A talent resume matching method and system based on artificial intelligence big data

Abstract

The present invention relates to the technical field of resume matching, and specifically discloses a talent resume matching method and system based on artificial intelligence big data, the method comprising: S100, obtaining resume information and recruitment demand information, and collecting a first data set in the resume information and a second data set in the recruitment demand information book through a preset intelligent model; S200, performing big data collection according to the first data set and the second data set, respectively, to obtain a first associated data set and a second associated data set; S300, respectively inputting the first associated data set and the second associated data set into a training model, to obtain a first label tree and a second label tree; evaluating the matching degree between the resume information and the recruitment demand information according to the comparison between the first label tree and the second label tree; the method can realize the matching degree evaluation process between the resume information and the professionalism of the recruitment demand, and assist the efficient screening process of recruiters and applicants.

Description

Technical Field

[0001] The present invention relates to the technical field of resume matching, and in particular to a talent resume matching method and system based on artificial intelligence big data. Background Art

[0002] The rapid development and popular application of Internet technology has enabled the existing corporate recruitment and hiring process to be gradually completed online. Through the transmission of information online, the two-way selection process of recruiters and applicants can be efficiently realized.

[0003] Existing resume matching methods are mainly implemented by screening information one by one. Through the comparison process of recruitment needs and recruiter resume information, matching solutions that meet the needs of applicants and applicants meet the requirements of recruiters can be screened out, thereby improving the success rate of the selection process for applicants and recruiters.

[0004] However, the existing screening process is mainly based on analysis and judgment of basic data, such as educational requirements, work experience requirements, the industry field, etc. Although this screening process can help recruiters narrow down the selection range, for more professional information, recruiters still need to review and judge one by one. Obviously, the existing screening method still has shortcomings in matching recruiters with recruitment needs, resulting in low screening efficiency. Summary of the invention

[0005] The purpose of the present invention is to provide a talent resume matching method and system based on artificial intelligence big data to solve the following technical problems:

[0006] How to achieve an efficient screening process to professionally match recruiters with recruitment needs.

[0007] The purpose of the present invention can be achieved through the following technical solutions:

[0008] A talent resume matching method based on artificial intelligence big data, the method comprising:

[0009] S100, obtaining resume information and recruitment demand information, and collecting a first data set in the resume information and a second data set in the recruitment demand information book respectively through a preset intelligent model;

[0010] S200, performing big data collection according to the first data set and the second data set respectively to obtain a first associated data set and a second associated data set;

[0011] S300, respectively inputting the first associated data set and the second associated data set into the training model to obtain a first label tree and a second label tree; and evaluating the matching degree between the resume information and the recruitment requirement information based on the comparison between the first label tree and the second label tree.

[0012] Furthermore, the calculation process of the matching degree between resume information and recruitment demand information includes:

[0013] Compare the first label tree and the second label tree for label content overlap, and obtain overlap label information;

[0014] By formula

[0015] Calculate the matching degree ;

[0016] Where N is the number of ladder levels of the first label tree and the second label tree, i∈[1,N];

[0017] is the number of label overlaps between the first label tree of the i-th layer and the second label tree of the i-th layer;

[0018] The default label tree for the first label tree and the second label tree of the i-th level ladder;

[0019] The number of labels in the Nth layer that overlap in the ladder is the same as the N-1th layer, from large to small;

[0020] M is the number of items in the sequence, j∈[1,M];

[0021] is the number of labels for the jth item;

[0022] f is a preset weight function, and f is monotonically increasing;

[0023] , It is a preset fixed adjustment factor.

[0024] Furthermore, the process of evaluating the matching degree between resume information and recruitment demand information includes:

[0025] Collect historical recruitment demand matching success data, and establish the corresponding piecewise function h(K) of the score range and quantity of each position type according to the position type;

[0026] By equation Get min{V};

[0027] Where, k∈[1,Y]; Y is the total number of intervals; V is the change value, and V≤Y; The total number of successful matches for the corresponding job type recruitment needs. is a preset fixed coefficient, and <1;

[0028] The matching degree will be obtained The Vth score interval corresponding to the position type To compare:

[0029] like , then the match degree is judged to be high;

[0030] like , then the matching degree is judged to be qualified;

[0031] like , then the match degree is judged to be low.

[0032] Furthermore, the process of establishing the training model includes:

[0033] Step 1: Manually label the training text data;

[0034] Step 2: Perform supervised training on the training text data, perform forward reasoning to obtain the output results of multiple models, label the model output results through manual annotation results, and use the labeled data to train the reward feedback model;

[0035] Step 3: Establish a policy network, generate output results through the policy network, calculate feedback through the reward feedback model, and feed the feedback back to the policy network to obtain output results.

[0036] Furthermore, the method further comprises:

[0037] S400, sort and evaluate the screened recruitment needs based on big data, and display them on the applicant side according to the evaluation results;

[0038] The ranking and evaluation process is determined based on the weight allocation strategy set by the applicant.

[0039] Furthermore, the ranking evaluation process includes:

[0040] Preset W groups of matching feature categories, score each group of matching feature categories according to the resume data of the applicant, and obtain the application feature sequence;

[0041] Based on big data, the corresponding information of each group of matching feature categories of recruitment needs is obtained, and scores are performed to obtain the recruitment feature sequence;

[0042] The applicant sets a judgment weight sequence, and based on the judgment weight sequence, compares and analyzes the application feature sequence and the recruitment feature sequence to obtain a score value;

[0043] All recruitment needs after screening the recruiter are sorted from high to low according to the score value and displayed on the applicant side.

[0044] Furthermore, the process of obtaining the scoring value includes:

[0045] By formula Calculate and obtain the score value P;

[0046] where x∈[1,W], Score the i-th recruitment feature sequence;

[0047] Score the i-th candidate feature sequence;

[0048] is the weight coefficient of the i-th recruitment feature sequence, and .

[0049] A talent resume matching system based on artificial intelligence big data, the system includes an information recruitment end, an application end, a collection module, a preset intelligent model and a training model;

[0050] The system is used to execute the talent resume matching method based on artificial intelligence big data.

[0051] Beneficial effects of the present invention:

[0052] (1) The present invention uses big data to collect related data on the resume information of the applicant and the recruitment demand information of the recruiting end, and obtains the corresponding label tree by identifying the collected information. By comparing the label trees, it can realize the matching evaluation process between the resume information and the professionalism of the recruitment demand, and assist the efficient screening process of recruiters and applicants. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The present invention will be further described below in conjunction with the accompanying drawings.

[0054] Figure 1 It is a flow chart of the steps of the talent resume matching method based on artificial intelligence big data of the present invention. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] See also Figure 1 As shown, in one embodiment, a talent resume matching method based on artificial intelligence big data is provided, and the method includes:

[0057] S100, obtaining resume information and recruitment demand information, and collecting a first data set in the resume information and a second data set in the recruitment demand information book respectively through a preset intelligent model;

[0058] S200, performing big data collection according to the first data set and the second data set respectively to obtain a first associated data set and a second associated data set;

[0059] S300, respectively inputting the first associated data set and the second associated data set into the training model to obtain a first label tree and a second label tree; and evaluating the matching degree between the resume information and the recruitment requirement information based on the comparison between the first label tree and the second label tree.

[0060] Through the above technical solution, this embodiment uses big data to collect related data of resume information on the applicant side and recruitment demand information on the recruitment side, and obtains the corresponding label tree by identifying the collected information. The judgment process of the patent matching degree between the recruiter and the recruitment demand is realized by comparing the label tree; specifically, firstly, a first data set in the establishment information and a second data set in the recruitment demand are collected through a preset intelligent model, wherein the preset intelligent model is implemented by the text keyword extraction technology in the prior art, and the first data set and the second data set can be established by identifying and extracting the relevant text information in the recruitment resume and the recruitment demand; secondly, a big data collection process is performed based on the obtained first data set and second data set, and the collected content is continuously screened through the data in the data set, so as to obtain professional data associated with the recruiter and the recruitment demand; for example, in the first data set Including the school information, major information, historical company information, historical position information, historical position responsibility information, etc. of the recruited personnel, the second data set includes company information, company industry information, company position responsibility information, etc., through the big data collection process based on the first data set and the second data set, it is possible to screen out the first associated data set that can represent the recruited personnel and the second associated data set that can represent the agency recruitment needs respectively; by inputting the first associated data set and the second associated data set into the training model to obtain the first label tree and the second label tree, wherein the training model is obtained by training based on a convolutional neural network text classification model according to massive labeled data, and its specific training method can be achieved through existing technology. Therefore, through the comparison process of the first label tree and the second label tree, the matching degree evaluation process of the resume information and the professionalism of the recruitment needs can be realized, which assists the efficient screening process of recruiters and applicants.

[0061] As an implementation mode of the present invention, the calculation process of the matching degree between resume information and recruitment demand information includes:

[0062] Compare the first label tree and the second label tree for label content overlap, and obtain overlap label information;

[0063] By formula Calculate the matching degree ;

[0064] Where N is the number of ladder levels of the first label tree and the second label tree, i∈[1,N];

[0065] is the number of label overlaps between the first label tree of the i-th layer and the second label tree of the i-th layer;

[0066] The default label tree for the first label tree and the second label tree of the i-th level ladder;

[0067] is the sequence of the number of labels in the N-1th layer with overlapping labels in the Nth layer from large to small; M is the number of items in the sequence, j∈[1,M];

[0068] is the number of labels for the jth item; f is the preset weight function, and f is monotonically increasing;

[0069] , It is a preset fixed adjustment factor.

[0070] Through the above technical solution, this embodiment provides a calculation process for the matching degree between resume information and recruitment demand information, that is, the judgment is made by comparing the overlap of the label content of the first label tree and the second label tree. Specifically, the formula Calculate the matching degree , where the number of layers N of the first label tree and the second label tree is a preset value, and its value range is generally 4-6 layers. By judging the label overlap status of the first label tree and the second label tree, the process of professional matching screening is realized. As the number of layers of the label tree increases, the content range corresponding to the label becomes smaller, so through The calculation process can evaluate the overlap status of the first label tree and the second label. The more overlaps there are, the higher the number of layers of the overlapping labels is, indicating that the professional fit between the applicant and the recruitment requirements is higher. At the same time, is the number of labels for the jth item, is the number of labels in the Nth layer that overlap in the ladder and belong to the N-1th layer from large to small; f is the preset weight function, which is selected and obtained after fitting based on empirical data, and f(k) increases with the increase of k value. Therefore, the more concentrated the distribution of labels at the bottom layer is, the more concentrated the corresponding professional range is. The larger the value of, the better the matching degree. The calculation process, combined with the preset fixed adjustment coefficient , The weighted analysis process can then be used to determine the match between applicants and recruitment requirements.

[0071] It should be noted that the preset fixed adjustment coefficient is set according to the test data fitting, which will not be described in detail here.

[0072] As an implementation mode of the present invention, the process of evaluating the matching degree between resume information and recruitment demand information includes:

[0073] Collect historical recruitment demand matching success data, and establish the corresponding piecewise function h(k) of the score range and quantity of each position type according to the position type;

[0074] By equation Get min{V};

[0075] Where, k∈[1,Y]; Y is the total number of intervals; V is the change value, and V≤Y; The total number of successful matches for the corresponding job type recruitment needs. is a preset fixed coefficient, and ;

[0076] The matching degree will be obtained The Vth score interval corresponding to the position type ] for comparison:

[0077] like , then the match degree is judged to be high;

[0078] like , then the matching degree is judged to be qualified;

[0079] like , then the match degree is judged to be low.

[0080] Through the above technical solution, this embodiment provides a matching The method for comparison and judgment, due to the differences in different industries and positions, the professional characteristics of each field are also different. Therefore, this embodiment collects historical recruitment demand matching success data and adaptively determines the judgment score interval according to the position type. Specifically, first collect historical recruitment demand matching success data, and establish a corresponding piecewise function h(k) of the score interval and quantity of each position type according to the position type. Through the equation Get min{V}, that is, get the number of people * The range of the interval, where the preset fixed coefficient <1, which is set according to empirical data, so the matching degree obtained by The Vth score interval corresponding to the position type A comparison is performed to determine the degree of match.

[0081] It should be noted that the process of establishing the piecewise function h(K) is divided and set according to the overall range interval of the score, and is not limited here.

[0082] As an implementation mode of the present invention, the process of establishing the training model includes:

[0083] Step 1: Manually label the training text data;

[0084] Step 2: Perform supervised training on the training text data, perform forward reasoning to obtain the output results of multiple models, label the model output results through manual annotation results, and use the labeled data to train the reward feedback model;

[0085] Step 3: Establish a policy network, generate output results through the policy network, calculate feedback through the reward feedback model, and feed the feedback back to the policy network to obtain output results.

[0086] Through the above technical solution, this embodiment provides a process for establishing a training model. First, a large amount of training text is manually annotated to mark the desired label data, and the labeled data is supervised trained by GPT-3 to obtain the output results of multiple models through forward reasoning. The model output results are marked by the manual annotation results, and the labeled data is used to train the reward feedback model; finally, the output results are generated by inputting the collected text and applying the policy strategy network, and then the feedback is calculated through the reward feedback model, and the feedback is fed back to the policy strategy network to obtain the output results, thereby realizing the process of establishing the training model.

[0087] It should be noted that the relevant specific training process in this embodiment is based on the existing technology of text AI training and will not be further described in this embodiment.

[0088] As an embodiment of the present invention, the method further includes:

[0089] S400, sort and evaluate the screened recruitment needs based on big data, and display them on the applicant side according to the evaluation results;

[0090] The ranking and evaluation process is determined based on the weight allocation strategy set by the applicant.

[0091] The ranking evaluation process includes:

[0092] Preset W groups of matching feature categories, score each group of matching feature categories according to the resume data of the applicant, and obtain the application feature sequence;

[0093] Based on big data, the corresponding information of each group of matching feature categories of recruitment needs is obtained, and scores are performed to obtain the recruitment feature sequence;

[0094] The applicant sets a judgment weight sequence, and based on the judgment weight sequence, compares and analyzes the application feature sequence and the recruitment feature sequence to obtain a score value;

[0095] All recruitment needs after screening the recruiter are sorted from high to low according to the score value and displayed on the applicant side.

[0096] Through the above technical scheme, the method in this embodiment also includes sorting and evaluating the screened recruitment needs, and displaying them on the applicant side according to the evaluation results; specifically, firstly, a matching feature category is selected, such as the growth potential of the position, the degree of overtime, the frequency of business trips, etc., and the specific number of groups W is set according to the selection. At the same time, different scores are set according to the different contents of the matching feature category, so through the comparison relationship between the content and the score, the application feature sequence and the recruitment feature sequence are obtained; in addition, a judgment weight sequence is set according to the applicant to judge the emphasis of the applicant on the position requirements, and then the score value obtained by comparing and analyzing the application feature sequence and the recruitment feature sequence according to the weight sequence is realized to realize the sorting process of all recruitment needs.

[0097] It should be noted that the points are divided according to the recruiter's setting of corresponding evaluation scores for the party with matching feature categories. For example, the better the growth potential of the position, the higher the score.

[0098] As an implementation manner of the present invention, the process of obtaining the scoring value includes:

[0099] By formula Calculate and obtain the score value P;

[0100] where x∈[1,W], Score the i-th recruitment feature sequence;

[0101] Score the i-th candidate feature sequence;

[0102] is the weight coefficient of the i-th recruitment feature sequence, and .

[0103] Through the above technical solution, this embodiment uses the formula Calculate the score value P, where is the weight coefficient of the i-th recruitment feature sequence, and Therefore, the recruiter will adjust the weight of the expected direction of the position accordingly during the process set by the recruiter. By combining the set weight coefficient sequence with the score difference for analysis and comparison, the score value P can be obtained. The sorting and corresponding display process can be achieved through the comparison process of the score value P.

[0104] In one embodiment, a talent resume matching system based on artificial intelligence big data is provided, the system comprising an information recruitment end, an application end, a collection module, a preset intelligent model and a training model; the system is used to execute the talent resume matching method based on artificial intelligence big data.

[0105] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A talent resume matching method based on artificial intelligence big data, characterized in that: The method comprises: S100, obtaining resume information and recruitment demand information, and collecting a first data set in the resume information and a second data set in the recruitment demand information book respectively through a preset intelligent model; S200, performing big data collection according to the first data set and the second data set respectively to obtain a first associated data set and a second associated data set; S300, respectively inputting the first associated data set and the second associated data set into the training model to obtain a first label tree and a second label tree; and evaluating the matching degree between the resume information and the recruitment requirement information according to the comparison between the first label tree and the second label tree; The calculation process of the matching degree between resume information and recruitment demand information includes: Compare the first label tree and the second label tree for label content overlap, and obtain overlap label information; By formula Calculate the matching degree ; Where N is the number of ladder levels of the first label tree and the second label tree, i∈[1,N]; is the number of label overlaps between the first label tree of the i-th layer and the second label tree of the i-th layer; The preset label tree is the first label tree of the i-th layer ladder and the second label tree of the i-th layer; is the sequence of the number of labels in the N-1th layer with overlapping labels in the Nth layer from large to small; M is the number of items in the sequence, j∈[1,M]; is the number of labels for the jth item; f is the preset weight function, and f is monotonically increasing; , It is a preset fixed adjustment coefficient; The process of evaluating the matching degree between resume information and recruitment demand information includes: Collect historical recruitment demand matching success data, and establish corresponding piecewise functions of score range and quantity for each position type according to the position type ; By equation Get min{V}; Where, k∈[1,Y]; Y is the total number of intervals; V is the change value, and V≤Y; The total number of successful matches for the corresponding job type recruitment needs. is a preset fixed coefficient, and ; The matching degree will be obtained The Vth score interval corresponding to the position type [ , ] for comparison: like , then the match degree is judged to be high; like , then the matching degree is judged to be qualified; like , then the match degree is judged to be low.

2. According to the method for matching talent resumes based on artificial intelligence big data according to claim 1, it is characterized in that: The process of establishing the training model includes: Step 1: Manually label the training text data; Step 2: Perform supervised training on the training text data, perform forward reasoning to obtain the output results of multiple models, label the model output results through manual annotation results, and use the labeled data to train the reward feedback model; Step 3: Establish a policy network, generate output results through the policy network, calculate feedback through the reward feedback model, and feed the feedback back to the policy network to obtain output results.

3. According to the method for matching talent resumes based on artificial intelligence big data in claim 1, it is characterized in that: The method further comprises: S400, sort and evaluate the screened recruitment needs based on big data, and display them on the applicant side according to the evaluation results; The ranking and evaluation process is determined based on the weight allocation strategy set by the applicants.

4. The talent resume matching method based on artificial intelligence big data according to claim 3 is characterized in that: The ranking evaluation process includes: Preset W groups of matching feature categories, score each group of matching feature categories according to the resume data of the applicant, and obtain the application feature sequence; Based on big data, the corresponding information of each group of matching feature categories of recruitment needs is obtained, and scores are performed to obtain the recruitment feature sequence; The applicant sets a judgment weight sequence, and based on the judgment weight sequence, compares and analyzes the application feature sequence and the recruitment feature sequence to obtain a score value; All recruitment needs after screening the recruiter are sorted from high to low according to the score value and displayed on the applicant side.

5. The talent resume matching method based on artificial intelligence big data according to claim 4 is characterized in that: The process of obtaining the score value includes: By formula Calculate and obtain the score value P; where x∈[1,W], Score the i-th recruitment feature sequence; Score the i-th candidate feature sequence; is the weight coefficient of the i-th recruitment feature sequence, and .

6. A talent resume matching system based on artificial intelligence big data, characterized in that: The system includes an information recruitment terminal, an application terminal, a collection module, a preset intelligent model and a training model; The system is used to execute the talent resume matching method based on artificial intelligence big data as described in any one of claims 1 to 5.

Images