Skill Word Evaluation Method and Apparatus, Electronic Device, Computer Readable Medium

Abstract

The present disclosure provides a method for evaluating skill words in a resume, which relates to the technical field of machine learning. The method includes: determining a first list of skill words to be evaluated from a resume document to be evaluated, where the first list of skill words includes multiple skill words; for each skill word in the first list of skill words, using a pre-trained skill word evaluation model and the context information of the skill word in the first list of skill words, predicting the probability value of the occurrence of the skill word, and the probability value is used to characterize the importance of the skill word. The present disclosure also provides a skill word evaluation device, an electronic device, and a computer-readable medium.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the technical field of machine learning, and in particular to a resume skill word evaluation method and device, an electronic device, and a computer-readable medium. Background Art

[0002] At present, in the process of corporate recruitment, recruiters are often faced with hundreds or thousands of resumes. On the one hand, recruiters often use manual identification, judgment, and screening methods, which causes recruiters to spend a lot of time identifying effective information in resumes in order to screen out talents that meet corporate needs from millions of resumes; on the other hand, companies often have different professional requirements for different positions, especially in terms of professional skills. Due to the limited knowledge of recruiters, they cannot effectively identify all professional skills in resumes, resulting in the loss of excellent resumes.

[0003] Therefore, how to help recruiters improve the efficiency and accuracy of resume screening and target talent screening has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] The embodiments of the present disclosure provide a resume skill word evaluation method and device, an electronic device, and a computer-readable medium.

[0005] In a first aspect, an embodiment of the present disclosure provides a method for evaluating skill words in a resume, the method comprising:

[0006] Determining a first skill word list to be evaluated from the resume document to be evaluated, wherein the first skill word list includes a plurality of skill words;

[0007] For each skill word in the first skill word list, a pre-trained skill word evaluation model and context information of the skill word in the first skill word list are used to predict a probability value of the skill word appearing, and the probability value is used to characterize the importance of the skill word.

[0008] In a second aspect, an embodiment of the present disclosure provides a skill word evaluation device, the skill word evaluation device comprising:

[0009] A skill word acquisition module, used to determine a first skill word list to be evaluated from the resume document to be evaluated, wherein the first skill word list includes a plurality of skill words;

[0010] A skill word evaluation module is used to predict the probability value of each skill word in the first skill word list by using a pre-trained skill word evaluation model and the context information of the skill word in the first skill word list, and the probability value is used to represent the importance of the skill word.

[0011] In a third aspect, embodiments of the present disclosure provide an electronic device, which includes:

[0012] One or more processors;

[0013] A memory storing one or more programs, which when executed by the one or more processors, cause the one or more processors to implement the skill word evaluation method provided in any embodiment of the present disclosure.

[0014] In a fourth aspect, embodiments of the present disclosure provide a computer-readable medium having a computer program stored thereon, wherein when the computer program is executed, it implements the skill word evaluation method provided in any embodiment of the present disclosure.

[0015] The skill word evaluation method and apparatus for resumes, electronic device, and computer-readable medium provided by the embodiments of the present disclosure improve the accuracy of skill word evaluation, enhance the resume screening efficiency, and greatly save the time cost of manual screening and evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings are used to provide a further understanding of the embodiments of the present disclosure, and constitute a part of the specification. Together with the embodiments of the present disclosure, they are used to explain the present disclosure, and do not constitute a limitation to the present disclosure. By describing the detailed exemplary embodiments with reference to the drawings, the above and other features and advantages will become more apparent to those skilled in the art. In the drawings:

[0017] Figure 1 Is a flowchart of a skill word evaluation method for a resume provided by an embodiment of the present disclosure;

[0018] Figure 2 Is Figure 1 A flowchart of a specific implementation manner of step 11 in;

[0019] Figure 3 Is Figure 2 A flowchart of a specific implementation manner of step 111 in;

[0020] Figure 4 Is a flowchart of a training method for a skill word evaluation model in an embodiment of the present disclosure;

[0021] Figure 5 Is a schematic diagram of a neural network structure of a word embedding model in an embodiment of the present disclosure;

[0022] Figure 6 Is a block diagram of a skill word evaluation apparatus provided by an embodiment of the present disclosure;

[0023] Figure 7 Is Figure 6Block diagram of the composition of a skill word acquisition module in

[0024] Figure 8 Block diagram of the composition of another skill word evaluation device provided by an embodiment of the present disclosure;

[0025] Figure 9 Block diagram of the composition of an electronic device provided by an embodiment of the present disclosure. Detailed implementation manners

[0026] To enable those skilled in the art to better understand the technical solutions of the present disclosure, the skill word evaluation method and device, electronic device, and computer-readable medium for resumes provided by the present disclosure will be described in detail below with reference to the accompanying drawings.

[0027] In the following, example embodiments will be described more fully with reference to the accompanying drawings, but the example embodiments may be embodied in different forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0028] In the case of no conflict, the various embodiments of the present disclosure and the various features in the embodiments may be combined with each other.

[0029] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0030] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. As used herein, the singular forms "a" and "the" are also intended to include the plural forms unless the context clearly indicates otherwise. It will also be understood that when the terms "include" and / or "consist of" are used in this specification, the specified features, wholes, steps, operations, elements, and / or components are present, but one or more other features, wholes, steps, operations, elements, components, and / or groups thereof are not excluded.

[0031] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and the present disclosure, and will not be interpreted as having an idealized or overly formal meaning unless clearly defined herein.

[0032] Figure 1 Flowchart of a skill word evaluation method for resumes provided by an embodiment of the present disclosure, as Figure 1As shown, this method can be executed by a skill word evaluation device, which can be implemented in software and / or hardware, and can be integrated into an electronic device such as a server. The skill word evaluation method includes Step 11 and Step 12.

[0033] Step 11: Determine a first list of skill words to be evaluated from the resume document to be evaluated. The first list of skill words includes multiple skill words.

[0034] In some application scenarios, after obtaining one or more resume documents of job seekers, the recruiter can send the resume documents to the skill word evaluation device for evaluation. Among them, the channels through which the recruiter obtains the resume documents can be through recruitment email, recruitment website, recruitment client, etc. In some application scenarios, after receiving the resume documents of job seekers, recruitment email, recruitment website, recruitment client, etc. can also automatically forward the resume documents to the skill word evaluation device. In some application scenarios, the skill word evaluation device can also obtain the resume documents of job seekers from recruitment email, recruitment website, recruitment client, etc. by actively querying at preset intervals (such as 10 minutes, 20 minutes, etc.). In some application scenarios, the resume document can also be a paper resume document. After the recruiter obtains the paper resume document, it can be sent to the skill word evaluation device after converting the paper resume document into an electronic resume document by scanning.

[0035] In the embodiments of the present disclosure, after receiving the resume document, the skill word evaluation device performs the operations of Step 11 and Step 12 for each resume document, so as to complete the automatic evaluation of the skill words of each resume document. In some embodiments, after completing the skill word evaluation of each resume document, the skill word evaluation device can also display the skill word evaluation results of each resume document to the recruiter through a suitable manner such as a human-computer interaction interface, so that the recruiter can quickly and accurately obtain the resume skill profile of the job seeker and complete the resume screening.

[0036] Figure 2 For Figure 1 a flowchart of a specific implementation manner of Step 11 in Figure 2 As shown, in some embodiments, Step 11 includes Step 111 to Step 113.

[0037] Step 111: Determine a second list of skill words from the resume document. The second list of skill words includes all the skill words that appear in the resume document.

[0038] Figure 3 For Figure 2 a flowchart of a specific implementation manner of Step 111 in Figure 3 As shown, in some embodiments, Step 111 includes Step 1111 and Step 1112.

[0039] Step 1111: Obtain resume text data from the resume document.

[0040] Specifically, in Step 1111, after obtaining the resume document, standardize and format the resume document to obtain the resume text data in the resume document, where the resume text data includes text data such as work experience descriptions, project experience descriptions, and personal professional skill descriptions.

[0041] Step 1112: Extract all skill words that appear in the resume text data from the resume text data to generate a second skill word list.

[0042] Specifically, in Step 1112, for the resume text data, first use a preset word segmentation tool to perform word segmentation on the resume text data to obtain the word segmentation result, and the word segmentation result contains each word in the resume text data.

[0043] Then, use a preset domain skill word library to screen out all skill words that appear in the resume text data from the word segmentation result. Specifically, the words obtained by word segmentation can be matched with the skill words in the skill word library. If the match is consistent, it indicates that the word is a skill word. Among them, the skill word can be a skill word in Chinese, an English skill word, or a Chinese-English abbreviation skill word.

[0044] In Step 1112, after filtering out the non-skill words in the resume text data through the domain skill word library, all skill words that appear in the resume text data are obtained, and a second skill word list is generated based on the all skill words.

[0045] Step 112: Determine the technical field to which each skill word in the second skill word list belongs.

[0046] In order to facilitate recruiters' understanding of skill words, in some embodiments, it is necessary to identify the technical fields to which the skill words belong. Specifically, in step 112, using a preset knowledge graph, the technical field to which each skill word in the second skill word list belongs is determined. Among them, the knowledge graph contains the corresponding relationship between skill words and their technical fields, and the technical field can include multiple skill words. For example, the skill: "TensorFlow" belongs to the skill in the "deep learning" field. Since recruiters may not be familiar with some skill words (such as TensorFlow), it may cause a large deviation in the recruiters' understanding of job seekers' resumes. Therefore, in some embodiments, by introducing a preset knowledge graph of technical fields and skill words, the hyponymy relationship, similarity relationship, etc. of skill words are extended, and the description of skill words is reasonably standardized, which can not only standardize the input of the model in subsequent steps, but also improve the readability of the model output results and strengthen the recruiters' understanding of the skill words in the resume.

[0047] Step 113: Generate a first skill word list based on all the skill words in the second skill word list and their corresponding technical fields, with each technical field serving as a skill word.

[0048] In some embodiments, in step 113, after obtaining all the skill words that appear in the resume document and identifying the technical fields to which each skill word belongs, the technical field itself is also regarded as a skill word, and a first skill word list is generated based on all the skill words that appear in the resume document and their corresponding technical fields. In the first skill word list, each technical field serves as a skill word.

[0049] Step 12: For each skill word in the first skill word list, use a pre-trained skill word evaluation model and the context information of the skill word in the first skill word list to predict the probability value of the appearance of the skill word, and this probability value is used to represent the importance of the skill word.

[0050] It can be understood that the context information of the skill word in the first skill word list includes other skill words in the first skill word list except this skill word. In step 12, the input of the pre-trained skill word evaluation model is the word vectors corresponding to other skill words except this skill word, and its output is the probability value of the appearance of this skill word, that is, the probability of the appearance of this skill word when other skill words in the resume document are known. This probability value can represent the importance of the corresponding skill word. The larger the probability value, the higher the importance of the skill word.

[0051] Specifically, in step 12, first, for each skill word in the first skill word list except this skill word, a word vector corresponding to each skill word except this skill word is generated. The word vector corresponding to each skill word can be generated by the one-hot encoding method.

[0052] Then, the word vectors corresponding to each skill word except this skill word are used as the input of a pre-trained skill word evaluation model, and the probability value of the occurrence of this skill word is predicted by using the skill word evaluation model.

[0053] Using the pre-trained skill word evaluation model, each skill word in the first skill word list is predicted to obtain the probability value of the occurrence of each skill word in the first skill word list.

[0054] Figure 4 The figure is a flowchart of a training method for a skill word evaluation model in an embodiment of the present disclosure. In some embodiments, as Figure 4 shown, the skill word evaluation model is trained through the following steps:

[0055] Step 21: Obtain a training data set. The training data set includes a training skill word list extracted from resume samples, and the training skill word list includes multiple training skill words.

[0056] Among them, the multiple training skill words include the skill words extracted from resume samples and their corresponding technical fields.

[0057] Step 22: Generate word vectors corresponding to each training skill word.

[0058] In some embodiments, by performing one-hot encoding processing on each training skill word, a word vector corresponding to each training skill word can be obtained.

[0059] Step 23: For each training skill word, use the word vectors corresponding to each training skill word except this training skill word as the input, and use a preset word embedding model for model training. The output of the word embedding model is the probability value of the occurrence of this training skill word.

[0060] Among them, the word vectors corresponding to each training skill word except this training skill word are respectively denoted as x 1 , x 2 , …, x C , and C represents the total number of other training skill words except this training skill word.

[0061] In some embodiments, the word embedding model includes a continuous bag-of-words neural network model (CBOW). Figure 5 The figure is a schematic diagram of the neural network structure of a word embedding model in an embodiment of the present disclosure. As Figure 5As shown, the word embedding model includes an input layer (Inputlayer), a hidden layer (Hidden layer), and an output layer (Output layer).

[0062] Among them, the input layer has inputs of C training skill words: {x 1 , x 2 , …, x C}, where the window size is C and the vocabulary length is V, and V represents the total number of skill words in the domain skill word library.

[0063] The hidden layer is an N-dimensional vector, where N is the number of neurons in the hidden layer. The output of the hidden layer h is expressed as follows:

[0064]

[0065] Among them, W T is an N*V-dimensional weight matrix from the input layer to the hidden layer, h is the output of the hidden layer, h represents the weighted average of the word vectors corresponding to C training skill words, and x 1 , x 2 , …, x C represent the word vectors corresponding to other training skill words except this training skill word respectively.

[0066] The input of the output layer is a V1-dimensional vector u, u = W′ T ·h, where W′ T is an N*V-dimensional weight matrix from the hidden layer to the output layer. The j-th element u j of the vector u is the inner product of the j-th column of W′ T and the output h of the hidden layer, that is Among them, represents the j-th column of W′ T , and u j represents the score of the j-th skill word in the domain skill word library. The skill word with the highest score is taken as the predicted output skill word, and the Softmax (logistic regression) function is used to normalize the vector u to between [0, 1], so as to predict the probability of the output skill word, and finally obtain the output vector y j of the output layer. Among them, the output vector y j is expressed as follows:

[0067]

[0068] Among them, x i represents the i-th skill word in the training skill word table, and contex(x i ) represents the remaining skill words in the training skill word table except x i , and P(x i |contex(xi )) represents the probability value of the occurrence of the i-th skill word in the output.

[0069] Step 24: Use the preset stochastic gradient algorithm to iteratively update the model parameters of the word embedding model to obtain a skill word evaluation model.

[0070] In step 24, during the model training process, the stochastic gradient descent algorithm is used to continuously update the model parameters W T and W' T , until the model converges, and finally the required skill word evaluation model is obtained.

[0071] The skill word evaluation method provided by the embodiments of the present disclosure automatically extracts skill information from resumes, and uses the context information of skill words and a pre-trained skill word evaluation model to predict the probability of the occurrence of skill words. The larger the probability value, the higher the importance of the skill word. Thus, the evaluation of each skill word in the resume is realized automatically, the accuracy of skill word evaluation is improved. At the same time, a skill profile of a resume can be quickly constructed, which can effectively help recruiters quickly extract skill information from resumes, understand the resume content, and complete resume screening, improve the resume screening efficiency, and greatly save the time cost of manual screening and evaluation.

[0072] Figure 6 is a block diagram of a skill word evaluation device provided by an embodiment of the present disclosure. As Figure 6 shown, the skill word evaluation device is used to implement the above-mentioned skill word evaluation method. The skill word evaluation device includes: a skill word acquisition module 31 and a skill word evaluation module 32.

[0073] Among them, the skill word acquisition module 31 is used to determine a first skill word list to be evaluated from the resume document to be evaluated, and the first skill word list includes multiple skill words.

[0074] The skill word evaluation module 32 is used to, for each skill word in the first skill word list, use the pre-trained skill word evaluation model and the context information of the skill word in the first skill word list to predict the probability value of the occurrence of the skill word, and the probability value is used to characterize the importance degree of the skill word.

[0075] Figure 7 is Figure 6 a block diagram of a skill word acquisition module in Figure 7 shown. In some embodiments, the skill word acquisition module 31 includes a skill word extraction sub-module 311, a skill field determination sub-module 312, and a skill word list generation sub-module 313.

[0076] Among them, the skill word extraction sub-module 311 is used to determine a second skill word list from the resume document, and the second skill word list includes all the skill words appearing in the resume document; the skill field determination sub-module 312 is used to determine the technical field to which each skill word in the second skill word list belongs; the skill word list generation sub-module 313 is used to generate a first skill word list according to all the skill words and the corresponding technical fields in the second skill word list, and each technical field serves as a skill word.

[0077] In some embodiments, the skill word extraction sub-module 311 is specifically configured to: obtain resume text data from the resume document; extract all the skill words appearing in the resume text data from the resume text data to generate a second skill word list.

[0078] In some embodiments, the skill word extraction sub-module 311 is specifically configured to: perform word segmentation processing on the resume text data by using a preset word segmentation tool; and screen out all the skill words appearing in the resume text data from the word segmentation processing result by using a preset domain skill word library.

[0079] In some embodiments, the skill field determination sub-module 312 is specifically configured to use a preset knowledge graph to determine the technical field to which each skill word in the second skill word list belongs.

[0080] Figure 8 The block diagram of another skill word evaluation device provided by the embodiments of the present disclosure is shown in Figure 8 As shown, the skill word evaluation device further includes a model training module 33.

[0081] Among them, the model training module 33 is used to: obtain a training data set, and the training data set includes a plurality of training skill words extracted from resume samples; generate word vectors corresponding to each training skill word; for each training skill word, use the word vectors corresponding to each training skill word except this training skill word as inputs, and perform model training by using a preset word embedding model, and the output of the word embedding model is the probability value of the appearance of this training skill word; and use a preset stochastic gradient algorithm to iteratively update the model parameters of the word embedding model to obtain a skill word evaluation model.

[0082] In some embodiments, the word embedding model includes a continuous bag of words neural network model.

[0083] In addition, the skill word evaluation device provided by the embodiments of the present disclosure is specifically used to implement the foregoing skill word evaluation method. For specific details, reference may be made to the description of the foregoing skill word evaluation method, which will not be elaborated herein.

[0084] Figure 9 The block diagram of an electronic device provided by the embodiments of the present disclosure is shown in Figure 9As shown, the electronic device includes: one or more processors 501; a memory 502 storing one or more programs, which, when executed by the one or more processors 501, cause the one or more processors 501 to implement the above-mentioned skill word evaluation method; and one or more I / O interfaces 503 connected between the processor 501 and the memory 502 and configured to implement information interaction between the processor 501 and the memory 502.

[0085] Embodiments of the present disclosure also provide a computer-readable storage medium storing a computer program, wherein the computer program, when executed, implements the foregoing skill word evaluation method.

[0086] According to an embodiment of the present disclosure, the present disclosure also provides a computer program product including a computer program that, when executed by a processor, implements any one of the above-mentioned skill word evaluation methods.

[0087] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and appropriate combinations thereof. In the hardware implementation, the division of the functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be executed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transitory medium). As is well known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data. Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, as is well known to those of ordinary skill in the art, a communication medium generally includes computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and may include any information delivery medium.

[0088] Example embodiments have been disclosed herein, and although specific terms are employed, they are used only and should be interpreted only as general illustrative meanings and not for the purpose of limitation. In some instances, it will be apparent to those skilled in the art that, unless otherwise expressly specified, features, characteristics, and / or elements described in connection with a particular embodiment may be used singly or in combination with features, characteristics, and / or elements described in connection with other embodiments. Accordingly, those skilled in the art will understand that various forms and details may be changed without departing from the scope of the disclosure as set forth by the appended claims.

Claims

1. A method for evaluating skill words in a resume, including: Determine a first list of skill words to be evaluated from the resume document to be evaluated, where the first list of skill words includes multiple skill words; For each skill word in the first list of skill words, use a pre-trained skill word evaluation model and the context information of the skill word in the first list of skill words to predict the probability value of the occurrence of the skill word, and this probability value is used to characterize the importance of the skill word; Among them, the context information of the skill word in the first list of skill words includes other skill words in the first list of skill words except this skill word. The input of the skill word evaluation model is the word vectors corresponding to other skill words except this skill word, and the output is the probability value of the occurrence of this skill word. The probability value of the occurrence of this skill word refers to the probability of the occurrence of this skill word when other skill words in the resume document are known; The skill word evaluation model is trained in the following way: For each training skill word used for model training, use the word vectors corresponding to each training skill word except this training skill word as the model input, and use a preset word embedding model for model training to obtain the skill word evaluation model.

2. The skill word evaluation method according to claim 1, wherein determining the first list of skill words to be evaluated from the resume document to be evaluated, including: Determine a second list of skill words from the resume document, and the second list of skill words includes all skill words appearing in the resume document; Determine the technical field to which each skill word in the second list of skill words belongs; Generate the first list of skill words according to all skill words in the second list of skill words and the corresponding technical fields, and each technical field is used as a skill word.

3. The skill word evaluation method according to claim 2, wherein determining the second list of skill words from the resume document, including: Obtain resume text data from the resume document; Extract all skill words appearing in the resume text data from the resume text data to generate the second list of skill words.

4. The skill word evaluation method according to claim 3, wherein extracting all skill words appearing in the resume text data from the resume text data, including: Use a preset word segmentation tool to perform word segmentation on the resume text data; Use a preset domain skill word library to screen out all skill words appearing in the resume text data from the word segmentation result.

5. The skill word evaluation method according to claim 2, wherein determining the technical field to which each skill word in the second list of skill words belongs, including: Use a preset knowledge graph to determine the technical field to which each skill word in the second list of skill words belongs.

6. The skill word evaluation method according to claim 1, wherein the skill word evaluation model is trained through the following steps: Obtain a training data set, and the training data set includes multiple training skill words extracted from resume samples; Generate word vectors corresponding to each training skill word; For each training skill word, use the word vectors corresponding to each training skill word except this training skill word as inputs, and use a preset word embedding model to train the model. The output of the word embedding model is the probability value of the occurrence of this training skill word; And use a preset stochastic gradient algorithm to iteratively update the model parameters of the word embedding model to obtain the skill word evaluation model.

7. The skill word evaluation method according to claim 6, wherein generating the word vectors corresponding to each training skill word includes: Perform one-hot encoding processing on each training skill word to obtain the word vector corresponding to each training skill word.

8. The skill word evaluation method according to claim 6, wherein the word embedding model includes a continuous bag-of-words neural network model.

9. The skill word evaluation method according to claim 1, wherein using the pre-trained skill word evaluation model and the context information of this skill word in the first skill word list to predict the probability value of the occurrence of this skill word includes: Generate the word vectors corresponding to each skill word in the first skill word list except this skill word; Use the word vectors corresponding to each skill word except this skill word as the input of the skill word evaluation model, and use the skill word evaluation model to predict the probability value of the occurrence of this skill word.

10. A skill word evaluation device comprises: A skill word acquisition module, configured to determine a first skill word list to be evaluated from a resume document to be evaluated, where the first skill word list includes multiple skill words; A skill word evaluation module, configured to, for each skill word in the first skill word list, use the pre-trained skill word evaluation model and the context information of this skill word in the first skill word list to predict the probability value of the occurrence of this skill word, and this probability value is used to characterize the importance of this skill word; Wherein, the context information of this skill word in the first skill word list includes other skill words in the first skill word list except this skill word. The input of the skill word evaluation model is the word vector corresponding to other skill words except this skill word, and the output is the probability value of the occurrence of this skill word. The probability value of the occurrence of this skill word refers to the probability of the occurrence of this skill word when other skill words in the resume document are known; The skill word evaluation model is obtained through the following method: For each training skill word used for model training, use the word vectors corresponding to each training skill word except this training skill word as model inputs, and use a preset word embedding model to train the model to obtain the skill word evaluation model.

11. The skill word evaluation device according to claim 10, wherein the skill word acquisition module includes a skill word extraction sub-module, a skill field determination sub-module, and a skill word list generation sub-module; The skill word extraction sub-module is configured to determine a second skill word list from the resume document, and the second skill word list includes all skill words appearing in the resume document; The skill field determination sub-module is configured to determine the technical field to which each skill word in the second skill word list belongs; The skill word list generation sub-module is used to generate the first skill word list according to all the skill words and corresponding technical fields in the second skill word list, with each technical field serving as a skill word.

12. The skill word evaluation device according to claim 11, wherein the skill word extraction sub-module is specifically configured to: obtain resume text data from the resume document; extract all the skill words that appear in the resume text data from the resume text data to generate the second skill word list.

13. The skill word evaluation device according to claim 12, wherein the skill word extraction sub-module is specifically configured to: perform word segmentation processing on the resume text data by using a preset word segmentation tool; and screen out all the skill words that appear in the resume text data from the word segmentation processing results by using a preset domain skill word library.

14. The skill word evaluation device according to claim 11, wherein the skill field determination sub-module is specifically configured to determine the technical field to which each skill word in the second skill word list belongs by using a preset knowledge graph.

15. The skill word evaluation device according to claim 10, further comprising a model training module; The model training module is used to: obtain a training data set, where the training data set includes a plurality of training skill words extracted from resume samples ； generate word vectors corresponding to the respective training skill words; For each training skill word, use the word vectors corresponding to each training skill word other than this training skill word as inputs, and perform model training by using a preset word embedding model, where the output of the word embedding model is the probability value of the occurrence of this training skill word; and use a preset stochastic gradient algorithm to iteratively update the model parameters of the word embedding model to obtain the skill word evaluation model.

16. The skill word evaluation device according to claim 15, wherein the word embedding model includes a continuous bag-of-words neural network model.

17. An electronic device, which comprises: one or more processors; a memory having stored thereon one or more programs, which when executed by the one or more processors, cause the one or more processors to implement the skill word evaluation method according to any one of claims 1-9.

18. A computer-readable medium having stored thereon a computer program, wherein, the computer program, when executed, implements the skill word evaluation method according to any one of claims 1-9.

19. A computer program product comprising a computer program, which when executed by a processor, implements the skill word evaluation method according to any one of claims 1-9.

Images