TWI776146B - Resume scoring method and system - Google Patents

Abstract

A resume scoring system includes a plurality of regular expressions, a vector generating model, and a scoring model, each regular expression has a predetermined keyword, the vector generating model is used to generate a text vector according to a text content, the scoring The model generates a score based on the results generated by the regular expressions and the text vector generated by the vector generating model. When receiving a resume related to a job applicant, for each regular expression, the system uses the The regular expression obtains a keyword combination including the keyword content in the resume corresponding to the predetermined keyword of the regular expression, and uses the vector generation model to generate a text vector corresponding to the resume, and according to each keyword combination and the text vector, using the scoring model to generate a score.

Description

Resume scoring method and system

The present invention relates to an office automation method, in particular to a method for automatically generating scores according to resumes.

In today's society, when general companies conduct talent recruitment operations, the first pass will be based on the resumes provided by job seekers. The human resources department reviews the contents of each resume one by one to screen suitable job applicants. This approach is not only laborious, but also may cause the problem of wrong resume selection due to different opinions or human negligence.

In view of this, it is necessary to propose a new solution to solve the problems of excessively time-consuming and labor-intensive reviewing of resumes and prone to wrong selection of resumes.

Therefore, the purpose of the present invention is to provide a resume scoring method for assisting in evaluating resumes

In addition, another object of the present invention is to provide a history scoring system that assists in evaluating history.

Therefore, the resume scoring method of the present invention is implemented by a server. The resume includes a school experience filled in by a job seeker and an autobiography introduced by a person. The server includes a plurality of regular expressions, a vector generation model, and A scoring model, each regular expression has a predetermined keyword, the vector generating model is used to generate a text vector according to a content composed of text, the scoring model is based on the result generated by the regular expression and the vector generating model The generated text vector generates a score. The resume scoring method includes a step (A), a step (B), and a step (C).

In the step (A), when the server receives the resume related to the job seeker, for each regular expression, the server uses the regular expression to obtain the content of the learning experience according to the content of the learning experience. The keyword content corresponding to the predetermined keyword of the regular expression, wherein the predetermined keyword of each regular expression and its corresponding keyword content constitute a corresponding keyword combination.

In step (B), the server uses the vector generation model to generate a pair of text vectors corresponding to the autobiographical content according to the autobiographical content.

In step (C), the server uses the scoring model to generate a pair of scores corresponding to the resume according to each keyword combination and the text vector.

Furthermore, the resume scoring system of the present invention is used to generate a score for a resume, which is connected to a management terminal through a communication network. The system includes a communication module, a storage module, and a processing module.

The communication module is connected to the communication network, and the storage module stores a plurality of regular expressions, a vector generation model, and a scoring model, each regular expression has a predetermined keyword, and the vector generation model is used for A text vector is generated from a content consisting of text, and the scoring model generates a score based on the result generated by the regular expression and the text vector generated by the vector generation model.

The processing module is electrically connected to the communication module and the storage module, wherein, when the processing module receives the resume related to the job seeker from the management terminal through the communication module, for each formal representation formula, the processing module uses the regular expression to obtain the keyword content corresponding to the predetermined keyword of the regular expression in the learning experience content according to the learning experience content, wherein the predetermined keyword of each regular expression and its The corresponding keyword content constitutes a corresponding keyword combination, and according to the autobiographical content, the vector generation model is used to generate a pair of text vectors corresponding to the autobiographical content, and the scoring model is used according to each keyword combination and the text vector. Generate a pair of scores that should be resumed.

The effect of the present invention is: by generating the score corresponding to the resume by the server, the relevant staff can refer to the score to screen job applicants, which not only saves the time and labor cost of reviewing each resume, but also avoids the problem of wrong resume selection caused by human factors. .

Before the present invention is described in detail, it should be noted that in the following description, similar elements are designated by the same reference numerals.

Referring to FIG. 1 , a first embodiment of the history scoring system of the present invention is implemented by a server 1. The server 1 is connected to a management terminal 2 through a communication network 100, and includes a communication module 11, a storage module 12, and a processing module 13.

The communication module 11 is connected to the management terminal 2 through the communication network 100 .

The storage module 12 stores multiple regular expressions, multiple historical records, a vector generation model, and a scoring model, each regular expression has a predetermined keyword, and each historical record has a The autobiographical content of the personal introduction of the historical job seeker, multiple keyword combinations related to the historical job seeker's academic experience, a text vector related to the autobiographical content, and a score related to the historical resume, the vector generation model is used to generate A text vector is generated from a content consisting of text, and the scoring model generates a score based on the result generated by the regular expression and the text vector generated by the vector generation model.

The processing module 13 is electrically connected to the communication module 11 and the storage module 12, and generates a pair of scores corresponding to a resume according to a resume related to a job seeker, wherein the resume includes a school experience content filled in by the job seeker , and an autobiography about the job seeker's personal introduction.

Referring to FIG. 2 , FIG. 3 , and FIG. 4 , the history scoring method of the present invention includes a vector generation model training program, a scoring model training program, and a scoring program.

Referring to FIG. 2, the vector generation model training program includes a step 301, a step 302, a step 303, a step 304, and a step 305, and describes how the processing module 13 adjusts and refines the vector generation according to the historical records Model.

In step 301, the processing module 13 uses a deep learning algorithm to establish a second training model for generating a text vector according to a content composed of text according to the autobiographical content and the text vector of the historical records , such as training models like BERT or XLNet.

In step 302, the processing module 13 uses the second training model to generate a plurality of training text vectors corresponding to each historical record according to the autobiographical content corresponding to each historical record.

In step 303, for each historical record, the processing module 13 determines whether the similarity between the text vector corresponding to the historical record and the training text vector is greater than a preset threshold. In the first embodiment, the processing module 13 obtains the similarity between the text vector of the historical record and the training text vector according to a similarity comparison algorithm, such as cosine similarity. When the processing module 13 determines that the similarity is not greater than the preset threshold, the processing module 13 adjusts the second training model and performs the step 302 again, that is, the step 304, when the processing module 13 determines When the similarity is greater than the preset threshold, the processing module 13 uses the second training model as the vector generation model for generating a text vector according to a content composed of text, that is, step 305 .

3, the scoring model training procedure includes a step 401, a step 402, a step 403, a step 404, and a step 405, and describes how the processing module 13 adjusts and refines the scoring model according to the historical records.

In step 401, the processing module 13 divides the stored multiple historical records into a training subset and a testing subset, wherein the historical records included in the training subset and the historical records included in the testing subset are The number of such historical records may be equal or different.

In step 402, the processing module 13 uses a machine learning algorithm, such as logistic regression, according to the keyword combination, the text vector, and the score corresponding to each historical record in the training subset ), Random Forest, Gradient Boosting, Artificial Neural Network, etc., to establish a first training model that generates a training score according to the keyword combinations and the text vector .

In step 403, the processing module 13 determines whether the first training model is overfit according to the keyword combination, the text vector, and the score corresponding to each historical record in the training subset and the test subset If the first training model is over-fit or under-fit, the processing module 13 adjusts the first training model and re-determines whether the adjusted first training model is over-fit or under-fit. Insufficient fit, that is, in step 404, when it is judged that the first training model is not overfitting or underfitting, the processing module 13 uses the first training model as the result generated according to the regular expression and the vector The text vector generated by the generative model generates a score for the scoring model, that is, step 405 .

Referring to FIG. 4 , the scoring procedure includes a step 501 , a step 502 , a step 503 , and a step 504 , and describes how the processing module 13 generates the score corresponding to the resume according to the resume.

In step 501, when the processing module 13 receives the resume from the management terminal 2 and is related to the job applicant through the communication module 11, for each regular expression, the processing module 13 according to The learning experience content of the resume uses the regular expression to obtain the keyword content of the learning experience content corresponding to the predetermined keyword of the regular expression, wherein the predetermined keyword of each regular expression and its corresponding keyword content Constitute a corresponding keyword combination. In the first embodiment, the predetermined keywords of the regular expressions are “highest education”, “work experience”, and “English ability” respectively, then the processing module 13 obtains the keywords according to the regular expressions The content of the words corresponds to "Master of Electrical Engineering of National Cheng Kung University", "3 years of TSMC R&D Department", "TOEIC 870 points", and "Highest Education: Master of Electrical Engineering of National Cheng Kung University" is a keyword corresponding to "Highest Education" Combinations, similarly, the keyword combinations corresponding to "work experience" and "English ability" are "work experience: 3 years in TSMC's R&D department" and "English ability: TOEIC 870 points".

In step 502, the processing module 13 uses the vector generation model to generate a pair of text vectors corresponding to the autobiographical content according to the autobiographical content of the resume. It is worth mentioning that the vector generation model may be the vector generation model confirmed in step 305 .

In step 503, the processing module 13 uses the scoring model to generate a pair of scores corresponding to the resume according to each keyword combination and the text vector. It is worth mentioning that the scoring model may be the scoring model confirmed in step 405 .

In step 504, the processing module 13 stores the autobiographical content of the resume, each keyword combination, the text vector, and the score as one of the historical resumes. In this way, the number of these historical records can be accumulated, so that the training samples of the scoring model and the vector generation model are more diverse, and the scoring model and the vector obtained by continuously tracking and adjusting the training and testing samples are refined. Generate the model.

Referring to FIG. 5 , a second embodiment of the resume scoring method of the present invention is similar to the first embodiment, and the similarities will not be repeated. The difference is that in the second embodiment, the storage module 12 also stores an abstract generation model for generating an abstract according to a content composed of text, and each historical record stored in the storage module 12 also has an autobiographical content related to the corresponding historical record. Abstract, and before the step 504, it also includes an abstract program, which describes how the processing module 13 generates a pair of abstracts of the autobiographical content of the resume according to the autobiographical content of the resume, and includes a step 601, a step 602, a Step 603 , a step 604 , a step 605 , and a step 606 .

In step 601, the processing module 13 uses a deep learning algorithm, such as a recursive neural network, to create a content generated based on a text based on the autobiographical contents and the abstracts of the historical records. A summary of the training model, such as models such as GPT-2 or Transformer.

In step 602, the processing module 13 uses the training model to generate a plurality of training abstracts corresponding to each historical record according to the autobiographical content corresponding to each historical record.

In step 603, for each historical record, the processing module 13 determines whether the similarity between the abstract corresponding to the historical record and the training abstract is greater than another preset threshold. In this embodiment, the processing module 13 uses a similarity comparison algorithm, such as cosine similarity, to obtain the similarity between the abstract of the historical record and the training abstract. When judging that the similarity is not greater than the other preset threshold, the processing module 13 adjusts the training model and repeats the step 602, that is, the step 604, when it is judged that the similarity is greater than the other preset threshold , the processing module 13 determines that the training model is an abstract generation model for generating an abstract according to a content composed of text, that is, step 605 .

In step 606, when the processing module 13 receives the resume related to the job applicant from the management terminal 2 through the communication module 11, the processing module 13 uses the autobiographical content of the resume to use The abstract generation model produces a pair of abstracts that should be autobiographical content.

In addition, in the second embodiment, in the step 504, the processing module 13 stores the autobiographical content of the resume, each keyword combination, the text vector, the score, and the abstract as the historical resumes one of them.

To sum up, in the resume scoring method of the present invention, the processing module 13 generates a pair of scores corresponding to the resume according to the regular expressions, the vector generation model, and the scoring model, thereby saving relevant staff. The time and labor cost of evaluating and screening each resume, and at the same time, because the processing module generates the score according to the scoring model, problems caused by different opinions or human errors are avoided, such as resumes. Mis-screening or mis-scoring, in addition, the processing module 13 generates a pair of summaries corresponding to the autobiographical content according to the summary generation model. The abstract can quickly learn more about the job seeker, thereby saving the time and labor cost of reading the entire resume, so the purpose of the present invention can be achieved indeed.

However, the above are only examples of the present invention, and should not limit the scope of implementation of the present invention. Any simple equivalent changes and modifications made according to the scope of the patent application of the present invention and the contents of the patent specification are still included in the scope of the present invention. within the scope of the invention patent.

1: Servo side 100: Communication Network 11: Communication module 12: Storage Module 13: Processing modules 2: Management terminal 301~305: Steps 401~405: Steps 501~504: Steps 601~606: Steps

Other features and effects of the present invention will be clearly presented in the embodiments with reference to the drawings, wherein: 1 is a block diagram illustrating a first embodiment of the resume scoring system of the present invention connected to a management terminal via a communication network; Fig. 2 is a flow chart illustrating a vector generation model training program of the resume scoring method of the present invention executed by the first embodiment; FIG. 3 is a flow chart illustrating a scoring model training procedure of one of the resume scoring methods of the present invention executed by the first embodiment; FIG. 4 is a flow chart illustrating a scoring procedure of the history scoring method of the present invention performed by the first embodiment; and FIG. 5 is a flow chart illustrating a summary procedure of the history scoring method of the present invention performed by a second embodiment.

1: Servo side

100: Communication Network

11: Communication module

12: Storage Module

13: Processing modules

2: Management terminal

Claims (8)

A resume scoring method is implemented by a server. The resume includes a school experience filled in by a job seeker and an autobiography introduced by a person. The server includes a plurality of regular expressions, a vector generation model, and a scoring model. , each regular expression has a predetermined keyword, the vector generating model is used to generate a text vector according to a content composed of words, the scoring model is based on the result generated by the regular expression and the text generated by the vector generating model The vector generates a score, and the resume scoring method includes the following steps: (A) when the server receives the resume related to the job seeker, for each regular expression, the server uses the regular expression according to the content of the learning experience. The expression obtains the keyword content corresponding to the predetermined keyword of the regular expression in the learning experience content, wherein the predetermined keyword of each regular expression and its corresponding keyword content constitute a corresponding keyword combination; (B ) According to the autobiographical content, the server utilizes the vector generation model to generate a pair of text vectors corresponding to the autobiographical content; (C) the server utilizes the scoring model to generate a pair of corresponding resumes according to each keyword combination and the text vector (D) The server stores the autobiographical content of the resume, each keyword combination, the text vector, and the score into a historical record; (E) The server divides the stored historical records into a training subset and a test subset; (F) the server uses a machine learning method according to the keyword combination, the text vector, and the score corresponding to each historical record in the training subset The learning algorithm establishes a first training model that generates a training score according to the keyword combinations and the text vector; (G) the server according to the training subset and the test subset corresponding to each historical record in the key The word combination, the text vector, and the score determine whether the first training model is over-fitting or under-fitting; (H) when the server determines that the first training model is over-fitting or under-fitting, the servo The terminal adjusts the first training model and performs the step (G) again; and (I) when the server determines that the first training model is not overfitting or underfitting, the server determines that the first training model is not overfitted or underfitted as the scoring model when performing this step (C). The resume scoring method according to claim 1, further comprising the following steps: (J) the server uses a deep learning algorithm to establish a The content produces a second training model of a text vector; (K) the server uses the second training model to generate a plurality of training text vectors corresponding to each history respectively according to the autobiographical content corresponding to each historical record; (L ) For each historical record, the server determines whether the similarity between the text vector and the training text vector corresponding to the historical record is greater than a preset threshold; (M) when the server determines that the similarity is not greater than the preset When the threshold is set, the server adjusts the second training model and performs the step (K) again; and (N) when the server determines that the similarity is greater than the preset threshold, the server The server uses the second training model as the vector generation model when performing the step (B). The resume scoring method according to claim 1, wherein the server further includes an abstract generation model, the abstract generation model is used to generate an abstract according to a content composed of text, and after the step (C), the following further includes: Step: (O) According to the autobiographical content, the server generates a pair of abstracts corresponding to the autobiographical content by using the abstract generation model. According to the resume scoring method described in claim 3, the server further includes a plurality of historical resumes, and each historical resume includes an autobiographical content and an abstract of the corresponding historical resume, wherein, before the step (0), the It includes the following steps: (P) the server uses a deep learning algorithm to establish a training model for generating an abstract according to a content composed of text according to the autobiographical content and the abstracts of the historical records; (Q) The server uses the training model to generate a plurality of training abstracts corresponding to each historical record according to the autobiographical content corresponding to each historical record; (R) for each historical record, the server determines the Whether the similarity between the digest and the training digest is greater than another preset threshold; (S) when the server determines that the similarity is not greater than the other preset threshold, the server adjusts the training model and performs the process again Step (Q); and (T) when the server determines that the similarity is greater than the other preset threshold, the server determines that the training model is the digest generation model. A resume scoring system for scoring a resume and passing a pass The information network is connected to a management terminal. The resume includes a school experience filled in by a job seeker and an autobiography introduced by a person. The resume scoring system includes: a communication module, connected to the communication network; a storage module, A plurality of regular expressions, a vector generation model, and a scoring model are stored, each regular expression has a predetermined keyword, the vector generation model is used to generate a text vector according to a content composed of text, and the scoring The model generates a score according to the result generated by the regular expression and the text vector generated by the vector generation model; and a processing module electrically connected to the communication module and the storage module; wherein, when the processing module passes the communication When the module receives the resume related to the job seeker from the management terminal, for each regular expression, the processing module uses the regular expression to obtain the content of the learning experience corresponding to the regular expression according to the content of the learning experience. The keyword content of the predetermined keyword of the expression, wherein the predetermined keyword of each regular expression and its corresponding keyword content constitute a corresponding keyword combination, and according to the autobiographical content, the vector generation model is used to generate a pair of The text vector of the autobiographical content, and according to each keyword combination and the text vector, the scoring model is used to generate a pair of scores corresponding to the resume, the processing module The autobiographical content of the resume, each keyword combination, The text vector and the score are stored in the storage module to form a historical record, and the processing module divides the stored multiple historical records into a training subset and a test subset by the storage module, and according to the The keyword combination, the text vector, and the score corresponding to each historical record in the training subset are created using a machine learning algorithm to generate a training score according to the keyword combination and the text vector. number of the first training model, and according to the keyword combination, the text vector, and the score corresponding to each historical record in the training subset and the test subset, determine whether the first training model is overfitting or overfitting. If it is judged that the first training model is over-fitting or under-fitting, adjust the first training model and re-judge whether it is over-fitting or under-fitting, and when it is judged that the first training model is not over-fitting When the fit is insufficient, the first training model is used as the scoring model. The resume scoring system as claimed in claim 5, wherein the processing module uses a deep learning algorithm to create a system based on the autobiographical content and text vectors of a plurality of historical resumes stored in the storage module. The constituted content generates a second training model of a text vector, and according to the autobiographical content corresponding to each historical resume, the second training model is used to generate a plurality of training text vectors corresponding to each historical resume, for each historical resume , the processing module judges whether the similarity between the text vector corresponding to the historical resume and the training text vector is greater than a preset threshold, and when it is judged that the similarity is not greater than the preset threshold, adjusts the second training model and regenerate the training text vectors and determine the similarity. When the similarity is determined to be greater than the preset threshold, the second training model is used as the vector generation model. The resume scoring system according to claim 5, wherein the storage module further stores an abstract generation model, the abstract generation model is used to generate an abstract according to a content composed of text, and the processing module is based on the autobiographical content, The abstract generation model is used to generate a pair of abstracts of the autobiographical content. The resume scoring system according to claim 7, wherein the storage module further stores a plurality of historical resumes, and each historical resume includes a corresponding historical resume An autobiographical content and an abstract of the history, the processing module uses a deep learning algorithm to establish a training model that generates an abstract according to a content composed of text according to the autobiographical content and the abstract, and according to each history For the autobiographical content corresponding to the resume, the training model is used to generate a plurality of training abstracts corresponding to each historical resume. For each historical resume, the processing module determines whether the similarity between the abstract corresponding to the historical resume and the training abstract is not greater than another preset threshold, when it is judged that the similarity is not greater than the other preset threshold, adjust the training model and regenerate the training summaries and judge the similarity, when it is judged that the similarity is greater than the other preset When the threshold is set, the trained model is used as the summary generation model.

Images