KR20210099745A - System and Method for evaluation of personal statement using natural language processing technology - Google Patents

Abstract

Provided are a system for evaluating a letter of self-introduction through classification of input sentences/documents, measurement of similarity, detection of specific expressions (words, phrases, etc.), and calculation of evaluation scores using natural language processing technology and the calculation of evaluation scores using the classification of input sentences/documents, measurement of similarity, detection of specific expressions (words, phrases, etc.), and calculation of evaluation scores, and a method thereof. According to one embodiment of the present invention, the system for evaluating a letter of self-introduction comprises: a first processor executing a self-introduction evaluation engine which analyzes a job seeker's letter of self-introduction to generate evaluation result information including evaluation scores and sentences used as a basis of evaluation score calculation; and a second processor providing an operation screen of the system and executing an application program interworking with the self-introduction evaluation engine. Accordingly, since the evaluation of a letter of self-introduction is automated by using natural language processing technology and an evaluation result is scored, the present invention can reduce time and cost required to review a letter of self-introduction and help to conduct consistent and fair evaluation without being influenced by the inclination or mood of a personnel manager.

Description

System and Method for evaluation of personal statement using natural language processing technology

The present invention relates to a self-introduction evaluation system and method, and more particularly, classification of input sentences/documents using natural language processing technology, measurement of similarity, detection of specific expressions (words, phrases, etc.) and using the same It relates to a system and method for evaluating a self-introduction by calculating an evaluation score.

In general, a self-introduction is a document written to provide detailed information to others by listing one's growth environment or strengths for the purpose of introducing oneself.

Such a self-introduction letter is written and submitted when joining a specific group, getting a job, or when selecting new students. Since it cannot be identified, it can be used to select suitable talents by evaluating basic characteristics or personality through the self-introduction letter.

Conventionally, the hiring process was done in such a way that the human resources manager of a company for the purpose of recruiting reviews a number of self-introductions and selects the right person for his or her company, but this is inconvenient in that it takes a lot of time and money.

In addition, there is a problem in that the evaluation score varies due to environmental factors at the time of evaluation even if the same HR manager evaluates the same self-introduction letter due to factors such as subjectivity of application of evaluation criteria when evaluating the self-introduction letter of the personnel manager.

Therefore, it is required to find a way to reduce the time and cost required to review the self-introduction, and to perform a consistent and fair evaluation without being influenced by the inclination or mood of the personnel manager.

The present invention has been devised to solve the above problems, and an object of the present invention is to provide a self-introduction evaluation system and method capable of automating self-introduction evaluation using natural language processing technology and scoring the results. .

In accordance with an embodiment of the present invention for achieving the above object, the self-introduction evaluation system is a self-introduction evaluation engine that analyzes the job seeker's self-introduction and generates evaluation result information including evaluation scores and evaluation score calculation basis sentences. a first processor executing; and a second processor that provides an operation screen of the system and runs an application program that interworks with the self-introduction evaluation engine.

And the self-introduction evaluation engine, a sentence unit evaluation module for performing evaluation by parsing the input self-introduction in sentence units; includes.

In addition, the sentence unit evaluation module may be trained using training sentences composed of sentences matching the self-introduction evaluation items.

In addition, the sentence unit evaluation module may classify the input sentence according to a preset evaluation item to generate class information, and perform evaluation using the generated class information.

In addition, the self-introduction evaluation engine, an additional point and deduction item derivation module for performing evaluation by deriving the additional points and deduction items included in the input self-introduction in sentence units; may further include.

In addition, the additional point and deduction item derivation module can derive content that is not related to the evaluation standard and third-party support as deduction items, and can derive content corresponding to experience or history that can prove professionalism as an additional point item.

In addition, the additional points and deductions item derivation module derives sentences with a high probability of contradicting facts or false facts among the contents of the self-introduction letter as a handwritten check item based on the resume and submitted documents of the same job seeker as the input self-introduction letter can do.

And the self-introduction evaluation engine, a document unit evaluation module for comparing the similarity between the input self-introduction and the evaluation completed past self-introduction; may further include.

In addition, the document unit evaluation module selects the past self-introduction that has the highest similarity to the input self-introduction letter, derives the evaluation score of the selected past self-introduction letter, and evaluates it using the similarity between the derived evaluation score and the self-introduction letter. can be done

And the self-introduction evaluation engine, the job suitability evaluation module for evaluating the input self-introduction by analyzing the relationship with the technical requirements of the recruitment notice or the job description of a job offer company; may further include.

In addition, the self-introduction evaluation system according to an embodiment of the present invention is implemented in a parallel structure with the self-introduction evaluation engine, and may further include a disposition analysis engine for analyzing the applicant's disposition based on the input self-introduction letter. .

And the self-introduction evaluation system according to an embodiment of the present invention calculates the final result by merging the evaluation result information and the propensity analysis information, which is the propensity analysis result, but is based on the technical requirements and other requirements of the recruitment group or the recruitment criteria of the recruitment agency. Accordingly, each weight may be assigned to the pre-merging evaluation result information and the propensity analysis result.

On the other hand, according to another embodiment of the present invention, the self-introduction evaluation method, the self-introduction evaluation system, analyzing the job seeker's self-introduction, generating evaluation result information including evaluation scores and evaluation score calculation basis sentences; and outputting, by the self-introduction evaluation system, the generated evaluation result information.

As described above, according to the embodiments of the present invention, the self-introduction evaluation is automated using natural language processing technology, and by scoring the results, the time and cost required for reviewing the self-introduction is reduced, and the tendency of the personnel manager It can help to conduct a consistent and impartial evaluation without being influenced by or mood.

1 to 2 are diagrams provided for the description of a self-introduction evaluation system using natural language processing technology according to an embodiment of the present invention;
3 to 4 are diagrams provided for the description of a method of evaluating a sentence unit using a self-introduction evaluation system using a natural language processing technology according to an embodiment of the present invention;
5 to 6 are diagrams provided for the description of a method for evaluating a document unit using a self-introduction evaluation system using a natural language processing technology according to an embodiment of the present invention;
7 is a view provided for explaining a method of summing a sentence unit evaluation score and a document unit evaluation score using a self-introduction evaluation system using a natural language processing technology according to an embodiment of the present invention, and
8 is a diagram illustrating a screen on which a self-introduction evaluation result is output.

Hereinafter, the present invention will be described in more detail with reference to the drawings.

1 to 2 are diagrams provided for explanation of a self-introduction evaluation system (hereinafter, collectively referred to as 'self-introduction evaluation system') using natural language processing technology according to an embodiment of the present invention.

The self-introduction evaluation system according to the present embodiment can automate the self-introduction evaluation by using natural language processing technology, and score the results.

To this end, the self-introduction evaluation system includes a communication unit 10 , a storage unit 20 , and a processor 30 .

The communication unit 10 may receive data necessary for the processor 30 to operate or transmit data generated by the processor 30 to the outside.

For example, the communication unit 10 may be provided with a communication module to receive self-introductions from the outside.

The processor 30 is provided to execute the application program 100 , the self-introduction evaluation engine 200 , and the tendency analysis engine 300 .

That is, the application program 100, the self-introduction evaluation engine 200 and the tendency analysis engine 300 are executed in one processor or each processor.

The application program 100 may provide an operation screen that allows the user to operate the self-introduction evaluation system.

Specifically, the application program 100, in conjunction with the self-introduction evaluation engine 200 and the propensity analysis engine 300, may output the original self-introduction letter, evaluation results, evaluation basis sentences, etc. on the screen.

Also, the application program 100 may include a manipulation unit 110 that is a user interface for controlling information output on the screen.

The self-introduction evaluation engine 200 is provided to perform automatic evaluation of the self-introduction letter.

Specifically, the self-introduction evaluation engine 200, when the job seeker's self-introduction is input, analyzes the input self-introduction using natural language processing technology, calculates an evaluation score based on the analysis result, and calculates the evaluation score. Evidence sentences (additional points or deduction items) can be derived.

That is, the self-introduction evaluation engine 200 may analyze the self-introduction letter and generate evaluation result information including the evaluation score and the evaluation score calculation basis sentence.

To this end, the self-introduction evaluation engine 200 includes a sentence unit evaluation module 210 , a document unit evaluation module 220 , a job suitability evaluation module 230 , an additional point and deduction item derivation module 240 , and a scoring module 250 . ) may be included.

The sentence unit evaluation module 210 may perform evaluation by parsing the input self-introduction letter by sentence unit. A more detailed description of the sentence unit evaluation module 210 will be described later with reference to FIGS. 3 to 4 .

The document unit evaluation module 220 compares the similarity between the input self-introduction and the evaluation completed past self-introductions, selects the past self-introductions with the highest similarity to the input self-introduction letter, and evaluates the selected past self-introductions can be derived, and evaluation can be performed using the degree of similarity between the derived evaluation score and the self-introduction letter. A more detailed description of the document unit evaluation module 220 will be described later with reference to FIGS. 5 to 6 .

The job suitability evaluation module 230 may analyze and evaluate the relevance of the input self-introduction letter with the technical requirements of a job advertisement or a job description of a job offer company.

The additional point and deduction item derivation module 240 may derive the additional point and deduction item for performing evaluation by deriving the additional point and deduction items included in the input self-introduction in sentence units.

Specifically, the additional point and deduction item derivation module 240 derives content that is not related to the evaluation standard and third-party support as deductible items, and derives content corresponding to experience or history that can prove professionalism as an additional point item can do.

For example, the additional point and deduction item derivation module 240 may detect the content corresponding to the deductible item, such as "I want to become a company B," among the contents of the self-introduction applied to the company A.

In addition, the additional point and deduction item derivation module 240, "I have a history of publishing a thesis in the field of A". or "I won the B contest." Experience and history that can prove professionalism, such as such, can be derived as additional points.

For another example, the additional point and deduction item derivation module 240 is based on the inputted self-introduction and the same job seeker's resume and submitted documents, sentences with a high possibility of contradicting facts or false facts among the contents of the self-introduction letter can be derived as a handwritten check item.

That is, the additional points and deductions item derivation module 240 described in the self-introduction that he graduated from a master's degree at University A, but the job seeker's resume and submitted documents indicate that he graduated from a bachelor's degree at University B. Based on the resume and the contents of the resume and submitted documents of the same job seeker as the introduction, sentences that are highly likely to contradict or falsely describe basic facts such as age, educational background, family relationship, etc. can be derived as a handwriting check item.

The scoring module 250 may add up the sentence unit evaluation score and the document unit evaluation score.

The disposition analysis engine 300 may analyze the disposition of the applicant based on the input self-introduction letter.

Specifically, the disposition analysis engine 300 is implemented in a parallel structure with the self-introduction evaluation engine 200, and based on the input self-introduction, the disposition of the applicant, such as extroversion, affinity, emotional ups and downs, openness and sincerity, is evaluated. It can be calculated as a score for each set item.

Here, the scoring module 250 may calculate a final result by merging the evaluation result information of the self-introduction letter and the disposition analysis information that is the disposition analysis result.

In this case, the pre-merger evaluation result information and propensity analysis result may be given weights according to the technical requirements and other requirements of the recruitment group or the recruitment criteria of a recruitment agency.

That is, when recruiting talents with openness, extroversion, and affinity rather than conscientiousness according to the job category, the weight of disposition analysis information is given higher to those who show higher openness, extroversion, and affinity than conscientiousness. can do.

The storage unit 20 may include a database 400 , and may store data necessary to drive the application program 100 , the self-introduction evaluation engine 200 , and the tendency analysis engine 300 .

Specifically, the database 400, the job seeker's self-introduction and evaluation results of the self-introduction letter may be identified and stored for each individual.

3 to 4 are diagrams provided for explaining a method of evaluating in units of sentences using the self-introduction evaluation system according to an embodiment of the present invention.

The sentence unit evaluation module 210 of this self-introduction evaluation system uses the per-sentence evaluation model to parse the self-introduction into sentence units, classifies the sentences according to the preset evaluation items, and provides class information. It can be created and evaluated using the generated class information.

At this time, the sentence unit evaluation model is trained using training sentences composed of sentences matching the self-introduction evaluation items.

Referring to FIG. 3, the sentence unit evaluation method using the sentence unit evaluation module 210 of the self-introduction evaluation system collects and refines the data to train the sentence unit evaluation model (S310), and generates training data. can be (S320).

At this time, when the training data is generated, the sentence unit evaluation module 210 may train the sentence unit evaluation model using the generated training sentence (S330), and, in addition, the sentence unit evaluation module 210 performs the test. and review the results based on this (S350).

Referring to FIG. 4, the generation, training, and testing process of the sentence unit evaluation model will be described in detail. The sentence unit evaluation module 210 generates a list of training sentences (S410), and sets the training sentence list as a training set and a test set. is separated (S420), the generation and training of the classifier is executed (S430), and the test of the classifier can be performed (S440).

At this time, when the classification result is generated (S450), the sentence unit evaluation module 210 determines whether the accuracy is greater than or equal to a preset target value (S460), and if the accuracy is greater than or equal to the preset target value (S460-Yes), ID of the classifier And credential information may be stored in the database (S470).

In addition, the sentence unit evaluation module 210, on the contrary, if the accuracy is less than the preset target value (S460-No), automatically extract the training sentence (S480), add the training sentence (S490), and re-perform the training. there is.

5 to 6 are diagrams provided for explaining a method of evaluating a document unit using a self-introduction evaluation system according to an embodiment of the present invention.

Referring to FIG. 5 , the document unit evaluation method using the document unit evaluation module 220 of the evaluation system may collect and refine data (S510) to generate and train a document similarity model (S520).

In addition, the document unit evaluation module 220, using the document similarity model, when selecting a past self-introduction (Sample B) with a high degree of similarity to a specific self-introduction (Sample A), the past selected using the document unit evaluation model An evaluation score of the self-introduction letter (Sample B) is derived, an evaluation test is performed using the similarity between the derived evaluation score and the self-introduction letter (S530), and the result can be reviewed based on this (S540).

At this time, the document unit evaluation model is trained using training data composed of self-introductions that have been evaluated in the past.

Referring to FIG. 6, the creation, training, and testing process of the document unit evaluation model will be described in detail. The document unit evaluation module 220 generates a list of self-introductions that have been evaluated in the past (S610), and a list of self-introductions. It can be divided into a training set and a test set (S620).

In addition, the document unit evaluation module 220 may perform a pre-processing operation such as POS tagging on each of the separated sets (S630). Here, POS tagging refers to an operation of tagging what kind of part-of-speech a morpheme is based on the meaning of the morpheme.

When the pre-processing operation is completed, the document unit evaluation module 220 may generate a plurality of candidate models of the document unit evaluation model ( S640 ), and individually perform a test using each candidate model ( S650 ).

In addition, the document unit evaluation module 220 calculates the accuracy of each candidate model, organizes it in the accuracy list (S660), and through this, the document unit evaluation model with the best accuracy can be selected (S670).

7 is a view provided for explaining a method of summing the evaluation score for each sentence and the evaluation score for each document using the self-introduction evaluation system according to an embodiment of the present invention.

Referring to FIG. 7 , the scoring module 250 may receive the sentence unit evaluation score and the document unit evaluation score from the sentence unit evaluation module 210 and the document unit evaluation module 220 , and may add them.

In this case, the sentence unit evaluation score and the document unit evaluation score before merging may be given weights W1 and W2, respectively, calculated through testing and optimization of evaluation models.

8 is a diagram illustrating a screen on which a self-introduction evaluation result is output.

The application program 100 is, as illustrated in FIG. 8, the evaluation score output area (a) of the self-introduction, the propensity analysis result output area (b), the core summary output area of the self-introduction (c) and the evaluation basis sentence (d) ) It is possible to output the self-introduction evaluation result output screen consisting of an output area, etc.

In particular, the core summary of the CV extracts the main sentences in the process of performing the evaluation of the CV, and can help the HR manager to manually evaluate the CV of the job seeker.

For example, in the application program 100, among the contents of the self-introduction letter, "I set a goal of winning a national invention contest since I entered university and worked hard.", "After hard work, I won the top prize at the 2019 national invention contest. " It is possible to extract the main sentences corresponding to the key summary, such as.

Through this, the time and cost required to review the personal statement can be reduced, and it can help to perform a consistent and fair evaluation without being influenced by the inclination or mood of the personnel manager.

On the other hand, it goes without saying that the technical idea of the present invention can also be applied to a computer-readable recording medium containing a computer program for performing the functions of the apparatus and method according to the present embodiment. In addition, the technical ideas according to various embodiments of the present invention may be implemented in the form of computer-readable codes recorded on a computer-readable recording medium. The computer-readable recording medium may be any data storage device readable by the computer and capable of storing data. For example, the computer-readable recording medium may be a ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical disk, hard disk drive, or the like. In addition, the computer-readable code or program stored in the computer-readable recording medium may be transmitted through a network connected between computers.

In addition, although preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above, and the technical field to which the present invention belongs without departing from the gist of the present invention as claimed in the claims Various modifications are possible by those of ordinary skill in the art, and these modifications should not be individually understood from the technical spirit or prospect of the present invention.

10: communication department
20: storage
30: processor
100 : application
110: control unit
200: self-introduction evaluation engine
210: sentence unit evaluation module
220: document unit evaluation module
230: job suitability evaluation module
240: Addition and deduction item derivation module
250: scoring module
300: propensity analysis engine
400 : database

Claims (13)

A first processor running a self-introduction evaluation engine that analyzes the job seeker's self-introduction letter and generates evaluation result information including evaluation scores and evaluation score calculation basis sentences; And
A self-introduction evaluation system comprising a; providing an operation screen of the system, and a second processor running an application program interworking with the self-introduction evaluation engine.
The method according to claim 1,
Self-introduction evaluation engine,
A self-introduction evaluation system comprising a; a sentence unit evaluation module for parsing the input self-introduction letter by sentence unit and performing evaluation.
3. The method according to claim 2,
The sentence unit evaluation module,
Self-introduction evaluation system, characterized in that the training (training) using a training sentence consisting of sentences matching the self-introduction evaluation items.
3. The method according to claim 2,
The sentence unit evaluation module,
A self-introduction evaluation system, characterized in that by classifying input sentences according to preset evaluation items, generating class information, and performing evaluation using the generated class information.
The method according to claim 1,
Self-introduction evaluation engine,
The self-introduction evaluation system, characterized in that it further comprises; an additional point and deduction item derivation module for performing evaluation by deriving the additional points and deduction items included in the input self-introduction letter by sentence unit.
6. The method of claim 5,
Addition and deduction item derivation module,
Contents that are not related to the evaluation criteria and contents supported by third parties are deducted as items,
Self-introduction evaluation system, characterized in that the content corresponding to the experience or history that can prove professionalism is derived as an additional point item.
6. The method of claim 5,
Addition and deduction item derivation module,
A self-introduction evaluation system, characterized in that, based on the inputted self-introduction and the resume and submitted documents of the same job seeker, sentences with a high possibility of contradicting facts or false facts among the contents of the self-introduction are drawn as handwritten check items.
The method according to claim 1,
Self-introduction evaluation engine,
The self-introduction evaluation system further comprising; a document unit evaluation module that compares the similarity between the input self-introduction and evaluation completed past self-introductions.
9. The method of claim 8,
The document unit evaluation module,
Self-introduction letter, characterized in that the evaluation is performed using the similarity between the derived evaluation score and the self-introduction letter by selecting the past self-introduction letter with the highest similarity to the input self-introduction letter, deriving the evaluation score of the selected past self-introduction letter evaluation system.
The method according to claim 1,
Self-introduction evaluation engine,
The self-introduction evaluation system, characterized in that it further comprises a; job suitability evaluation module to analyze and evaluate the input of the personal statement in relation to the technical requirements of the recruitment notice or the job description of the recruiting company.
The method according to claim 1,
The self-introduction evaluation system, characterized in that it further comprises a; is implemented in a parallel structure with the self-introduction evaluation engine, and analyzes the candidate's disposition based on the input self-introduction letter.
12. The method of claim 11,
The final result is calculated by merging the evaluation result information and the propensity analysis information, which is the result of the propensity analysis. Self-introduction evaluation system, characterized in that given.
generating, by the self-introduction evaluation system, the evaluation result information including the evaluation score and the evaluation score calculation basis sentence by analyzing the job seeker's self-introduction letter; and
A self-introduction evaluation method comprising; outputting, by the self-introduction evaluation system, the generated evaluation result information.

Images