obtain talent training suggestions corresponding to target scientific and technological talents;

[0127] Talent portraits and talent development suggestions are visualized and displayed at the same time.

[0128] In one embodiment, the device further comprises:

[0129] A training unit, used to obtain multi-dimensional talent data corresponding to multiple scientific and technological talents and perform pre-processing;

[0130] Extract features from the multi-dimensional talent data corresponding to the pre-processed multiple scientific and technological talents to obtain multiple target features, and add labels to the multiple target features;

[0131] Divide each target feature after adding labels to obtain a training set and a validation set;

[0132] Build the recurrent neural network model to be trained;

[0133] Using the training set, training the recurrent neural network model to be trained to obtain a trained recurrent neural network model;

[0134] Use the validation set to evaluate the trained recurrent neural network model;

[0135] If the evaluation passes, the trained recurrent neural network model is determined to be obtained.

[0136] Based on the above-mentioned talent portrait construction device disclosed in the embodiment of the present invention, a multi-dimensional data template is used to collect multi-dimensional talent data of target scientific and technological talents; feature extraction is performed on each multi-dimensional talent data to obtain a talent label; for each multi-dimensional talent data, the initial score input by the expert is obtained, and the initial score is multiplied by the field coefficient of the corresponding field to obtain multiple secondary indicator scores; each field belongs to one of the multiple data dimensions that need to be included in the talent portrait; a corresponding weight is preset for each data dimension; for each data dimension, the secondary indicator scores corresponding to each field in the data dimension are multiplied by the weight corresponding to the data dimension and then summed to obtain the talent score corresponding to the data dimension; based on the talent label, the talent score corresponding to each data dimension and the ECharts chart library technology, a talent portrait is generated. In this solution, by collecting multi-dimensional talent data and integrating and analyzing it using a unified data specification, using feature extraction technology to generate personalized talent labels, and automatically calculating the scores of scientific and technological talents in multiple dimensions, and finally combining the scores of multiple dimensions of the talent label to generate a talent portrait, the purpose of improving the comprehensiveness and objectivity of the talent portrait is achieved.

[0137] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0138] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0139] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for constructing a talent portrait, characterized in that: The method comprises: Use multi-dimensional data templates to collect multi-dimensional talent data of target scientific and technological talents; Extracting features from each of the multi-dimensional talent data to obtain a talent label; For each of the multi-dimensional talent data, an initial score input by an expert is obtained, and the initial score is multiplied by the field coefficient of the corresponding field to obtain multiple secondary indicator scores; each of the fields belongs to one of the multiple data dimensions that need to be included in the talent portrait; a corresponding weight is preset for each data dimension; For each of the data dimensions, the scores of the secondary indicators corresponding to each field in the data dimension are multiplied by the weight corresponding to the data dimension and then summed to obtain the talent score corresponding to the data dimension; A talent portrait is generated based on the talent label, the talent score corresponding to each of the data dimensions and the ECharts chart library technology.

2. The method according to claim 1, characterized in that Before generating talent labels based on the preset label generation rules and the multi-dimensional talent data, the method further includes: Based on the preset verification rules, verify whether the multi-dimensional talent data meets the requirements; If not, return to the step of using the multi-dimensional data template to collect the multi-dimensional talent data input by the target scientific and technological talents until it is verified that the multi-dimensional talent data meets the requirements.

3. The method according to claim 1, characterized in that: The extracting features of each of the multi-dimensional talent data to obtain talent labels includes: Preprocess each of the multi-dimensional talent data to remove interference information; Using natural language processing technology to extract features from each of the pre-processed multi-dimensional talent data to obtain multiple talent features; Classification processing is performed based on each of the talent characteristics to obtain the talent label corresponding to the target scientific and technological talent.

4. The method according to claim 1, characterized in that The method further comprises: Input each of the multi-dimensional talent data into a pre-trained recurrent neural network model to obtain talent training recommendations corresponding to the target scientific and technological talents; The talent portrait and the talent training suggestions are visualized and displayed at the same time.

5. The method according to claim 4, characterized in that The training process of the recurrent neural network model includes: Obtain multi-dimensional talent data corresponding to multiple scientific and technological talents and perform pre-processing; Extracting features from the preprocessed multi-dimensional talent data corresponding to the multiple scientific and technological talents to obtain multiple target features, and adding labels to the multiple target features; Dividing each of the target features after adding the labels to obtain a training set and a validation set; Build a recurrent neural network model to be trained; Using the training set, training the recurrent neural network model to be trained to obtain a trained recurrent neural network model; Using the validation set to evaluate the trained recurrent neural network model; If the evaluation passes, the trained recurrent neural network model is determined to be obtained.

6. A device for constructing a talent portrait, characterized in that: The device comprises: A collection unit, used to collect multi-dimensional talent data of target scientific and technological talents using a multi-dimensional data template; An extraction unit, configured to extract features from each of the multi-dimensional talent data to obtain a talent label; The first scoring unit is used to obtain an initial score input by an expert for each of the multi-dimensional talent data, and multiply the initial score by a field coefficient of a corresponding field to obtain multiple secondary indicator scores; each of the fields belongs to one of the multiple data dimensions that need to be included in the talent profile; and a corresponding weight is preset for each data dimension; A second scoring unit is used to, for each of the data dimensions, multiply the secondary indicator scores corresponding to each field in the data dimension by the weight corresponding to the data dimension and then sum them up to obtain the talent score corresponding to the data dimension; The portrait generation unit is used to generate a talent portrait based on the talent label, the talent score corresponding to each of the data dimensions and the ECharts chart library technology.

7. The device according to claim 6, characterized in that The device also includes: The verification unit is used to verify whether the multi-dimensional talent data meets the requirements based on preset verification rules before generating talent labels based on preset label generation rules and the multi-dimensional talent data; if not, return to execute the step of using the multi-dimensional data template to collect the multi-dimensional talent data input by the target scientific and technological talents until it is verified that the multi-dimensional talent data meets the requirements.

8. The device according to claim 6, characterized in that The extraction unit is specifically used for: Preprocess each of the multi-dimensional talent data to remove interference information; Using natural language processing technology to extract features from each of the pre-processed multi-dimensional talent data to obtain multiple talent features; Classification processing is performed based on each of the talent characteristics to obtain the talent label corresponding to the target scientific and technological talent.

9. The device according to claim 6, characterized in that The device also includes: A suggestion generating unit, used for inputting each of the multi-dimensional talent data into a pre-trained recurrent neural network model to obtain talent training suggestions corresponding to the target scientific and technological talents; The talent portrait and the talent training suggestions are visualized and displayed at the same time.

10. The device according to claim 9, characterized in that The device also includes: A training unit, used to obtain multi-dimensional talent data corresponding to multiple scientific and technological talents and perform pre-processing; Extracting features from the preprocessed multi-dimensional talent data corresponding to the multiple scientific and technological talents to obtain multiple target features, and adding labels to the multiple target features; Dividing each of the target features after adding the labels to obtain a training set and a validation set; Build a recurrent neural network model to be trained; Using the training set, training the recurrent neural network model to be trained to obtain a trained recurrent neural network model; Using the validation set to evaluate the trained recurrent neural network model; If the evaluation passes, the trained recurrent neural network model is determined to be obtained.

Images