Intelligent resume generation system based on natural language processing

Abstract

The invention belongs to the field of resume automatic generation, and discloses an intelligent resume generation system based on natural language processing, which comprises a position information acquisition module, a resume information storage module, a template storage module and a resume generation module, the position information acquisition module is used for acquiring recruitment information from a recruitment website; the resume information storage module is used for storing work experience information of job seekers; the template storage module is used for storing a resume template, the resume template comprises a basic information filling area and a working experience filling area, and basic information of a job seeker is filled in the basic information filling area in advance; and the resume generation module is used for performing screening from the work experience information by adopting a natural language processing technology according to the recruitment information to obtain screening information, and filling the screening information in a salary experience filling area in the resume template so as to generate a resume. According to the invention, intelligent generation of resumes is realized, and the efficiency of batch delivery of resumes by job seekers is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of automatic resume generation, and in particular to an intelligent resume generation system based on natural language processing. Background Art

[0002] In the traditional resume generation process, job seekers usually manually modify the content of the resume template according to the requirements of the job position to obtain a resume for submission. For the same type of position, different companies have different focuses, so job seekers need to manually modify their resumes to make them better suited to different recruitment requirements. For job seekers who want to get more interview opportunities by submitting a large number of resumes, modifying a large number of resumes takes a lot of time. Summary of the invention

[0003] The purpose of the present invention is to disclose an intelligent resume generation system based on natural language processing to solve the technical problem of how to assist job seekers in generating resumes more quickly.

[0004] In order to achieve the above object, the present invention adopts the following technical solution:

[0005] The present invention provides an intelligent resume generation system based on natural language processing, including a position information acquisition module, a resume information storage module, a template storage module and a resume generation module;

[0006] The job information acquisition module is used to obtain job information from job websites;

[0007] The resume information storage module is used to store the job seeker's work experience information;

[0008] The template storage module is used to store resume templates. The resume templates include a basic information filling area and a work experience filling area. The basic information of the job seeker has been filled in advance in the basic information filling area.

[0009] The resume generation module is used to filter work experience information based on recruitment information using natural language processing technology, obtain screening information, and fill the screening information into the salary and experience filling area in the resume template to generate a resume.

[0010] Preferably, the position information acquisition module includes a control unit and a crawler unit;

[0011] The control unit is used to adaptively calculate the access interval based on the crawling records;

[0012] The crawler unit is used to visit the recruitment website based on the access interval and crawl the recruitment information published by the recruitment website.

[0013] Preferably, the position information acquisition module further includes a comparison unit and a storage unit;

[0014] The comparison unit is used to compare all the recruitment information obtained by crawling this time with all the recruitment information obtained by crawling last time, and determine whether there is any newly added recruitment information. If so, the newly added recruitment information is stored and sent to the storage unit for storage.

[0015] Preferably, the crawling record includes the crawling time, the server response time, the received HTTP status code and the number of newly added recruitment information obtained in the crawling.

[0016] Preferably, the recruitment information includes basic information, job description, educational requirements, work experience requirements, skill requirements, and certificate requirements;

[0017] The recruitment website contains multiple pages, and each page contains multiple recruitment information.

[0018] Preferably, the basic information includes the job title, work location and company name of the position being recruited;

[0019] The job description includes the job content of the position;

[0020] Work experience requirements include project experience requirements and job experience requirements;

[0021] Skill requirements are the skills that the person recruited for this position needs to possess;

[0022] The skill requirements are the certificates that the person recruited for this position needs to have.

[0023] Preferably, the resume generation module includes a keyword acquisition unit and an extraction unit;

[0024] The keyword acquisition unit is used to acquire a keyword set from the recruitment information;

[0025] The extraction unit is used to obtain screening information from the job seeker's work experience information based on a keyword set.

[0026] Preferably, a keyword set is obtained from the recruitment information, including:

[0027] Perform word segmentation on the text of the recruitment information to obtain a word set;

[0028] Calculate the ranking coefficient of each word in the word set respectively;

[0029] Get keyword set words based on ranking coefficients.

[0030] Preferably, the text of the recruitment information is segmented to obtain a word set, including:

[0031] Remove useless characters from the text of the recruitment information to obtain the text to be segmented;

[0032] Use a word segmentation algorithm to perform word segmentation on the text to be segmented, obtain multiple words, and store the obtained words into a word set;

[0033] Remove stop words from a set of words.

[0034] Preferably, obtaining a keyword set based on a ranking coefficient includes:

[0035] The first S×N words with the largest ranking coefficients are stored as keywords in the keyword set, where N represents the total number of words in the keyword set and S represents a preset ratio.

[0036] Beneficial effects:

[0037] The present invention automatically obtains recruitment information from recruitment websites, and adopts natural language processing technology to automatically obtain screening information for filling in the work experience filling area in the resume template from the work experience information of job seekers based on the recruitment information, thereby realizing intelligent generation of resumes. The present invention can quickly generate a resume text when job seekers need to submit resume information in batches, which can effectively improve the efficiency of job seekers submitting resumes in batches, and the generated resumes are all targeted resumes, rather than submitting the same resume to all companies, so that better resume submission effects can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0039] Figure 1 A schematic diagram of an intelligent resume generation system based on natural language processing according to the present invention.

[0040] Figure 2 A schematic diagram of the process of acquiring a word set in the present invention. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0042] like Figure 1 In one embodiment shown, the present invention provides an intelligent resume generation system based on natural language processing, including a position information acquisition module, a resume information storage module, a template storage module and a resume generation module;

[0043] The job information acquisition module is used to obtain job information from job websites;

[0044] The resume information storage module is used to store the job seeker's work experience information;

[0045] The template storage module is used to store resume templates. The resume templates include a basic information filling area and a work experience filling area. The basic information of the job seeker has been filled in advance in the basic information filling area.

[0046] The resume generation module is used to filter work experience information based on recruitment information using natural language processing technology, obtain screening information, and fill the screening information into the salary and experience filling area in the resume template to generate a resume.

[0047] The above process automatically obtains recruitment information from recruitment websites and uses natural language processing technology to automatically obtain screening information for filling in the work experience filling area in the resume template from the job seeker's work experience information based on the recruitment information, thereby realizing intelligent generation of resumes. When job seekers need to submit resume information in batches, a resume text can be quickly generated, which can effectively improve the efficiency of job seekers submitting resumes in batches. Moreover, the generated resumes are targeted resumes, rather than submitting the same resume to all companies, so that better resume submission results can be obtained.

[0048] Specifically, the basic information filling area of ​​the resume template is used to fill in the basic information of the job seeker, including age, contact information, name, ID number, education and certificates, etc.

[0049] The basic information filling area is located above the work experience filling area.

[0050] Preferably, the position information acquisition module includes a control unit and a crawler unit;

[0051] The control unit is used to adaptively calculate the access interval based on the crawling records;

[0052] The crawler unit is used to visit the recruitment website based on the access interval and crawl the recruitment information published by the recruitment website.

[0053] Acquiring recruitment information based on adaptive access intervals can effectively reduce the probability of the problem of too long or too short access intervals in fixed access interval storage. If the access interval is too long, the latest job posting recruitment information cannot be obtained in time, so that job seekers cannot submit applications for the job in time; and if the access interval is set too short, it is easy to be identified as a crawler by the server, triggering the anti-crawler mechanism, so that operations such as changing the IP address are required to continue crawling recruitment information.

[0054] Preferably, the crawling record includes the crawling time, the server response time, the received HTTP status code and the number of newly added recruitment information obtained in the crawling.

[0055] Specifically, the server response time refers to the time required for the client to send an access request to the recruitment website's server to complete the request and return the response data.

[0056] HTTP status code refers to a three-digit code returned by the server in response to a client request in the HTTP protocol. These status codes are used to inform the client of the processing result of the request so that the client can take corresponding processing measures according to different status codes.

[0057] The number of newly added job postings obtained during this crawling refers to the number of new job postings posted on the recruitment website when this visit is compared with the previous visit.

[0058] Preferably, the access interval is adaptively calculated based on the crawling records, including:

[0059] After the previous crawl is completed, the visit interval for the next visit to the recruitment website is calculated:

[0060] Use S h represents the interval between the hth and h-1th visits to the recruitment website, then S h The calculation formula is:

[0061]

[0062] S h-1 represents the interval between the h-1th and h-2th visits to the recruitment website, bsh-1 and bs h-2 They represent the visit feature values ​​of the h-1th and h-2th visits to the recruitment website, sas represents the duration set in advance, S mi and S ma Indicates the minimum and maximum values ​​of the access interval set in advance;

[0063] When the value of h is greater than 2, the values ​​of S1 and S2 are S ma ;

[0064] For the i-th visit to the recruitment website, i∈{h-1,h-2}, the calculation formula of the visit feature value is:

[0065]

[0066] bs i represents the visit feature value of the i-th visit to the recruitment website, resp i represents the server response time when the i-th visit to the recruitment website, resp cmp Indicates the preset response time comparison value, wrgnum i Indicates the number of HTTP status codes of preset types among all types of HTTP status codes returned by the server during the i-th visit to the recruitment website. wrgtypnum indicates the total number of all types of HTTP status codes returned by the server during the i-th visit to the recruitment website. newnum i represents the number of newly added job information crawled during the i-th visit to the job website, newnum cmp Indicates the preset quantity comparison value, d1 indicates the response time calculation factor, d2 indicates the status code calculation factor, and d3 indicates the new quantity calculation factor.

[0067] The access interval of the present invention is changed based on the previous access interval, which can reduce the impact of occasional high-frequency updates on the overall access interval, so that the access interval can be better adapted to different access results. Specifically, when the access characteristic value of the h-1th access is less than the access characteristic value of the h-2th access, the access interval of the present invention will be reduced accordingly, otherwise, the access interval of the present invention will be increased accordingly. And the increase or decrease amplitude is the same as bs h-1 and bs h-2 The absolute value of the difference between h-1 and bs h-2The larger the absolute value of the difference is, the greater the increase or decrease is, so that the access interval can change more closely with the access results, thereby improving the effectiveness of the access interval. While reducing the probability of being identified by the anti-crawler mechanism, it can avoid the access interval being too large, resulting in the inability to obtain newly released recruitment information on the recruitment website in a timely manner.

[0068] The access characteristic value of the present invention is obtained by comprehensive calculation from multiple angles based on the response time, HTTP status code and the number of newly added recruitment information, so that the access characteristic value can better indicate the urgency of visiting the recruitment website. Specifically, when the response time is longer, the number of HTTP status codes of preset types among all types of HTTP status codes received from the server is larger, and the number of newly added recruitment information is smaller, it means that the access interval of the next visit should be extended; when the response time is shorter, the number of HTTP status codes of preset types among all types of HTTP status codes received from the server is smaller, and the number of newly added recruitment information is larger, it means that the access interval of the next visit should be shortened.

[0069] Such a characteristic value calculation method can not only effectively reduce the access pressure on the recruitment website caused by the crawling process of the recruitment website of the present invention, but also obtain updated recruitment information more timely.

[0070] Preferably, the values ​​of d1, d2 and d3 are and

[0071] Preferably, the preset response time comparison value is 5 seconds.

[0072] Preferably, the quantitative comparison value is 1000.

[0073] Preferably, the preset types of HTTP status codes include 429 status code and 430 status code.

[0074] Preferably, the minimum and maximum values ​​of the access interval are 1 minute and 12 minutes, respectively.

[0075] Preferably, the duration set in advance is 2 minutes.

[0076] Preferably, visiting a recruitment website based on the access interval and crawling the recruitment information published on the recruitment website includes:

[0077] After the h-1th crawl is completed, the access interval of the hth visit to the recruitment website is calculated;

[0078] After calculating the hth visit interval to the recruitment website, the duration is S hThe countdown,

[0079] After the countdown ends, the recruitment website is visited to crawl the recruitment information on the recruitment website.

[0080] Preferably, the position information acquisition module further includes a comparison unit and a storage unit;

[0081] The comparison unit is used to compare all the recruitment information obtained by crawling this time with all the recruitment information obtained by crawling last time, and determine whether there is any newly added recruitment information. If so, the newly added recruitment information is stored and sent to the storage unit for storage.

[0082] Specifically, when crawling, in order to avoid missing some recruitment information due to excessive recruitment information published on the recruitment website, the present invention adopts a holistic crawling method, that is, each crawling crawls all the recruitment information about the set positions on the website.

[0083] All recruitment information obtained in this crawling and all recruitment information obtained in the last crawling can be stored in two texts respectively. The part of the text of all recruitment information obtained in this crawling that belongs to the insertion operation can be obtained through the diff algorithm to obtain the newly added recruitment information.

[0084] The Diff algorithm is a common text comparison algorithm that can find the differences between two texts. Many programming languages ​​have libraries that implement this algorithm, such as Python's difflib module.

[0085] The main purpose of the Diff algorithm is to find differences by comparing the lines or characters between two texts, and assign an operation (insert, delete, or modify) to each difference. Classic Diff algorithms include the greedy algorithm and the longest common subsequence (LCS) algorithm. The greedy algorithm compares the lines or characters of the two texts, matches the same parts as much as possible, and divides the different parts into deletion or insertion operations. The LCS algorithm finds the longest common subsequence between the two texts, finds the same parts as much as possible, and divides the different parts into deletion or insertion operations.

[0086] Preferably, the recruitment information includes basic information, job description, educational requirements, work experience requirements, skill requirements, and certificate requirements;

[0087] The recruitment website contains multiple pages, and each page contains multiple recruitment information.

[0088] Specifically, when a job title is entered in a search box on a recruitment website, the search results are usually displayed in a list format. In order to avoid excessive data volume on a single page, the search results are usually displayed in a paginated format.

[0089] Preferably, the basic information includes the job title, work location and company name of the position being recruited;

[0090] The job description includes the job content of the position, such as what the person is mainly responsible for;

[0091] Work experience requirements include project experience requirements and years of experience in the position. The project experience requirements are intended to recruit people with relevant experience so that they can start work directly and reduce the training time. Many positions will have years of experience requirements, such as having worked in human resources management positions for more than 5 years.

[0092] Skill requirements are the skills that the person recruited for the position needs to have. Different positions require different skills. For example, for programmers, job seekers are generally required to have proficiency in one or several programming algorithms.

[0093] The skill requirements are the certificates that the person recruited for the position needs to have. For example, for lawyers, a lawyer's practice certificate is generally required.

[0094] Preferably, the resume generation module includes a keyword acquisition unit and an extraction unit;

[0095] The keyword acquisition unit is used to acquire a keyword set from the recruitment information;

[0096] The extraction unit is used to obtain screening information from the job seeker's work experience information based on a keyword set.

[0097] The keyword acquisition algorithm can extract keywords from recruitment information, so that these obtained keywords can be used to filter information and generate targeted texts.

[0098] Preferably, the resume generation module further includes a filling unit, which is used to fill the screening information into the salary and experience filling area in the resume template.

[0099] Preferably, a keyword set is obtained from the recruitment information, including:

[0100] Perform word segmentation on the text of the recruitment information to obtain a word set;

[0101] Calculate the ranking coefficient of each word in the word set respectively;

[0102] Get keyword set words based on ranking coefficients.

[0103] Preferably, the text of the recruitment information is segmented to obtain a word set, such as Figure 2 As shown, including:

[0104] Remove useless characters from the text of the recruitment information to obtain the text to be segmented;

[0105] Use a word segmentation algorithm to segment the text to be segmented, obtain multiple words, and store the obtained words in a word set;

[0106] Delete the stop words in the word set.

[0107] Specifically, the useless characters include punctuation marks.

[0108] Specifically, the word segmentation algorithm can be the forward maximum matching method, the reverse maximum matching method, the bidirectional maximum matching method, etc.

[0109] The forward maximum matching method matches several consecutive characters in the text to be segmented with the word list from left to right. If the match is successful, a word is segmented. The core idea of the forward maximum matching method is to use the longest word as much as possible during the segmentation process to achieve the best segmentation effect.

[0110] The basic principle of the reverse maximum matching method is similar to that of the forward maximum matching algorithm, but the segmentation direction is opposite. The reverse maximum matching method starts from the end of the document to be processed for matching and scanning. Each time, the last i characters (the longest number of words in the dictionary) are taken as the matching field. If the match fails, the first character of the matching field is removed and the matching continues. The word segmentation dictionary used by the reverse maximum matching method is a reverse dictionary, and each entry in it will be stored in reverse order.

[0111] The bidirectional maximum matching method is an algorithm that combines the forward maximum matching method and the reverse maximum matching method, aiming to improve the accuracy and efficiency of word segmentation. This method first segments the text using the forward maximum matching method and the reverse maximum matching method respectively, and then compares the results of the two methods, and selects the one with fewer word segmentations as the final result.

[0112] Specifically, the stop words include "de", "le", "shi", etc.

[0113] The steps to delete the stop words in the word set are as follows:

[0114] Establish a stop word list before word segmentation: First, a stop word list needs to be prepared, which contains those words that are not very important in the analysis, frequently appear but contribute little to understanding the text (such as "de", "le", "shi", etc.).

[0115] Filter the word segmentation results: Compare the segmented words with the stop word list and remove the words in the list.

[0116] Preferably, obtain the keyword set words based on the ranking coefficient, including:

[0117] The first S×N words with the largest ranking coefficients are stored as keywords in the keyword set, where N represents the total number of words in the keyword set and S represents a preset ratio.

[0118] For example, the preset ratio is 5%.

[0119] When the present invention obtains keywords, it does not calculate by setting a fixed ranking coefficient threshold, because the more times a word appears, the more important it is, and the more it should be selected as a keyword. If a fixed ranking coefficient threshold is set, then when the length of the text of the recruitment information is too long, the calculated ranking information of the word will be too small, which may result in the absence of words greater than the ranking coefficient threshold, and the acquisition of keywords fails.

[0120] Preferably, the calculation formula of the ranking coefficient is as follows:

[0121]

[0122] ordval v Indicates the ranking coefficient of word v, numloca v Indicates the total number of words v contained in the local scope, locarenum indicates the total number of words in the local scope, numove v represents the total number of words v contained in the text of the recruitment information; ovenum represents the total number of words contained in the text of the recruitment information, and α represents the adaptive weight.

[0123] The ranking coefficient of the present invention is calculated comprehensively from the number of words v in the partial range and the number of words v in the text of the entire recruitment information, which can further improve the accuracy of the extracted keywords. If the ranking coefficient is calculated only based on the local range, then if the number of a word in the local range is small, and the number in the text of the entire recruitment information is large, it is easy for the word to not be selected as a keyword, which is obviously wrong. If the ranking coefficient is calculated only based on the global range, it is easy to use some commonly used words with a high frequency of occurrence in the industry as the final selected keywords, and it is impossible to select more targeted words as keywords based on these keywords, so that the generated resume is less targeted.

[0124] Specifically, the words counted by locarenum and ovenum are words in the word set.

[0125] Preferably, the calculation formula of the adaptive weight is:

[0126]

[0127] Θ represents a preset value.

[0128] The weight of the present invention is not a fixed value, because a fixed value cannot be well applied to different text distribution situations. When v belongs to a keyword for recruitment, if the word v, which originally belongs to the keyword for job recruitment, only appears frequently in a local range, and appears less frequently in other areas, if the weight is set too large, it is easy to make the calculated ranking coefficient too large. If the weight is set too small, it is easy to refer too much to the characteristics of the local distribution of the word, without considering the global distribution characteristics, thereby reducing the accuracy of the keywords finally obtained. Therefore, the present invention calculates adaptive weights based on the distribution characteristics of the words in the text, effectively improving the applicability of the calculated weights, and then improving the applicability of the calculated ranking coefficients.

[0129] Preferably, the preset value is 1.

[0130] Preferably, the local range is determined in the following manner:

[0131] When v first appears in the text of the recruitment information at the Lth v When the local range corresponds to the text of the recruitment information, the number of lines in the text is [L v -NL,L v +NL], NL represents the preset number of lines.

[0132] Preferably, the line division method of the recruitment information text is the line division method of the recruitment information in the original webpage at a resolution of 1080p. For example, the original webpage stipulates that at a resolution of 1080p, the number of words in each line is M. Here, M can be determined according to the size of the text. The larger the text, the smaller M.

[0133] For example, if the text width is 10.8, then the M value is 100.

[0134] Preferably, the preset number of rows is 5.

[0135] Preferably, obtaining screening information from the job seeker's work experience information based on a keyword set includes:

[0136] The job seeker's work experience information includes multiple project experience information;

[0137] All the project experience information that contains keywords in the keyword set in the text of the project experience information is used as the screening information.

[0138] Project experience information includes the project name, the work content in the project, the project introduction and the project results.

[0139] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An intelligent resume generation system based on natural language processing, characterized in that: It includes a position information acquisition module, a resume information storage module, a template storage module and a resume generation module; The job information acquisition module is used to obtain job information from job websites; The resume information storage module is used to store the job seeker's work experience information; The template storage module is used to store resume templates. The resume templates include a basic information filling area and a work experience filling area. The basic information of the job seeker has been filled in advance in the basic information filling area. The resume generation module is used to filter work experience information based on recruitment information using natural language processing technology, obtain screening information, and fill the screening information into the salary and experience filling area in the resume template to generate a resume.

2. According to the intelligent resume generation system based on natural language processing according to claim 1, it is characterized in that: The job information acquisition module includes a control unit and a crawler unit; The control unit is used to adaptively calculate the access interval based on the crawling records; The crawler unit is used to visit the recruitment website based on the access interval and crawl the recruitment information published by the recruitment website.

3. According to claim 2, the intelligent resume generation system based on natural language processing is characterized in that: The position information acquisition module also includes a comparison unit and a storage unit; The comparison unit is used to compare all the recruitment information obtained by crawling this time with all the recruitment information obtained by crawling last time, and determine whether there is any newly added recruitment information. If so, the newly added recruitment information is stored and sent to the storage unit for storage.

4. According to claim 2, the intelligent resume generation system based on natural language processing is characterized in that: The crawling records include the crawling time, the server response time, the received HTTP status code, and the number of newly added job information obtained during the crawling.

5. The intelligent resume generation system based on natural language processing according to claim 1 is characterized in that: Recruitment information includes basic information, job description, education requirements, work experience requirements, skills requirements and certification requirements; The recruitment website contains multiple pages, and each page contains multiple recruitment information.

6. The intelligent resume generation system based on natural language processing according to claim 5, characterized in that: Basic information includes the job title, job location and company name of the position being recruited; The job description includes the job content of the position; Work experience requirements include project experience requirements and job experience requirements; Skill requirements are the skills that the person recruited for this position needs to possess; The skill requirements are the certificates that the person recruited for this position needs to have.

7. The intelligent resume generation system based on natural language processing according to claim 1 is characterized in that: The resume generation module includes a keyword acquisition unit and an extraction unit; The keyword acquisition unit is used to acquire a keyword set from the recruitment information; The extraction unit is used to obtain screening information from the job seeker's work experience information based on a keyword set.

8. The intelligent resume generation system based on natural language processing according to claim 7 is characterized in that: Get a set of keywords from job postings, including: Perform word segmentation on the text of the recruitment information to obtain a word set; Calculate the ranking coefficient of each word in the word set respectively; Get keyword set words based on ranking coefficients.

9. The intelligent resume generation system based on natural language processing according to claim 8, characterized in that: Perform word segmentation on the text of the recruitment information to obtain a word set, including: Remove useless characters from the recruitment information to obtain the text to be segmented; Use a word segmentation algorithm to perform word segmentation on the text to be segmented, obtain multiple words, and store the obtained words into a word set; Remove stop words from a set of words.

10. The intelligent resume generation system based on natural language processing according to claim 8, characterized in that: Get keyword set words based on ranking coefficients, including: The first S×N words with the largest ranking coefficients are stored as keywords in the keyword set, where N represents the total number of words in the keyword set and S represents a preset ratio.

Images