CN115934899A - IT industry resume recommendation method and device, electronic equipment and storage medium - Google Patents
IT industry resume recommendation method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN115934899A CN115934899A CN202310175551.0A CN202310175551A CN115934899A CN 115934899 A CN115934899 A CN 115934899A CN 202310175551 A CN202310175551 A CN 202310175551A CN 115934899 A CN115934899 A CN 115934899A
- Authority
- CN
- China
- Prior art keywords
- text
- resume
- recruitment
- job seeker
- enterprise
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 77
- 230000007115 recruitment Effects 0.000 claims abstract description 163
- 238000007781 pre-processing Methods 0.000 claims abstract description 15
- 238000004364 calculation method Methods 0.000 claims description 26
- 238000012549 training Methods 0.000 claims description 15
- 230000011218 segmentation Effects 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 9
- 238000012216 screening Methods 0.000 claims description 3
- 238000004422 calculation algorithm Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000013135 deep learning Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 3
- 230000003245 working effect Effects 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a resume recommendation method and device in IT industry, electronic equipment and a storage medium, wherein the method comprises the following steps: acquiring an enterprise recruitment text and a plurality of job seeker resume texts and preprocessing the enterprise recruitment text and the resume texts; dividing the preprocessed enterprise recruitment text and job seeker resume text into a plurality of different text blocks, and setting corresponding recruitment preference weighted values for the different text blocks according to recruitment requirements; calculating the matching degree between the text block of each enterprise recruitment text and the text block of the corresponding resume text of the job seeker; and multiplying the matching degree of each text block by the corresponding recruitment preference weighted value, adding the multiplying degrees to obtain the similarity of the plurality of enterprise recruitment texts and the job seeker resume texts, sorting in a descending order according to the similarity, and selecting the job seeker resume texts with the top sorting as recommended resumes. The invention solves the problem that the resume in the IT industry is difficult to be recommended individually, realizes the efficient and accurate matching of the resume and the recruitment post, and improves the working efficiency of enterprise recruiters.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a resume recommendation method and device in the IT industry, electronic equipment and a storage medium.
Background
With the continuous development of a big data system, online recruitment becomes a main recruitment mode of IT enterprises, particularly the Internet IT industry. In the face of a large number of candidates, how the recruiter of the enterprise selects the talents meeting the requirements of the company is the biggest problem faced by the recruiter of the enterprise. However, the mainstream recruitment website usually adopts a fuzzy matching algorithm, and only relevant keywords are matched in the resume received by the enterprise recruiter; meanwhile, the time for the enterprise recruiter to screen resumes is too long, and negative experience is brought to job seekers.
Currently, resume recommendation methods for the IT industry mainly include three types: resume recommendation based on a personalized recommendation algorithm, resume recommendation based on ontology knowledge and resume recommendation based on a deep learning algorithm.
1) Resume recommendation based on personalized recommendation algorithm: namely, content-based recommendations and collaborative filtering-based recommendations and hybrid model-based recommendations are applied to resume recommendations. The disadvantages of this approach: the first is in the problems of data cold start and data sparseness; secondly, in a real scene, the recruiter generally judges comprehensively according to resume of job hunting and recruitment information of posts, while the resume based on the personalized recommendation algorithm only utilizes the resume and the information of word level of the recruitment information and fails to utilize knowledge in other fields, which results in inaccurate recommendation result.
2) Resume recommendation based on ontology domain knowledge: firstly, an ontology domain knowledge system, such as an IT professional skill knowledge tree, needs to be established; secondly, calculating the matching degree of the resume and the recruitment text before different attribute modules according to a knowledge level knowledge system of the ontology field; and finally, carrying out sequencing recommendation according to the matching degree calculation result. The biggest problems with this approach are: only the professional skills of the workplace, the academic calendar, the colleges and the IT in the resume and the recruitment text are extracted, and the text information such as the work experience and the project experience is ignored, so that the quality of the recommendation result is not high.
3) Resume recommendation based on deep learning algorithm: most of the job seekers recommend suitable job seekers by collecting behavior information of users, and resume and recruitment texts are not utilized.
Therefore, in the conventional resume recommendation algorithm of the IT industry, no scheme can be completely suitable for the resume recommendation characteristic of the IT industry, so that effective resume recommendation can be performed according to the recruitment characteristic of the IT industry.
Disclosure of Invention
The invention provides a resume recommendation method, a resume recommendation device, electronic equipment and a storage medium in the IT industry, and aims to solve the problem that in the prior art, the resume recommendation quality is not high due to the fact that the matching degree of the resume of the IT industry and a recruitment post is low.
The invention provides an IT industry resume recommendation method, which comprises the following steps:
acquiring an enterprise recruitment text and resume texts of a plurality of job seekers, and preprocessing the texts;
dividing the preprocessed enterprise recruitment text and the job seeker resume text into a plurality of different text blocks, and setting corresponding recruitment preference weighted values for the different text blocks according to different recruitment requirements of each enterprise;
respectively calculating the matching degree between the text block of each enterprise recruitment text and the corresponding text block of the job seeker resume text according to the text block type;
and multiplying the matching degree of each text block by the corresponding recruitment preference weighted value, adding the multiplying degrees to obtain the similarity of the plurality of enterprise recruitment texts and the job seeker resume texts, sorting in a descending order according to the similarity, and selecting the job seeker resume texts with the top sorting as recommended resumes.
According to the IT industry resume recommendation method provided by the invention, the text block types comprise simple texts, numerical texts and unstructured texts,
correspondingly, respectively calculating the matching degree between the text block of each enterprise recruitment text and the text block of the corresponding job seeker resume text according to the text block types, wherein the method comprises the following steps:
calculating a first matching degree of the text block of which the text block type is a simple text by adopting a method of traversing a knowledge tree of a domain ontology;
calculating a second matching degree for the text block with the text block type being a numerical text by adopting a numerical matching method;
and inputting the text block with the type of the unstructured text into a pre-trained similarity question-answer model to obtain a third matching degree output by the similarity question-answer model.
According to the IT industry resume recommendation method provided by the invention, for the text block of which the text block type is a simple text, a first matching degree is calculated by a method of traversing a knowledge tree of a domain ontology, and the method comprises the following steps:
dividing the domain ontology into Chinese regional knowledge, industry position name knowledge, chinese higher college knowledge and college professional knowledge;
constructing knowledge trees of ontologies in different fields;
traversing nodes on a knowledge tree by adopting a node assignment method for a knowledge tree in higher schools and specialized knowledge trees in colleges and universities to calculate a first matching degree;
and traversing nodes on the knowledge tree by adopting a node comparison method for the Chinese regional knowledge tree and the industry position name knowledge tree to calculate the first matching degree.
According to the IT industry resume recommendation method provided by the invention, for the text block of which the text block type is a numerical text, a numerical matching method is adopted to calculate a second matching degree, and the method comprises the following steps:
discretizing salary and work experience requirements of the enterprise recruitment text and expected salary and working age of the resume text of the job seeker respectively, and dividing the discretization into a plurality of numerical value intervals;
sequentially assigning values to salaries and work experience requirements in the numerical interval and expected salaries and work years;
and respectively carrying out second matching degree calculation on the salary numerical value text and the working year numerical value text according to the numerical value size relationship.
According to the IT industry resume recommendation method provided by the invention, the text block of which the text block type is the unstructured text is input into a pre-trained similarity question-answer model to obtain a third matching degree output by the similarity question-answer model, and the method comprises the following steps:
extracting the post description and the occupational requirement of the enterprise recruitment text to serve as a problem text;
extracting the working experience and the project experience of the plurality of job seeker resume texts corresponding to the enterprise recruitment positions as a first candidate answer text set;
further screening the candidate answer text set to obtain a second candidate answer text set;
and taking the question text and the second candidate answer text set as the input of the similarity question-answer model to obtain a third matching degree between the enterprise recruitment text and the unstructured text of the job seeker resume text.
According to the IT industry resume recommendation method provided by the invention, before the text block of which the text block type is the unstructured text is input to the pre-trained similarity question-answer model, the similarity question-answer model is trained, and the method specifically comprises the following steps:
acquiring an original sample data set, wherein the original sample data set comprises a real sample, a positive sample and a negative sample, the positive sample is a resume text of a job seeker recommended by an expert, and the negative sample is a resume text of a job seeker randomly extracted;
inputting the original sample data set into the similarity question-answering model, respectively calculating the matching degrees between a true sample and a positive sample and between the true sample and a negative sample to obtain a contrast loss, substituting the contrast loss and a corresponding sample label into a first loss function, and finishing training when the first loss function is converged.
According to the IT industry resume recommendation method provided by the invention, the text preprocessing is carried out on the enterprise recruitment text and the resume texts of a plurality of job seekers, and the method comprises the following steps:
and performing word segmentation and stop word removal on the enterprise recruitment text and the plurality of job seeker resume texts respectively.
The invention also provides a resume recommendation device in IT industry, comprising:
the acquisition module is used for acquiring an enterprise recruitment text and a plurality of job seeker resume texts and performing text preprocessing;
the text dividing module is used for dividing the preprocessed enterprise recruitment text and the job seeker resume text into a plurality of different text blocks and setting corresponding recruitment preference weighted values for the different text blocks according to different recruitment requirements of each enterprise;
the matching degree calculation module is used for calculating the matching degree between the text block of each enterprise recruitment text and the corresponding text block of the job seeker resume text according to the text block type;
and the resume recommendation module is used for multiplying the matching degree of each text block by the corresponding recruitment preference weighted value, adding the multiplied matching degrees to obtain the similarity of the plurality of enterprise recruitment texts and the resume texts of job seekers, sorting the similarity in a descending order, and selecting the resume texts of the job seekers with the top sorting as recommended resumes.
The invention also provides an electronic device, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the program to realize the IT industry resume recommendation method.
The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the IT industry resume recommendation method as described in any of the above.
According to the resume recommendation method, device, electronic equipment and storage medium in the IT industry, provided by the invention, the enterprise recruitment text and the resume texts of a plurality of job seekers are obtained, and the texts are preprocessed; dividing the preprocessed enterprise recruitment text and job seeker resume text into a plurality of different text blocks, and setting corresponding recruitment preference weighted values for the different text blocks according to different recruitment requirements of each enterprise; respectively calculating the matching degree between the text block of each enterprise recruitment text and the text block of the corresponding job seeker resume text according to the text block types; and multiplying the matching degree of each text block by the corresponding recruitment preference weighted value, adding the multiplying degrees to obtain the similarity of the plurality of enterprise recruitment texts and the job seeker resume texts, sorting in a descending order according to the similarity, and selecting the job seeker resume texts with the top sorting as recommended resumes. According to the technical scheme, different recruitment preference weighted values are set according to different recruitment requirements of each enterprise in the IT industry, and the resume with high matching degree is automatically obtained through calculation, so that the problem that the resume is difficult to be personalized and recommended is solved, and the working efficiency of enterprise recruiters is improved.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 is a schematic flow chart diagram of an IT industry resume recommendation method provided by the present invention;
fig. 2 is a schematic diagram illustrating division of an enterprise recruitment text and a job seeker resume text according to text block types in the method for recommending resumes in the IT industry according to the present invention;
FIG. 3 is a model training schematic diagram of a similarity question-answer model of the IT industry resume recommendation method provided by the invention;
FIG. 4 is a schematic structural diagram of an IT industry resume recommending apparatus according to the present invention;
fig. 5 is a schematic structural diagram of an electronic device provided in the present invention.
Reference numerals:
21: an acquisition module; 22: a text dividing module; 23: a matching degree calculation module; 24: and a resume recommending module.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without inventive step based on the embodiments of the present invention, are within the scope of protection of the present invention.
Example one
Referring to fig. 1, the method for recommending the resume in the IT industry provided by this embodiment includes:
step S1: acquiring an enterprise recruitment text and a plurality of job seeker resume texts, and preprocessing the texts;
in the step, word segmentation and stop word removal are carried out on the enterprise recruitment text and the resume texts of a plurality of job seekers. The Chinese word segmentation means that a Chinese character sequence is segmented into a single word, the Chinese word segmentation is the basis of text mining, the text mining is used for finding a Chinese word segment based on text information, and the effect of enabling a computer to automatically recognize the meaning of a sentence can be achieved by successfully carrying out Chinese word segmentation on the input Chinese word segment. For stop words, a stop word list can be created by self, the stop word list is imported when needed, and after word segmentation processing, stop words are introduced to optimize word segmentation results, so that words without practical meanings are removed.
Step S2: dividing the preprocessed enterprise recruitment text and job seeker resume text into a plurality of different text blocks, and setting corresponding recruitment preference weighted values for the different text blocks according to different recruitment requirements of each enterprise;
in the step, the multiple attributes of the enterprise recruitment text and the job seeker resume text are classified and divided into a plurality of different text blocks. And the corresponding recruitment preference weighted values can be conveniently set for different text blocks according to different recruitment requirements.
And step S3: respectively calculating the matching degree between the text block of each enterprise recruitment text and the text block of the corresponding job seeker resume text according to the text block types;
in this step, different algorithms are used to calculate the matching degree according to whether the text block type is a structured text or an unstructured text.
And step S4: and multiplying the matching degree of each text block by the corresponding recruitment preference weighted value, adding the multiplying degrees to obtain the similarity of the plurality of enterprise recruitment texts and the job seeker resume texts, sorting in a descending order according to the similarity, and selecting the job seeker resume texts with the top sorting as recommended resumes.
Specifically, an enterprise recruitment text and a plurality of job seeker resume texts are obtained, and text preprocessing is carried out; dividing the preprocessed enterprise recruitment text and job seeker resume text into a plurality of different text blocks, and setting corresponding recruitment preference weighted values for the different text blocks according to different recruitment requirements of each enterprise; respectively calculating the matching degree between the text block of each enterprise recruitment text and the text block of the corresponding job seeker resume text according to the text block types; and multiplying the matching degree of each text block by the corresponding recruitment preference weighted value, adding the multiplying degrees to obtain the similarity of the plurality of enterprise recruitment texts and the job seeker resume texts, sorting in a descending order according to the similarity, and selecting the job seeker resume texts with the top sorting as recommended resumes. According to the method, different recruitment preference weighted values are set according to different recruitment requirements of each enterprise, and the resume with high matching degree is automatically obtained through calculation, so that the problem that the resume is difficult to be personalized and recommended is solved, and the working efficiency of enterprise recruiters is improved.
Referring to fig. 2, in the present embodiment, text block types include simple text, numeric text and unstructured text,
correspondingly, respectively calculating the matching degree between the text block of each enterprise recruitment text and the text block of the corresponding job seeker resume text according to the text block types, wherein the matching degree comprises the following steps:
calculating a first matching degree of a text block of which the text block type is a simple text by adopting a method of traversing a domain ontology knowledge tree;
calculating a second matching degree of the text block with the text block type of a numerical text by adopting a numerical matching method;
and inputting the text block with the type of the unstructured text into a pre-trained similarity question-answer model to obtain a third matching degree output by the similarity question-answer model.
Specifically, the enterprise recruitment text and the job seeker resume text are divided into structured text and unstructured text, and the structured text comprises: expected job location, job age, expected job category, expected job position, expected industry, academic, college, specialty, expected salary expected in the resume text of the job seeker; and the job site, the working age, the academic requirement, the job position category, the job position name and the salary in the enterprise recruitment text. Unstructured text includes: the job experience and the project experience in the resume text of the job seeker; post description and job assignment requirements in the enterprise recruitment text. The working years and expected salaries in the resume text of the job seeker and the working years and the expected salaries in the recruitment text of the enterprise are collectively referred to as numerical texts. The expected work place, expected job position category, expected job position, academic calendar, colleges and universities in the resume text of the job seeker and the work place, academic requirement, job position category, job position name and the like in the enterprise recruitment text are collectively called as simple texts.
In this embodiment, the text block of the enterprise recruitment text is divided into: working age, expected working place, working property, position name, academic calendar, colleges, professions, position description and job assignment requirements, monthly salaries and skills 10 text blocks, considering that each text block is different according to the recruitment requirements of each enterprise, and setting different recruitment preference weighted values lambda for each text block i I = {1,2,3,4,5,6,7,8,9,10}, where λ 1 …λ 10 And respectively representing the recruitment preference weighted values of the enterprise recruiter to different text blocks.
The similarity calculation formula of the enterprise recruitment text and the job seeker resume text is as follows:
wherein the content of the first and second substances,Rfor the resume text of the job seeker,Ja text for the recruitment of the enterprise,k i a simple text for the job seeker resume text,l i a simple text for the enterprise recruiting text,Sim（k i , l i ）for using knowledge in the ontology domainThe method calculates a first degree of matching,m i numerical text for the job seeker resume text,n i a numeric text for the enterprise recruitment text,Sim（m i , n i ）a second degree of matching is calculated for the numerical matching method,punstructured text that is the job seeker resume text,qunstructured text for enterprise recruitment text,Sim（p, q）is the third matching degree output by the similarity question-answering model. The invention discloses three matching degree calculation methods of knowledge tree, numerical matching and similarity question-answer model calculation semantic similarity based on domain ontology construction, and realizes efficient and accurate matching of resume and recruitment posts.
In this embodiment, for a text block of which the text block type is a simple text, calculating a first matching degree by a method of traversing a knowledge tree of a domain ontology, including:
dividing the domain ontology into Chinese regional knowledge, industry position name knowledge, chinese higher college knowledge and college professional knowledge;
constructing knowledge trees of ontologies in different fields;
traversing nodes on a knowledge tree of China higher schools and a professional knowledge tree of China schools by adopting a node assignment method to calculate a first matching degree;
and traversing nodes on the knowledge tree by adopting a node comparison method for the Chinese regional knowledge tree and the industry position name knowledge tree to calculate the first matching degree.
Specifically, let simple text of the resume text of the job seeker bek i ，i=｛1,2,3,4,5,6｝，k 1 …k 6 And sequentially representing six text blocks of an expected work place, an expected work property, a job title, a academic calendar, an institution and a specialty. Simple text for setting enterprise recruitment text isl i ，i=｛1,2,3,4｝，l 1 …l 4 And four text blocks of work place, work property, job title and academic requirement are sequentially represented. Adopting domain ontology knowledge construction tool (project) to determine Chinese regional knowledge, industry position name knowledge and Chinese heightAnd establishing a hierarchical knowledge tree structure for the knowledge of colleges and universities and the professional knowledge of colleges and universities. For simple texts, different knowledge trees are constructed according to ontologies in different fields, and different calculation methods are adopted to calculate the first matching degree, and the specific steps are as follows:
working site matching degree calculation
The work sites are divided into: north China, northeast China, east China, south China, southwest China and northwest China. According to the Chinese regional knowledge tree, we willk 1 Andl 1 each viewed as a node on the knowledge tree. Traverse the knowledge tree ifk 1 And withl 1 In the same area, thenk 1 And withl 1 The same father node exists, and the similarity of the father node and the father node is 0.5; if it is notk 1 Andl 1 not in the same area, i.e.k 1 Andl 1 the same father node does not exist, and the two father nodes are similar to each other and are 0; if it is notk 1 Andl 1 if the two nodes are the same node, the similarity is 1, and the calculation formula is as follows:
(ii) working property matching degree calculation
The working properties are divided into: full-time, part-time and practice. Will be provided withk 2 Andl 2 respectively assigned the value 0,1,2. Wherein 0 represents the full-time, 1 represents the part-time, and 2 represents practice. If it is notk 2 And withl 2 If equal, the similarity between the two is 1; otherwise, the similarity between the two is 0, and the calculation formula is as follows:
(iii) Job name match calculation
The industry professional names of IT enterprises are divided into: front-end development, back-end developmentMobile research and development, big data, test engineers, operation and maintenance support, product manager, project manager and design nine major categories. According to the knowledge tree of the job title, willk 3 And withl 3 Each viewed as a node on the knowledge tree. Traverse the knowledge tree ifk 3 Andl 3 the job title is the same, i.e.k 3 Andl 3 is the same node, and the similarity of the two nodes is 1; otherwise, the similarity between the two is 0, and the calculation formula is as follows:
(iv) computation of degree of matching of learned calendar
First, colleges and universities in China are divided into seven levels: 985 colleges and universities, 211 colleges and universities, one batch, two batches department of civil affairs, general high-post college and college officer, civil high-post college and college officer. According to the knowledge tree of China colleges and universitiesk 5 And (7) assigning values. If it is notk 5 Is "985 colleges" then isk 5 The value is assigned to 1; if it is usedk 5 Is "211 colleges", thenk 5 An assignment of 0.8; if it is notk 5 The parent node of is "this family batch", thenk 5 Has an assignment of 0.6; if it is usedk 5 The parent node of (1) is 'two lots of family', then k 5 An assigned value of 0.4; if it is usedk 5 Is "other universities", thenk 5 Has a value of 0.2.
Secondly, there are 13 subjects related to IT profession in China higher colleges and universities, 61 professional classifications. According to the specialized knowledge tree of colleges and universities, isk 6 And (4) assigning values. If it is notk 6 Exists in a specialized knowledge tree of colleges and universities, thenk 6 Is 1, otherwise it is 0.
Finally, according to the division of the academic calendar: primary school, junior middle school, senior high vocation, major specialty, basic department, master, doctor, pairk 4 Andl 4 performing a numerical process, i.e. onk 4 Andl 4 in turn assigned the value 1,2,3,4,5,6,7. If it is notk 4 Is greater thanl 4 Then λ is set to 1, otherwise it is set to 0. If there is a matchk 4 Andl 4 need to take comprehensive considerationk 5 ，k 6 The calculation formula of the assignment of the two file blocks is as follows:
in this embodiment, for a text block whose text block type is a numeric text, calculating a second matching degree by using a numeric matching method includes:
discretizing salary and work experience requirements of the enterprise recruitment text and expected salary and working age of the resume text of the job seeker respectively, and dividing the discretization into a plurality of numerical value intervals;
sequentially assigning values to salaries and work experience requirements in the numerical interval and expected salaries and work years;
and respectively carrying out second matching degree calculation on the salary numerical value text and the working year numerical value text according to the numerical value size relationship.
Specifically, discretization processing is carried out on expected salaries and working years in a resume text of the job seeker, salaries and working experience requirements in an enterprise recruitment text. Salaries are divided into five intervals: less than 2k, 2k to 6k, 6k to 10k, 10k to 25k and more than 25k; the working years are divided into three intervals: 1-3 years, 3-5 years and more than 5 years. Assume the numerical text in the resume of job hunting asm i The numerical text in the enterprise recruitment isn i ，i=｛1，2｝。m 1 ，m 2 Representing expected salaries and operational years, respectively.n 1 ，n 2 Respectively representing monthly salaries and work experience requirements. Sequentially assigning values between monthly salaries and work experiencesm 1 ，m 2 ，n 1 ，n 2 . And respectively carrying out second matching degree calculation on the salary numerical value text and the working year numerical value text according to the numerical value size relationship, wherein the calculation formula is as follows:
in this embodiment, inputting the text block of which the text block type is the unstructured text into the pre-trained similarity question-answer model to obtain a third matching degree output by the similarity question-answer model, includes:
extracting post description and job requirements of the enterprise recruitment text as a problem text;
extracting the working experience and the project experience of a plurality of job seeker resume texts corresponding to the enterprise recruitment positions as a first candidate answer text set;
further screening the candidate answer text set to obtain a second candidate answer text set;
and taking the question text and the second candidate answer text set as the input of a similarity question-and-answer model to obtain a third matching degree between the enterprise recruitment text and the unstructured text of the job seeker resume text.
Specifically, the post description and the job request of the enterprise recruitment text are regarded as questions in a QA question-answer, the working experience and the project experience of the job seeker resume text are regarded as answers in the QA question-answer, and the job seeker resume text is delivered to the resume set { R) of the job according to the questions in the QA question-answer 1 ，R 2 ，…，R k And the second candidate answer text set is used as an answer in the QA question and answer, and the first candidate answer text set is further screened to obtain a second candidate answer text set { Rs, …, rm }, wherein s is more than or equal to 1 and less than or equal to m and less than or equal to k.
The similarity question-answering model adopts a deep learning network model combining LSTM and CNN to solve the semantic representation and semantic similarity calculation problems of unstructured texts of enterprise recruitment texts and job seeker resume texts. The parameters of the deep learning network model are as follows: the parameter value of the threshold M is 0.05, the parameter value of the filter _ sizes is [1,2,3,5], the parameter value of the iteration number n _ epochs is 100, the parameter value of the learning rate learning _ rate is 0.05, the parameter value of the word vector embedding _ size is 300, the parameter value of the filter number num _ filters is 500, the whole training set is divided into n _ batch groups in each training, and the batch _ size of each training set comprises 256 training samples.
In this embodiment, before inputting the text block of which the text block type is the unstructured text into the pre-trained similarity question-and-answer model, the method further includes training the similarity question-and-answer model, and specifically includes:
acquiring an original sample data set, wherein the original sample data set comprises a real sample, a positive sample and a negative sample, the positive sample is a job seeker resume text recommended by an expert, and the negative sample is a randomly extracted job seeker resume text;
inputting an original sample data set into a similarity question-answering model, respectively calculating the matching degrees between a true sample and a positive sample and between the true sample and a negative sample to obtain a contrast loss, substituting the contrast loss and a corresponding sample label into a first loss function, and finishing training when the first loss function is converged.
Specifically, enterprise recruitment information of the IT industry is crawled from various recruitment websites through Python and is used as an enterprise recruitment text original data set, resume data of job seekers in the IT industry is collected and is used as a resume text original data set of the job seekers. And extracting work experience and project experience in the job seeker resume text and post description and job-undertaking requirements in the enterprise recruitment text, and performing word segmentation and stop word removal on the job seeker resume text and the enterprise recruitment text by adopting an NLPIR word segmentation tool. And randomly extracting 1000 enterprise recruitment texts, matching the job descriptions and the job requirements in the enterprise recruitment text with the job experiences and project experiences in the resume text of the job seeker according to the position descriptions and the job requirements in the enterprise recruitment text, and manually recommending 20 resumes as true samples for the recruitment enterprise. And (3) pairing the resume texts of the job seekers recommended by the experts with the positions one by one to form a resume-recruitment pair as a positive sample, and similarly pairing the randomly extracted resume texts of the job seekers one by one to serve as a negative sample. The method comprises the steps of training resume-recruitment pairs formed by pairing enterprise recruitment texts and job seeker resume texts in a one-to-one mode as input of a model, training word2vec word vectors obtained through training as word embedding of a neural network model, performing supervised training on the network, respectively calculating matching degrees among true samples, positive samples and negative samples, obtaining contrast loss, bringing the contrast loss and corresponding sample labels into a first loss function, adjusting model parameters when the first loss function is converged, and storing a trained similarity question-answer model. And calculating the matching degree of the complex information between the unstructured text in the enterprise recruitment text and the job seeker resume text by using the trained similarity question-answer model, wherein the training process of the network model is shown in fig. 3.
In the embodiment, word segmentation and stop word removal are respectively carried out on the enterprise recruitment text and the resume texts of a plurality of job seekers.
Preferably, an NLPIR word segmentation tool is adopted to perform word segmentation and stop word removal on the enterprise recruitment text and the resume texts of a plurality of job seekers respectively.
Example two
Referring to fig. 4, the present embodiment provides an IT industry resume recommending apparatus, including:
the acquisition module 21 is configured to acquire an enterprise recruitment text and a plurality of job seeker resume texts, and perform text preprocessing;
the text dividing module 22 is configured to divide the preprocessed enterprise recruitment text and job seeker resume text into a plurality of different text blocks, and set corresponding recruitment preference weighted values for the different text blocks according to different recruitment requirements of each enterprise;
the matching degree calculation module 23 is configured to calculate a matching degree between the text block of each enterprise recruitment text and the text block of the corresponding job seeker resume text according to the text block type;
and the resume recommending module 24 is used for multiplying and adding the matching degree of each text block and the corresponding recruitment preference weighted value to obtain the similarity of the plurality of enterprise recruitment texts and the resume texts of job seekers, sorting the similarity in a descending order, and selecting the resume texts of the job seekers with the top sorting as the recommended resumes.
Further, the obtaining module 21 includes: the first acquisition unit is used for acquiring the enterprise recruitment text; and the text preprocessing unit is used for performing text preprocessing on the acquired enterprise recruitment text and job seeker resume text.
The text division module 22 includes: the text dividing unit is used for dividing the preprocessed enterprise recruitment text and job seeker resume text into a plurality of different text blocks; the preference weight setting unit is used for setting corresponding recruitment preference weights for different text blocks according to different recruitment requirements of each enterprise;
the matching degree calculation module 23 includes: the first matching degree calculating unit is used for calculating a first matching degree of a text block of which the text block type is a simple text by adopting a method of ontology domain knowledge; the second matching degree calculating unit is used for calculating a second matching degree of the text block of which the text block type is a numerical text by adopting a numerical matching method; the third matching degree calculation unit is used for inputting the text blocks of which the text block types are unstructured texts into a pre-trained similarity question-answer model to obtain a third matching degree output by the similarity question-answer model;
the resume recommendation module 24 includes: the similarity calculation unit is used for multiplying the matching degree of each text block by the corresponding recruitment preference weighted value and adding the multiplication results to obtain the similarity of the plurality of enterprise recruitment texts and the job seeker resume texts; and the sorting unit is used for sorting in a descending order according to the similarity and selecting the resume text of the job seeker with the top sorting as the recommended resume.
The implementation processes of the functions and actions of each module in the apparatus are specifically described in the implementation processes of the corresponding steps in the method, so that the relevant parts can be referred to the partial description of the method embodiment, and are not described herein again. The above-described system embodiments are merely illustrative, and some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present invention.
EXAMPLE III
As shown in fig. 5, the present embodiment provides an electronic apparatus including: a processor (processor) 310, a communication Interface (communication Interface) 320, a memory (memory) 330 and a communication bus 340, wherein the processor 310, the communication Interface 320 and the memory 330 communicate with each other via the communication bus 340. The processor 310 may call the logic instructions in the memory 330, and the processor 310 executes the IT industry resume recommendation method according to the above method embodiment, where the method includes:
acquiring an enterprise recruitment text and resume texts of a plurality of job seekers, and preprocessing the texts;
dividing the preprocessed enterprise recruitment text and job seeker resume text into a plurality of different text blocks, and setting corresponding recruitment preference weighted values for the different text blocks according to different recruitment requirements of each enterprise;
respectively calculating the matching degree between the text block of each enterprise recruitment text and the text block of the corresponding job seeker resume text according to the text block types;
and multiplying the matching degree of each text block by the corresponding recruitment preference weighted value, adding the multiplying degrees to obtain the similarity of the plurality of enterprise recruitment texts and the job seeker resume texts, sorting in a descending order according to the similarity, and selecting the job seeker resume texts with the top sorting as recommended resumes.
In addition, the logic instructions in the memory 330 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solution of the present invention or a part thereof which substantially contributes to the prior art may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk, and various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product, the computer program product including a computer program, the computer program being storable on a non-transitory computer readable storage medium, wherein when the computer program is executed by a processor, the computer is capable of executing the method for recommending IT industry resume as described in the above method embodiment, the method comprising:
acquiring an enterprise recruitment text and resume texts of a plurality of job seekers, and preprocessing the texts;
dividing the preprocessed enterprise recruitment text and job seeker resume text into a plurality of different text blocks, and setting corresponding recruitment preference weighted values for the different text blocks according to different recruitment requirements of each enterprise;
respectively calculating the matching degree between the text block of each enterprise recruitment text and the text block of the corresponding job seeker resume text according to the text block types;
and multiplying the matching degree of each text block by the corresponding recruitment preference weighted value, adding the multiplying degrees to obtain the similarity of the plurality of enterprise recruitment texts and the job seeker resume texts, sorting in a descending order according to the similarity, and selecting the job seeker resume texts with the top sorting as recommended resumes.
Example four
The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements an IT industry resume recommendation method as described in the above method embodiments, the method comprising:
acquiring an enterprise recruitment text and a plurality of job seeker resume texts, and preprocessing the texts;
dividing the preprocessed enterprise recruitment text and job seeker resume text into a plurality of different text blocks, and setting corresponding recruitment preference weighted values for the different text blocks according to different recruitment requirements of each enterprise;
respectively calculating the matching degree between the text block of each enterprise recruitment text and the text block of the corresponding job seeker resume text according to the text block types;
and multiplying and adding the matching degree of each text block and the corresponding recruitment preference weighted value to obtain the similarity of the plurality of enterprise recruitment texts and the job seeker resume texts, sorting in a descending order according to the similarity, and selecting the job seeker resume text which is sorted in the front as the recommended resume.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. An IT industry resume recommendation method is characterized by comprising the following steps:
acquiring an enterprise recruitment text and a plurality of job seeker resume texts, and preprocessing the texts;
dividing the preprocessed enterprise recruitment text and the job seeker resume text into a plurality of different text blocks, and setting corresponding recruitment preference weighted values for the different text blocks according to different recruitment requirements of each enterprise;
respectively calculating the matching degree between the text block of each enterprise recruitment text and the corresponding text block of the job seeker resume text according to the text block type;
and multiplying the matching degree of each text block by the corresponding recruitment preference weighted value, adding the multiplying degrees to obtain the similarity of the plurality of enterprise recruitment texts and the job seeker resume texts, sorting in a descending order according to the similarity, and selecting the job seeker resume texts with the top sorting as recommended resumes.
2. The IT industry resume recommendation method of claim 1, wherein the text block types comprise simple text, numeric text, and unstructured text,
correspondingly, respectively calculating the matching degree between the text block of each enterprise recruitment text and the text block of the corresponding job seeker resume text according to the text block types, wherein the matching degree comprises the following steps:
calculating a first matching degree of the text block of which the text block type is a simple text by adopting a method of traversing a knowledge tree of a domain ontology;
calculating a second matching degree for the text block with the text block type being a numerical text by adopting a numerical matching method;
and inputting the text block with the type of the unstructured text into a pre-trained similarity question-answer model to obtain a third matching degree output by the similarity question-answer model.
3. The IT industry resume recommendation method of claim 2, wherein for a text block of which the text block type is a simple text, calculating a first matching degree by traversing a knowledge tree of a domain ontology, comprises:
dividing the domain ontology into Chinese regional knowledge, industry position name knowledge, chinese higher college knowledge and college professional knowledge;
constructing knowledge trees of ontologies in different fields;
traversing nodes on a knowledge tree of China higher schools and a professional knowledge tree of China schools by adopting a node assignment method to calculate a first matching degree;
and traversing nodes on the knowledge tree by adopting a node comparison method for the Chinese regional knowledge tree and the industry position name knowledge tree to calculate a first matching degree.
4. The IT industry resume recommendation method of claim 2, wherein for the text block with the text block type of numerical text, a numerical matching method is adopted to calculate a second matching degree, comprising:
discretizing salary and work experience requirements of the enterprise recruitment text and expected salary and working age of the resume text of the job seeker respectively, and dividing the discretization into a plurality of numerical value intervals;
sequentially assigning values to salaries and work experience requirements in the numerical interval and expected salaries and work years;
and respectively carrying out second matching degree calculation on the salary numerical value text and the working year numerical value text according to the numerical value size relationship.
5. The IT industry resume recommendation method of claim 2, wherein the step of inputting the text block of which the text block type is the unstructured text into a pre-trained similarity question-answer model to obtain a third matching degree output by the similarity question-answer model comprises the steps of:
extracting the post description and the job requirements of the enterprise recruitment text as a problem text;
extracting the working experience and the project experience of the plurality of job seeker resume texts corresponding to the enterprise recruitment positions as a first candidate answer text set;
further screening the candidate answer text set to obtain a second candidate answer text set;
and taking the question text and the second candidate answer text set as the input of the similarity question-answer model to obtain a third matching degree between the enterprise recruitment text and the unstructured text of the job seeker resume text.
6. The IT industry resume recommendation method according to claim 2, wherein before inputting the text block of which the text block type is an unstructured text into a pre-trained similarity question-answer model, training the similarity question-answer model further comprises:
the method comprises the steps of obtaining an original sample data set, wherein the original sample data set comprises a real sample, a positive sample and a negative sample, the positive sample is a job seeker resume text recommended by experts, and the negative sample is a randomly extracted job seeker resume text;
inputting the original sample data set into the similarity question-answering model, respectively calculating the matching degrees between a true sample and a positive sample and between the true sample and a negative sample to obtain a contrast loss, substituting the contrast loss and a corresponding sample label into a first loss function, and finishing training when the first loss function is converged.
7. The IT industry resume recommendation method of claim 1, wherein the text preprocessing of the enterprise recruitment text and the plurality of job seeker resume texts comprises:
and performing word segmentation and stop word removal on the enterprise recruitment text and the plurality of job seeker resume texts respectively.
8. An IT industry resume recommendation device, comprising:
the acquisition module is used for acquiring an enterprise recruitment text and a plurality of job seeker resume texts and performing text preprocessing;
the text dividing module is used for dividing the preprocessed enterprise recruitment text and the job seeker resume text into a plurality of different text blocks and setting corresponding recruitment preference weighted values for the different text blocks according to different recruitment requirements of each enterprise;
the matching degree calculation module is used for calculating the matching degree between the text block of each enterprise recruitment text and the corresponding text block of the job seeker resume text according to the text block type;
and the resume recommending module is used for multiplying the matching degree of each text block by the corresponding recruitment preference weighted value, adding the multiplying degrees to obtain the similarity of the plurality of enterprise recruitment texts and the resume texts of job seekers, sorting the similarity in a descending order, and selecting the resume texts of the job seekers with the top sorting as the recommended resumes.
9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements the IT industry resume recommendation method of any of claims 1-7.
10. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the IT industry resume recommendation method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310175551.0A CN115934899A (en) | 2023-02-28 | 2023-02-28 | IT industry resume recommendation method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310175551.0A CN115934899A (en) | 2023-02-28 | 2023-02-28 | IT industry resume recommendation method and device, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115934899A true CN115934899A (en) | 2023-04-07 |
Family
ID=86701010
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310175551.0A Pending CN115934899A (en) | 2023-02-28 | 2023-02-28 | IT industry resume recommendation method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115934899A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116503031A (en) * | 2023-06-29 | 2023-07-28 | 中国人民解放军国防科技大学 | Personnel similarity calculation method, device, equipment and medium based on resume analysis |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107590133A (en) * | 2017-10-24 | 2018-01-16 | 武汉理工大学 | The method and system that position vacant based on semanteme matches with job seeker resume |
| CN112100999A (en) * | 2020-09-11 | 2020-12-18 | 河北冀联人力资源服务集团有限公司 | Resume text similarity matching method and system |
| CN115564393A (en) * | 2022-10-24 | 2023-01-03 | 深圳今日人才信息科技有限公司 | Recruitment requirement similarity-based job recommendation method |
-
2023
- 2023-02-28 CN CN202310175551.0A patent/CN115934899A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107590133A (en) * | 2017-10-24 | 2018-01-16 | 武汉理工大学 | The method and system that position vacant based on semanteme matches with job seeker resume |
| CN112100999A (en) * | 2020-09-11 | 2020-12-18 | 河北冀联人力资源服务集团有限公司 | Resume text similarity matching method and system |
| CN115564393A (en) * | 2022-10-24 | 2023-01-03 | 深圳今日人才信息科技有限公司 | Recruitment requirement similarity-based job recommendation method |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116503031A (en) * | 2023-06-29 | 2023-07-28 | 中国人民解放军国防科技大学 | Personnel similarity calculation method, device, equipment and medium based on resume analysis |
| CN116503031B (en) * | 2023-06-29 | 2023-09-08 | 中国人民解放军国防科技大学 | Personnel similarity calculation method, device, equipment and medium based on resume analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9990609B2 (en) | Evaluating service providers using a social network | |
| CN105608477B (en) | Method and system for matching portrait with job position | |
| US20200193382A1 (en) | Employment resource system, method and apparatus | |
| CN111831905B (en) | Recommendation method and device based on team scientific research influence and sustainability modeling | |
| Alshmrany | Adaptive learning style prediction in e-learning environment using levy flight distribution based CNN model | |
| CN113656687B (en) | Teacher portrait construction method based on teaching and research data | |
| CN113851020A (en) | Self-adaptive learning platform based on knowledge graph | |
| CN110321421A (en) | Expert recommendation method and computer storage medium for website Knowledge Community system | |
| Cui | Application of deep learning and target visual detection in English vocabulary online teaching | |
| CN115934899A (en) | IT industry resume recommendation method and device, electronic equipment and storage medium | |
| Kuznetsov et al. | Reducing cold start problems in educational recommender systems | |
| CN108153818B (en) | Big data based clustering method | |
| Arafath et al. | Predicting career using data mining | |
| CN113313470A (en) | Employment type evaluation method and system based on big data | |
| CN115115483B (en) | Student comprehensive ability evaluation method integrating privacy protection | |
| CN114896975A (en) | Online education intelligent teaching-aid-oriented autonomous evolution method and system | |
| CN109885647B (en) | User history verification method, device, electronic equipment and storage medium | |
| Thali et al. | Survey on job recommendation systems using machine learning | |
| Mühlbauer et al. | Machine learning for labour market matching | |
| Roy et al. | Automated resume classification using machine learning | |
| Pawade et al. | Automated scoring system for online discussion forum using machine learning and similarity measure | |
| CN111191108A (en) | Software crowdsourcing project recommendation method and system based on reinforcement learning | |
| Khan et al. | Session-Based Course Recommendation Frameworks Using Deep Learning. | |
| XIAa | Personalized Recommendation for Network Teaching Courses Based on Combined Filtering of Deep Learning and K-Means | |
| Chen et al. | Management of Ideological and Political Education of College Based on Data Mining |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20230407 |