Information fusion model training method and position and resume matching method

Abstract

The invention discloses an information fusion model training method and a position and resume matching method. The training method of the information fusion model comprises the steps of obtaining first user information corresponding to a plurality of job hunting users and second user information corresponding to a plurality of recruitment users; generating a training sample set according to the plurality of pieces of first user information, the plurality of pieces of second user information, a job hunting relationship between each job hunting user and a position, and a feedback relationship between a recruitment user corresponding to each position and each job hunting user; and according to the training sample set, training a pre-constructed information fusion model, and obtaining a target information fusion model until an output result of the information fusion model meets a preset training condition. The embodiment of the invention is beneficial to finding out the feature information representing the willingness of the job hunting user and the recruitment user, and is beneficial to improving the satisfaction degree of the two parties to the recommended information.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a training method for an information fusion model and a method for matching positions with resumes. Background Art

[0002] With the popularization of the Internet and the increasing maturity of data processing technology, more and more convenience has been brought to users' lives and work. For example, in the recruitment industry, recruiters can post recruitment information on the recruitment platform and find the required job seekers on the recruitment platform. Job seekers can post resumes on the recruitment platform and find suitable positions on the recruitment platform. Based on this, the recruitment platform can accumulate a large number of resumes and positions.

[0003] In order to provide better services to users at both ends and obtain better service experience as well as to increase job search rate or recruitment rate, in the related technologies, when matching positions to job seekers or matching resumes to recruiters, the positions and resumes are often matched by extracting keywords to obtain matching results, and information is recommended to users at both ends based on the matching results. However, based on the current matching methods, it is difficult to find feature information that represents the wishes of both job seekers and recruiters, and it is not conducive to improving the matching success rate between job seekers and positions, resulting in matching results that are often unsatisfactory. Summary of the invention

[0004] In view of this, the embodiments of the present application provide a training method for an information fusion model and a method for matching positions and resumes, which are helpful in finding characteristic information that characterizes the intentions of both job seekers and recruiters, and are helpful in improving the satisfaction of both parties with the recommended information.

[0005] The embodiment of the present application provides a training method for an information fusion model, comprising: obtaining first user information corresponding to a plurality of job seekers and second user information corresponding to a plurality of recruiting users, wherein the first user information includes job viewing behavior and resume submission behavior, and the second user information includes resume feedback behavior; determining the job-seeking relationship between each job seeker and each job, and the feedback relationship between the recruiting user corresponding to each job and each job seeker based on the plurality of first user information and the plurality of second user information; generating a training sample set based on the plurality of first user information, the plurality of second user information, the job-seeking relationship between each job seeker and the job, and the feedback relationship between the recruiting user corresponding to each job and each job seeker, wherein the training sample set includes a plurality of training samples and a sample label corresponding to each training sample, and the sample label includes Job search relationship label and feedback label, wherein the job search relationship label is a delivery label corresponding to a job seeker who has delivered a resume, or a non-delivery label corresponding to a job seeker who has not delivered a resume, and the feedback label is a positive feedback label or a negative feedback label, a positive feedback label indicates that the job seeker corresponding to the position has marked the resume as suitable, and a negative feedback label indicates that the job seeker corresponding to the position has not marked the resume as suitable; according to the training sample set, a pre-constructed information fusion model is trained, and the output result of the information fusion model is obtained until a preset training condition is met to obtain a target information fusion model, wherein the output result of the information fusion model includes a delivery rate between the job seeker corresponding to the resume and the position and a positive feedback rate between the job seeker corresponding to the position and the resume, and the preset training condition is that the error of the delivery rate output by the model is less than a first threshold, and the positive feedback rate output by the model is less than a second threshold.

[0006] Optionally, the method according to the embodiment of the present application determines the job-seeking relationship between each job-seeking user and each position, as well as the feedback relationship between the recruiting user corresponding to each position and each job-seeking user based on multiple first user information and multiple second user information, including: determining a set of job-seeking users with to-be-confirmed relationships and a set of positions with to-be-confirmed relationships based on multiple first user information and multiple second user information, the job-seeking user set including multiple job-seeking users and resumes corresponding one-to-one to the job-seeking users, and the position set including multiple recruiting users and positions corresponding one-to-one to the recruiting users; for each job-seeking user, determining, based on the first user information, a position having a first relationship with the resume corresponding to the job-seeking user, a position having a second relationship with the resume corresponding to the job-seeking user relationship with the job seeker's corresponding resume, and a position with a third relationship with the job seeker's corresponding resume; for each recruiting user, determine the resume with a fourth relationship with the recruiting user's corresponding position, and a resume with a fifth relationship with the recruiting user's corresponding position based on the second user information; the first relationship is used to indicate that the job seeker's corresponding resume and the position are in a viewed but not submitted relationship; the second relationship is used to indicate that the job seeker's corresponding resume and the position are in a viewed and submitted relationship; the third relationship is used to indicate that the job seeker's corresponding resume and the position are in a unviewed and unsubmitted relationship; the fourth relationship is used to indicate that the recruiting user's corresponding position and the resume are in a positive feedback relationship; the fifth relationship is used to indicate that the recruiting user's corresponding position and the resume are in a negative feedback relationship.

[0007] Optionally, according to the method of the embodiment of the present application, each first user information also includes: user basic information, access behavior information, and context information of the access behavior information; each second user information also includes: position basic information, behavior information of the recruiting user corresponding to the position, and context information of the behavior information of the recruiting user corresponding to the position.

[0008] An embodiment of the present application provides a method for matching positions with resumes, including: obtaining multiple job seekers to be matched and user information of each job seeker, as well as position information corresponding to multiple positions, wherein the user information includes user basic information and resume information; inputting the user information corresponding to the multiple job seekers and the position information corresponding to the multiple positions into an information fusion model, and outputting the delivery rate between each job seeker and each position, as well as the positive feedback rate between the recruiting user corresponding to each position and the job seeker through the information fusion model, wherein the target information fusion model is trained by the above information fusion model training method; determining the matching degree between each job seeker and each position according to the delivery rate between each resume and each position, as well as the positive feedback rate between the recruiting user corresponding to each position and the resume.

[0009] Optionally, according to the method of the embodiment of the present application, the delivery rate corresponds to a first weight parameter, and the positive feedback rate corresponds to a second weight parameter; according to the delivery rate between each job seeker corresponding to each resume and each position, and the positive feedback rate between each recruiting user corresponding to each position and the resume, the matching degree between each job seeker and each position is determined, including: according to the delivery rate and the first weight parameter between each job seeker corresponding to each resume and each position, and the positive feedback rate and the second weight parameter between each recruiting user corresponding to each position and the resume, the matching degree between each job seeker and each position is determined.

[0010] Optionally, according to the method of an embodiment of the present application, the method also includes: in response to the target object visiting the recruitment platform, determining recommendation information to the target object based on the degree of matching; wherein, when the target object is a job seeker, the recommended information includes the top M positions with the highest degree of matching with the job seeker; when the target object is a recruiting user, the recommended information includes the top N job seekers with the highest degree of matching with the corresponding recruitment positions.

[0011] The embodiment of the present application provides a training device for an information fusion model, including: a first acquisition module, used to acquire first user information corresponding to multiple job seekers and second user information corresponding to multiple recruiting users, wherein the first user information includes job viewing behavior and resume submission behavior, and the second user information includes resume feedback behavior; a first processing module, used to determine the job-seeking relationship between each job seeker and each job, and the feedback relationship between the recruiting user corresponding to each job and each job seeker based on the multiple first user information and the multiple second user information; the first processing module is also used to generate a training sample set based on the multiple first user information, the multiple second user information, the job-seeking relationship between each job seeker and the job, and the feedback relationship between the recruiting user corresponding to each job and each job seeker, wherein the training sample set includes multiple training samples and samples corresponding to each training sample Labels, sample labels include job-seeking relationship labels and feedback labels, wherein the job-seeking relationship label is a delivery label corresponding to a job-seeking user who has delivered a resume, or a non-delivery label corresponding to a job-seeking user who has not delivered a resume, and the feedback label is a positive feedback label or a negative feedback label, a positive feedback label indicates that the job-seeking user corresponding to the position has marked the resume as suitable, and a negative feedback label indicates that the job-seeking user corresponding to the position has not marked the resume as suitable; the first processing module is also used to train a pre-constructed information fusion model according to a training sample set, and the output result of the information fusion model is obtained until a preset training condition is met to obtain a target information fusion model, wherein the output result of the information fusion model includes a delivery rate between the job-seeking user corresponding to the resume and the position and a positive feedback rate between the job-seeking user corresponding to the position and the resume, and the preset training condition is that the error of the delivery rate output by the model is less than a first threshold, and the positive feedback rate output by the model is less than a second threshold.

[0012] An embodiment of the present application provides a device for matching positions and resumes, including: a second acquisition module, used to obtain multiple job seekers to be matched and user information of each job seeker, as well as position information corresponding to multiple positions, wherein the user information includes basic user information and resume information; a second processing module, used to input the user information corresponding to the multiple job seekers and the position information corresponding to the multiple positions into an information fusion model, and output the delivery rate between each job seeker and each position, as well as the positive feedback rate between the recruiting user corresponding to each position and the job seeker through the information fusion model, wherein the target information fusion model is trained based on the training method of the information fusion model as above; the second processing module is also used to determine the matching degree between each job seeker and each position based on the delivery rate between each resume and each position, as well as the positive feedback rate between the recruiting user corresponding to each position and the resume.

[0013] An embodiment of the present application provides an electronic device, which includes a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the steps of the above method are implemented.

[0014] An embodiment of the present application provides a computer-readable storage medium, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the steps of the above method are implemented.

[0015] An embodiment of the present application provides a computer program product, which includes computer program instructions. When the computer program instructions are executed by a processor, the steps of the above method are implemented.

[0016] According to an embodiment of the present application, by obtaining the job-seeking behavior and resume submission behavior of job seekers and the resume feedback information of recruiting users, the job-seeking relationship between job seekers and positions, as well as the feedback relationship between recruiting users and job seekers are found to construct training samples. When constructing training samples, the job-seeking relationship and feedback relationship between recruiting users and job seekers are combined to generate training samples that integrate the intentions of both parties, and the training targets are constructed based on the submission rate between resumes and positions and the positive feedback rate between positions and recruiting users and resumes. Based on this, after using the samples to train the pre-constructed information fusion model, the model can deeply mine the intention information between job seekers and recruiting users. In the subsequent use process, the intention information between job seekers and recruiting users can be mined. In this way, based on the intention information of both parties as a reference factor for matching calculation, the matching objects provided to job seekers and recruiting users can make the matching objects more in line with the wishes of both parties, which is conducive to promoting a handshake between the two parties and improving recruitment efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solution of the embodiments of the present application, the following briefly introduces the drawings in the embodiments of the present application.

[0018] Figure 1 It is a schematic diagram of the system architecture of an embodiment of the present application.

[0019] Figure 2 It is a flowchart of the training method of the information fusion model of the embodiment of the present application.

[0020] Figure 3 It is a flowchart of the method for matching positions with resumes in an embodiment of the present application.

[0021] Figure 4 It is a structural diagram of the information fusion model of an embodiment of the present application.

[0022] Figure 5 It is a structural block diagram of the training device of the information fusion model of the embodiment of the present application.

[0023] Figure 6 It is a structural block diagram of a device for matching positions and resumes according to an embodiment of the present application.

[0024] Figure 7 It is a schematic diagram of an electronic device used to implement an embodiment of the present application. DETAILED DESCRIPTION

[0025] The principles and spirit of the present application will be described below with reference to several exemplary embodiments. It should be understood that the purpose of providing these embodiments is to make the principles and spirit of the present application clearer and more thorough, so that those skilled in the art can better understand and implement the principles and spirit of the present application. The exemplary embodiments provided herein are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments herein, all other embodiments obtained by ordinary technicians of the art without creative work are within the scope of protection of this application.

[0026] It should be noted that the acquisition, storage, use, and processing of data in the embodiments of the present application are in compliance with the relevant provisions of national laws and regulations.

[0027] In this document, terms such as first, second, third, etc. are only used to distinguish one entity (or operation) from another entity (or operation), but not to require or imply any order or relationship between these entities (or operations).

[0028] Embodiments of the present application relate to terminal devices and / or servers. Those skilled in the art will appreciate that the embodiments of the present application may be implemented as a system, apparatus, device, method, computer-readable storage medium, or computer program product. Therefore, the present disclosure may be specifically implemented in at least one of the following forms: complete hardware, complete software, or a combination of hardware and software. According to the embodiments of the present application, the present application requests protection for a training method for an information fusion model, a person-job matching method, an apparatus, an electronic device, a computer-readable storage medium, and a computer program product.

[0029] With the popularization of the Internet and the increasing maturity of data processing technology, more and more convenience has been brought to users' lives and work. For example, in the recruitment industry, recruiters can post recruitment information on the recruitment platform and find the required job seekers on the recruitment platform. Job seekers can post resumes on the recruitment platform and find suitable positions on the recruitment platform. Based on this, the recruitment platform can accumulate a large number of resumes and positions.

[0030] In order to provide better services to users at both ends and obtain better service experience and improve the efficiency of the entire job search and recruitment process, in the relevant technology, when matching positions to job seekers or matching resumes to recruiters, the positions and resumes are often matched by extracting keywords to obtain the matching degree, and information is recommended to users at both ends based on the matching degree. However, those skilled in the art have found that in the recruitment industry, a complete recruitment process has a long link. An effective recruitment process involves not only the willingness of job seekers, but also the willingness of recruiters. However, based on the current matching method, it is difficult to find characteristic information that represents the willingness of both job seekers and recruiters, and it is not conducive to improving the matching success rate between job seekers and positions, resulting in the recommendation information often being unsatisfactory.

[0031] Based on the above considerations, the embodiments of the present application provide a method for training an information fusion model and a method for matching positions and resumes. Through the job-seeking users' position viewing behavior and resume submission behavior, as well as the recruiting users' resume feedback behavior, the intention information between job-seeking users and recruiting users is deeply mined, so that the output results of the model have more reference value. Furthermore, based on the results of the model output, the matching objects provided to both job-seeking users and recruiting users can be more in line with the wishes of both parties, which is conducive to promoting a handshake between the two parties and improving recruitment efficiency.

[0032] Figure 1 A schematic diagram of a system architecture of an embodiment of the present application is shown. Figure 1As shown, the system includes a terminal device and a server 104. The terminal device may include at least one of the following: a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart TV, various wearable devices, an augmented reality AR device, a virtual reality VR device, and the like. The terminal device may use the recruitment software provided by the recruitment platform. For example, a client may be installed on the terminal device. The client may be a client that specifically performs a specific function (such as an application app), or a client that is embedded with a variety of application applets (with different functions), or a client that logs in through a browser. Job seekers may perform operations on the terminal device 102a. For example, job seekers may open the terminal device 102a, use the recruitment software, and perform operations such as viewing positions and posting resumes in the recruitment software. Optionally, the recruitment user may also perform operations on another terminal device 102b, such as posting positions and position-related introductions in the recruitment software.

[0033] In an embodiment of the present application, the executor of the information fusion model training method and the position and resume matching method may be an electronic device with data processing capabilities, such as a server. Optionally, the server may be a physical server, a virtual server, or a cloud server. No specific restrictions are imposed on the form of the server.

[0034] As a specific example, the server 104 can provide storage and retrieval functions, for example, it can store information such as resumes of job seekers and positions of recruiters. Exemplarily, after the terminal device 102a receives the instruction input by the job seeker, the request information including the instruction is sent to the server 104. After receiving the request information, the server 104 performs corresponding processing, and then returns the processing result information to the terminal device 102a. The user instruction is completed through a series of data processing and information interaction. Based on these instructions, the server 104 can obtain the job search user's position viewing behavior and resume delivery behavior, as well as the resume feedback behavior of the recruiter, deeply mine the intention information between the job seeker and the recruiter, so that the output result of the model is more valuable for reference, and then according to the output result of the model, the matching objects provided to both the job seeker and the recruiter can be more in line with the wishes of both parties, which is conducive to promoting the handshake between the two parties and improving the recruitment efficiency.

[0035] The following is an introduction to the recruitment method provided by the embodiment of the present application in conjunction with the accompanying drawings. Figure 2 A flow chart of a method for training an information fusion model according to an embodiment of the present application is shown. The method includes the following steps 201 to 204 .

[0036] Step 201: Obtain first user information corresponding to a plurality of job seekers and second user information corresponding to a plurality of recruiters.

[0037] Among them, the first user information includes job viewing behavior and resume submission behavior, and the second user information includes resume feedback behavior.

[0038] Step 202: Determine the job-seeking relationship between each job-seeking user and each position, and the feedback relationship between the recruiting user corresponding to each position and each job-seeking user based on the plurality of first user information and the plurality of second user information.

[0039] Step 203: Generate a training sample set based on the information of multiple first users, the information of multiple second users, the job-seeking relationship between each job-seeker and the position, and the feedback relationship between the recruiting user corresponding to each position and each job-seeker.

[0040] Among them, the training sample set includes multiple training samples and a sample label corresponding to each training sample. The sample label includes a job-seeking relationship label and a feedback label. Among them, the job-seeking relationship label is a submission label corresponding to a job-seeking user who has submitted a resume, or a non-submission label corresponding to a job-seeking user who has not submitted a resume. The feedback label is a positive feedback label or a negative feedback label. The positive feedback label indicates that the job-seeking user corresponding to the position has marked the resume as suitable, and the negative feedback label indicates that the job-seeking user corresponding to the position has not marked the resume as suitable.

[0041] Step 204: train the pre-built information fusion model according to the training sample set, and obtain the target information fusion model until the output result of the information fusion model meets the preset training conditions.

[0042] Among them, the output results of the information fusion model include the delivery rate between job seekers and positions corresponding to resumes and the positive feedback rate between recruiting users and resumes corresponding to positions. The preset training conditions are that the error of the delivery rate output by the model is less than the first threshold, and the positive feedback rate output by the model is less than the second threshold.

[0043] The above steps are described in detail below in conjunction with specific embodiments, as shown below.

[0044] In relation to the above step 201, the first user information may include the behavior information of the job seeker, such as job search behavior and resume submission behavior. The second user information may include the behavior information of the recruiter, such as resume feedback information.

[0045] Exemplarily, the behavioral information of job seekers and the behavioral information of recruiters can be obtained from the service equipment of the recruitment platform respectively. For example, the behaviors of job seekers viewing positions and submitting resumes in the recruitment software, and the behaviors of recruiters posting resumes, viewing resumes for a certain position, communicating with job seekers, and sending interview invitations can all be recorded in the service equipment of the recruitment platform.

[0046] After the first user information and the second user information are obtained, step 202 is involved. According to the first user information, positions viewed by the job seeker and positions for which the job seeker has submitted a resume can be found, thereby finding the job search relationship between each job seeker and each position.

[0047] The recruiting user can receive resumes submitted by job seekers for a certain position or resumes recommended by the recruiting platform through the recruitment platform, and thus, in combination with the second user information, the feedback relationship between the recruiting user and the resume can be found. It can be understood that each resume corresponds to a job seeker, so the feedback relationship between the recruiting user and the resume can also be determined as the feedback relationship between the position corresponding to the recruiting user and each job seeker.

[0048] Next, step 203 involves generating a training sample set based on the multiple first user information, multiple second user information, the job-seeking relationship between each job seeker and the position, and the feedback relationship between the recruiting user corresponding to each position and each job seeker.

[0049] For example, based on the first user information, multiple job seekers and their corresponding resumes can be found, and based on the second user information, multiple recruiters and their posted positions can be found. Based on the resumes corresponding to these job seekers and the positions corresponding to the recruiters, each training sample generated includes at least the resumes of the job seekers and the positions posted by the recruiters.

[0050] The sample label corresponding to each training sample can be determined based on the job-seeking relationship between each job-seeking user and each position, and the feedback relationship between each position's corresponding recruiting user and each job-seeking user. For example, the job-seeking relationship between a job-seeking user and a position can be that the job-seeking user has submitted a resume or that the job-seeking user has not submitted a resume; if the job-seeking user has a job-seeking relationship with the position in the sample in which the resume is submitted, and the recruiting user in the sample marks the resume as suitable, then the sample label of the training sample is a positive feedback label.

[0051] If the job seeker in the training sample has a job-seeking relationship of submitting a resume to the position in the sample, and the recruiting user in the sample does not mark the resume as suitable, then the sample label of the training sample is a negative feedback label.

[0052] It is understandable that if the job seeker in the training sample and the position in the sample have a job-seeking relationship in which the resume has not been submitted, and the recruiting user has not marked the resume as suitable, then the sample label of the training sample is a negative feedback label.

[0053] Based on this, by obtaining the user information of the recruiting users and the job-seeking users respectively, and determining the job-seeking relationship and feedback relationship between the two parties, a training sample set that integrates the intentions of both parties is generated.

[0054] After the training sample set is generated, step 204 is involved in which the pre-built information fusion model is trained using the training samples in the training sample set and the sample labels corresponding to the samples.

[0055] Specifically, the training objectives of the training samples include the delivery rate between the resume-corresponding job seeker and the position and the positive feedback rate between the position-corresponding job seeker and the resume. Accordingly, the preset training conditions are that the error of the delivery rate output by the model is less than the first threshold, and the positive feedback rate output by the model is less than the second threshold.

[0056] The output results of the information fusion model include the delivery rate between resumes and positions, and the positive feedback rate between positions and resumes. Based on this, after the output results of the information fusion model meet the preset training conditions, the target information fusion model obtained can mine the intention information between job seekers and recruiters in the subsequent use process. In this way, the intention information of both parties is used as a reference factor for matching calculations, and the matching objects provided to both job seekers and recruiters can be more in line with the wishes of both parties, which is conducive to promoting a handshake between the two parties and improving recruitment efficiency.

[0057] In some optional embodiments of the present application, based on the plurality of first user information and the plurality of second user information, the job search relationship between each job seeker and each position, and the feedback relationship between the recruiting user corresponding to each position and each job seeker are determined. Specifically, the following steps may be referred to:

[0058] Determine, based on the plurality of first user information and the plurality of second user information, a set of job seekers whose relationships are to be confirmed and a set of positions whose relationships are to be confirmed, wherein the set of job seekers includes a plurality of job seekers and resumes corresponding to the job seekers one by one, and the set of positions includes a plurality of recruiters and positions corresponding to the recruiters one by one;

[0059] For each job seeker, according to the first user information, determine a position having a first relationship with the resume corresponding to the job seeker, a position having a second relationship with the resume corresponding to the job seeker, and a position having a third relationship with the resume corresponding to the job seeker;

[0060] For each recruiting user, a resume having a fourth relationship with the position corresponding to the recruiting user and a resume having a fifth relationship with the position corresponding to the recruiting user are determined according to the second user information.

[0061] Specifically, the first relationship is used to indicate that the relationship between the resume of the job seeker and the position is a viewed-undelivered relationship.

[0062] The second relationship is used to indicate that the relationship between the resume of the job seeker and the position is a view and submit relationship.

[0063] The third relationship is used to indicate that the relationship between the job seeker's corresponding resume and the position is an unreviewed and unsubmitted relationship.

[0064] The fourth relationship is used to indicate that there is a positive feedback relationship between the corresponding position of the recruiting user and the resume.

[0065] The fifth relationship is used to indicate that there is a negative feedback relationship between the corresponding position of the recruiting user and the resume.

[0066] Exemplarily, the relationship between job seekers and recruiters can be shown in Table 1.

[0067] Table 1

[0068] Job seekers Recruitment users Whether to deliver Is it positive feedback? Job seeker A Recruitment User A 1 1 Job seeker B Recruitment User A 1 0 Job seeker C Recruitment User A 0 0 Job seeker A Recruit User B 1 0 Job seeker B Recruit User B 1 1 Job seeker C Recruit User B 1 0 …… …… …… ……

[0069] In the "Submission" column of Table 1, 1 indicates that the job seeker has submitted a resume to the position corresponding to the recruiter, and 0 indicates that the job seeker has not submitted a resume to the position corresponding to the recruiter; in the "Positive Feedback" column, 1 indicates that the recruiter has marked the job seeker's resume as suitable, which is a positive feedback relationship, and 0 indicates that the recruiter has not marked the job seeker's resume as suitable, which is a negative feedback relationship.

[0070] It is understandable that if a recruiting user marks a resume received for a position as suitable, it may be that the job seeker has already submitted the resume and the recruiting user marks the resume as suitable; or it may be that the job seeker has not submitted the resume and the recruiting user actively checks the resume and marks it as suitable.

[0071] If the recruiting user does not mark the resume received for the position as suitable, it may be that the job seeker has already submitted the resume but the recruiting user marked the resume as unsuitable; it may also be that the job seeker has already submitted the resume but the recruiting user did not mark the resume within a preset time period, where the preset time period may be 1 month or other time periods; it may also be that the job seeker has not submitted the resume and the recruiting user has not actively checked the resume.

[0072] According to the embodiment of the present application, by combining the job-seeking relationship and feedback relationship between recruiting users and job-seeking users when constructing training samples, training samples that integrate the wishes of both parties are generated, so that the trained model can comprehensively consider the wishes of both parties, promote job-seeking users and recruiting users to reach a handshake, and increase the number of effective recruitments.

[0073] In some embodiments of the present application, each first user information also includes: user basic information, access behavior information, and context information of the access behavior information; each second user information also includes: position basic information, behavior information of the recruiting user corresponding to the position, and context information of the behavior information of the recruiting user corresponding to the position.

[0074] Exemplarily, basic user information may include static features of job seekers, such as name, gender, etc., and dynamic features, such as age, contact information, permanent address, etc.; access behavior information, such as job viewing behavior, resume submission behavior, etc.; contextual information of access behavior information, such as login time of the recruitment platform, access IP address and other related information.

[0075] Exemplarily, the first user information may be as shown in Table 2.

[0076] Table 2

[0077]

[0078] The second user information also includes: basic information about the position, such as the time when the recruiting user posted the position; behavioral information about the recruiting user corresponding to the position, such as resume marking behavior, favorite resumes, and replied resumes, and may also include statistical behavioral information, such as the number of resumes viewed, etc.; contextual information about the behavioral information of the recruiting user corresponding to the position, such as the login time of the recruiting user to the recruitment platform, access IP address, and other related information.

[0079] Exemplarily, the second user information may be as shown in Table 3.

[0080] Table 3

[0081]

[0082] Based on this, the above information can be included in the generated training samples, so that the model can more accurately understand the skills, experience and preferences of job seekers, as well as the requirements and characteristics of positions. Based on this, the trained model can recommend positions that are more suitable for job seekers' backgrounds and interests, and recommend job seekers who are more suitable for job requirements to recruiters, thereby improving the accuracy of finding positions and job seekers' respective intentions, and improving user experience and satisfaction.

[0083] Based on the information fusion model training method provided in the embodiment of the present application, a target information fusion model is obtained by training. The embodiment of the present application also provides a method for matching positions with resumes. Specifically, Figure 3 This is a flowchart of a method for matching positions and resumes provided in an embodiment of the present application, combined with Figure 3 As shown, the method for matching positions with resumes may include the following steps 301 to 303.

[0084] Step 301, obtaining multiple job seekers to be matched and user information of each job seeker, as well as job information corresponding to multiple jobs, wherein the user information includes basic user information and resume information;

[0085] Step 302, inputting the user information corresponding to the multiple job seekers and the position information corresponding to the multiple positions into the information fusion model, and outputting the delivery rate between each job seeker and each position, and the positive feedback rate between the recruiting user corresponding to each position and the job seeker through the information fusion model.

[0086] Step 303, determining the matching degree between each job seeker and each position according to the delivery rate between each resume corresponding to the job seeker and each position, and the positive feedback rate between each position corresponding to the recruiting user and the resume.

[0087] The above steps are described in detail below in conjunction with specific embodiments, as shown below.

[0088] Specifically, when it is necessary to recommend a job position to a job seeker, or when it is necessary to recommend a resume to a job seeker, a job position and resume matching method may be implemented to provide recommendation information for both parties.

[0089] In step 301, the multiple job seekers to be matched may be job seekers who are active on the recruitment platform within a period of time, and the positions to be matched may be positions posted on the recruitment platform within a period of time. Optionally, the period of time may be a week or other preset time periods.

[0090] Among them, the actions of job seekers in the recruitment software, such as viewing positions and submitting resumes, and the actions of recruiters in posting resumes, viewing resumes for a certain position, communicating with job seekers, and sending interview invitations, can all be recorded in the service equipment of the recruitment platform. Based on this, it is convenient to obtain multiple job seekers to be matched and the user information of each job seeker, as well as the position information corresponding to multiple positions in the recruitment platform.

[0091] Next, in step 302 and step 303, the user information corresponding to the multiple job seekers and the position information corresponding to the multiple positions are input into the information fusion model, and the information fusion model outputs the delivery rate between each job seeker and each position, as well as the positive feedback rate between the recruiting user corresponding to each position and the job seeker.

[0092] According to the embodiments of the present application, by determining the application rate of job seekers for a position, it is possible to understand the interest and activity of job seekers in the position. Therefore, by referring to the application rate, it is helpful to find the job search preferences of job seekers. By determining the positive feedback rate between the recruiting users and resumes corresponding to the position, it is possible to understand which resumes are more attractive to the recruiting users of the position. Therefore, by referring to the positive feedback rate, it is helpful to find job seekers who are more in line with the recruiting users' wishes.

[0093] After determining the delivery rate between each job seeker and each position, and the positive feedback rate between each position and the recruiter, the matching degree between each job seeker and each position can be further calculated.

[0094] In some optional embodiments, the delivery rate corresponds to a first weight parameter, and the positive feedback rate corresponds to a second weight parameter; the matching degree between each job seeker and each position is determined based on the delivery rate between each resume and each position, and the positive feedback rate between each position and the recruiting user and the resume. Specifically, the matching degree between each job seeker and each position can be determined based on the delivery rate and the first weight parameter between each resume and each position, and the positive feedback rate and the second weight parameter between each position and the recruiting user and the resume.

[0095] Exemplarily, the calculation process may be as shown in formula (1).

[0096] score=ctr α1 ×cvr α2 (1)

[0097] Among them, α1 is the first weight parameter, α2 is the second weight parameter, ctr is the delivery rate, and cvr is the positive feedback rate. Among them, α1 and α2 can be preset, and score represents the matching degree between the job seeker and the position. The higher the score value, the higher the matching degree.

[0098] According to the embodiment of the present application, by calculating the matching degree, it is possible to help job seekers find positions that are more suitable for their background and interests, and it is also possible to help recruiters find job seekers that are more suitable for the job requirements. Afterwards, the matching degree is referred to to recommend information to both job seekers and recruiters, so as to promote a handshake between the two parties and improve recruitment efficiency.

[0099] In some embodiments of the present application, information recommendation for both job seekers and recruiters can be specifically performed by referring to the following steps: in response to a target object visiting a recruitment platform, determining recommendation information to the target object based on a degree of match; wherein, when the target object is a job seeker, the recommended information includes the top M positions with the highest degree of match with the job seeker; and when the target object is a recruiter, the recommended information includes the top N job seekers with the highest degree of match with the corresponding recruitment position.

[0100] For example, Table 4 is

[0101] Table 4

[0102]

[0103]

[0104] According to the embodiments of the present application, the intention information between job seekers and recruiters can be mined. In this way, the intention information of both parties is used as a reference factor for matching calculations, and the matching objects provided to both job seekers and recruiters can be more in line with the wishes of both parties, which is conducive to promoting a handshake between the two parties and improving recruitment efficiency.

[0105] In order to more clearly introduce the training method of the information fusion model and the method of matching positions and resumes provided by the present application, a specific embodiment is introduced below.

[0106] Specifically, in the training method of the information fusion model, the user information of the job seeker and the user information of the recruiter are obtained, as shown in Table 2 and Table 3 respectively. By obtaining this information, training samples can be generated, and the sample label corresponding to each training sample can be determined.

[0107] Construct an information fusion model. Optionally, the information fusion model can be a (Multi-gate Mixture-of-Experts, MMOE) model, or other DNN networks can be used. No specific restrictions are made on the specific information fusion model.

[0108] The following is an introduction to the information fusion model built on the basis of the MMOE model. Figure 4 As shown in FIG, the information fusion model constructed based on the MMOE model includes: input layer, shared bottom layer, expert layers, gate layers, combination of expert outputs, and task-specific top layers.

[0109] The input layer inputs features into the shared bottom layer, from which high-level feature representations are extracted. These feature representations are sent to all expert layers, and each expert layer processes the input independently. The gating layer for each task calculates the weight of each expert layer based on the input features. The output of the expert layer is weighted averaged according to the weights given by the gating layer to obtain the feature representation for the current task. Finally, the feature representation of each task enters the task-specific top layer to generate the final prediction result.

[0110] Specifically, combined Figure 4 As shown in the figure, the input layer contains all the input features of the model. For example, the user information of job seekers and the user information of recruiters. Input features can be numerical, categorical, or text data.

[0111] The shared bottom layer can include one or more fully connected (Dense) layers, whose purpose is to extract high-level abstract features from the original input features. These features are the basic representations shared by all tasks and help capture common information between different tasks.

[0112] The expert layer is a series of small fully connected networks, combining Figure 4 As shown in Figure 1, the expert layer can include multiple small fully connected networks, each of which is responsible for capturing different aspects of the input data. Each expert layer can be regarded as a sub-model specialized for a certain type of feature or pattern. The output of the expert layer can be used by different tasks, but which expert layer's output is more important for which task is determined by the subsequent gating layer.

[0113] Each task has a corresponding gating layer, which can assign weights to the output of each expert layer. The gating layer dynamically adjusts the weights by calculating the contribution of each expert layer to the current task. Specifically, the gating layer uses a softmax function to generate a probability distribution, which determines the importance of the output of each expert layer in the current task. In an embodiment of the present application, it may include the task of determining the delivery rate between each job seeker and each position, and the task of determining the positive feedback rate between the recruiting user and the job seeker corresponding to each position.

[0114] The comprehensive expert output layer can perform a weighted average of the outputs of all expert layers according to the weights calculated by the gating layer of each task. The result is a vector that integrates the knowledge of multiple experts and contains the best feature representation for the current task.

[0115] Each task has one or more task-specific top layers, combined with Figure 4 As shown in the figure, Tower A and Tower B. These layers are used to further process the results obtained from the comprehensive expert output and finally generate the predicted value of the task. This part can include additional fully connected layers or other types of layers, depending on the requirements of the task. Ultimately, Output A can represent the delivery rate between each job seeker and each position, and Output B can represent the positive feedback rate between the recruiter and the job seeker for each position.

[0116] In an embodiment of the present application, the training samples and the sample labels corresponding to the training samples are input into a pre-built information fusion model, and each record in the sample table is input into the constructed model. Since the model can only recognize data, after entering the model, all features are first converted into data through various methods through the shared bottom layer. Then the extracted features (embedding) are sent to each expert in the expert layer, and each expert will capture information from different aspects of the input data. At the same time, the abstracted features (embedding) are also input into the gated layer to calculate the weights of different experts in each task. Each task takes its own weight value and the output value of each expert as a weighted average to obtain the input of the task-specific top layer. After passing through the task-specific top layer, the final score of each task is obtained, and then fused together to obtain the final total target value. According to the difference between the target value and the actual value, each parameter in the model is reversely iterated until the model converges and reaches the preset training conditions, the training is stopped, and the target information fusion model is obtained.

[0117] Afterwards, in the method of matching positions with resumes, the position information and the user information of the job seekers can be obtained for a certain position and a group of job seekers; the position information and the user information of the job seekers can also be obtained for a group of positions and a group of job seekers.

[0118] By inputting these job information and user information into the target information fusion model, we can obtain the delivery rate between each resume and each job seeker, and the positive feedback rate between each job seeker and the resume, and then determine the matching degree between each job seeker and each job.

[0119] Corresponding to the method embodiment of the present application, the present application also provides a training device for an information fusion model, such as Figure 5 As shown, the information fusion model training device 500 includes a first acquisition module 501 and a first processing module 502 .

[0120] The first acquisition module 501 is used to obtain first user information corresponding to multiple job seekers and second user information corresponding to multiple recruiting users, wherein the first user information includes job viewing behavior and resume submission behavior, and the second user information includes resume feedback behavior.

[0121] The first processing module 502 is used to determine the job-seeking relationship between each job-seeking user and each position, and the feedback relationship between the recruiting user corresponding to each position and each job-seeking user according to the multiple first user information and the multiple second user information.

[0122] The first processing module 502 is also used to generate a training sample set based on multiple first user information, multiple second user information, the job-seeking relationship between each job-seeking user and the position, and the feedback relationship between the recruiting user corresponding to each position and each job-seeking user, wherein the training sample set includes multiple training samples and sample labels corresponding to each training sample, and the sample labels include job-seeking relationship labels and feedback labels, wherein the job-seeking relationship label is a submission label corresponding to the job-seeking user who has submitted a resume, or a non-submission label corresponding to the job-seeking user who has not submitted a resume, and the feedback label is a positive feedback label or a negative feedback label, and the positive feedback label indicates that the recruiting user corresponding to the position has marked the resume as suitable, and the negative feedback label indicates that the recruiting user corresponding to the position has not marked the resume as suitable.

[0123] The first processing module 502 is also used to train the pre-constructed information fusion model according to the training sample set, and the output result of the information fusion model is obtained until the preset training conditions are met to obtain the target information fusion model, wherein the output result of the information fusion model includes the delivery rate between the job seeker corresponding to the resume and the position and the positive feedback rate between the job seeker corresponding to the position and the resume, and the preset training conditions are that the error of the delivery rate output by the model is less than the first threshold, and the positive feedback rate output by the model is less than the second threshold.

[0124] In some embodiments of the present application, optionally, the first processing module 502 is also used to determine a set of job seekers whose relationships are to be confirmed and a set of positions whose relationships are to be confirmed based on multiple first user information and multiple second user information, the job seeker set including multiple job seekers and resumes corresponding one-to-one to the job seekers, and the position set including multiple recruiting users and positions corresponding one-to-one to the recruiting users.

[0125] The first processing module 502 is also used to determine, for each job seeker, based on the first user information, a position having a first relationship with the job seeker's corresponding resume, a position having a second relationship with the job seeker's corresponding resume, and a position having a third relationship with the job seeker's corresponding resume.

[0126] The first processing module 502 is further configured to determine, for each recruiting user, a resume having a fourth relationship with the position corresponding to the recruiting user and a resume having a fifth relationship with the position corresponding to the recruiting user according to the second user information.

[0127] The first relationship is used to indicate that the relationship between the resume of the job seeker and the position is a viewed-undelivered relationship.

[0128] The second relationship is used to indicate that the relationship between the resume of the job seeker and the position is a view and submit relationship.

[0129] The third relationship is used to indicate that the relationship between the job seeker's corresponding resume and the position is an unreviewed and unsubmitted relationship.

[0130] The fourth relationship is used to indicate that there is a positive feedback relationship between the corresponding position of the recruiting user and the resume.

[0131] The fifth relationship is used to indicate that there is a negative feedback relationship between the corresponding position of the recruiting user and the resume.

[0132] In some embodiments of the present application, each first user information may optionally include: basic user information, access behavior information, and context information of the access behavior information. Each second user information may also include: basic position information, behavior information of the recruiting user corresponding to the position, and context information of the behavior information of the recruiting user corresponding to the position.

[0133] It can be understood that the training device of the information fusion model in the embodiment of the present application can correspond to the execution entity of the training method of the information fusion model provided in the embodiment of the present application. The specific details of the operation and / or function of each module / unit of the model generation device can refer to the description of the corresponding parts of the training method of the information fusion model provided in the above-mentioned embodiment of the present application. For the sake of brevity, they will not be repeated here.

[0134] Corresponding to the method embodiment of the present application, the present application also provides a training device for an information fusion model, such as Figure 6 As shown, the information fusion model training device 600 includes a second acquisition module 601 and a second processing module 602 .

[0135] The second acquisition module 601 is used to acquire multiple job seekers to be matched and user information of each job seeker, as well as position information corresponding to multiple positions, wherein the user information includes basic user information and resume information.

[0136] The second processing module 602 is used to input user information corresponding to multiple job seekers and position information corresponding to multiple positions into an information fusion model, and output the delivery rate between each job seeker and each position, as well as the positive feedback rate between the recruiting user corresponding to each position and the job seeker through the information fusion model, wherein the target information fusion model is trained based on the training method of the information fusion model provided in the embodiment of the present application.

[0137] The second processing module 602 is further used to determine the matching degree between each job seeker and each position according to the delivery rate between each resume corresponding to the job seeker and each position, and the positive feedback rate between each position corresponding to the recruiting user and the resume.

[0138] In some embodiments of the present application, optionally, the delivery rate corresponds to a first weight parameter, and the positive feedback rate corresponds to a second weight parameter.

[0139] The second processing module 602 is also used to determine the matching degree between each job seeker and each position based on the delivery rate and the first weight parameter between each resume corresponding to the job seeker and each position, and the positive feedback rate and the second weight parameter between each position corresponding to the recruiting user and the resume.

[0140] In some embodiments of the present application, optionally, the second processing module 602 is further configured to determine recommendation information to the target object according to the matching degree in response to the target object visiting the recruitment platform. Wherein, when the target object is a job seeker, the recommendation information includes the top M positions with the highest matching degree with the job seeker. When the target object is a recruiting user, the recommendation information includes the top N job seekers with the highest matching degree with the corresponding recruitment position.

[0141] It can be understood that the position and resume matching device of the embodiment of the present application can correspond to the execution entity of the position and resume matching method provided in the embodiment of the present application. The specific details of the operation and / or function of each module / unit of the position and resume matching device can be found in the description of the corresponding parts of the position and resume matching method provided in the above-mentioned embodiment of the present application. For the sake of brevity, they will not be repeated here.

[0142] The electronic device in the embodiment of the present application may be a user terminal device, a server, other computing devices, or a cloud server. Figure 7 A schematic diagram of the hardware structure of an electronic device according to an embodiment of the present application is shown. The electronic device may include a processor 701 and a memory 702 storing computer program instructions. When the processor 701 executes the computer program instructions, the process or function of any of the above-mentioned embodiments of the method is implemented.

[0143] Specifically, the processor 701 may include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application. The memory 702 may include a large-capacity memory for data or instructions. For example, the memory 702 may be at least one of the following: a hard disk drive (HDD), a read-only memory (ROM), a random access memory (RAM), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a tape, a universal serial bus (USB) drive or other physical / tangible memory storage device. For another example, the memory 702 may include a removable or non-removable (or fixed) medium. For another example, the memory 702 may be inside or outside the integrated gateway disaster recovery device. The memory 702 may be a non-volatile solid-state memory. In other words, the memory 702 usually includes a tangible (non-transitory) computer-readable storage medium (such as a memory device) encoded with computer-executable instructions, and when the software is executed (such as executed by one or more processors), the operations described in the method of the embodiment of the present application can be performed. The processor 701 implements the process or function of any method in the above embodiments by reading and executing the computer program instructions stored in the memory 702.

[0144] In one example, Figure 7 The electronic device shown may also include a communication interface 703 and a bus 710. Among them, the processor 701, the memory 702, and the communication interface 703 are connected through the bus 710 and complete the communication between each other. The communication interface 703 is mainly used to realize the communication between each module, device, unit and / or device in the embodiment of the present application. The bus 710 includes hardware, software or both, and can couple the components of the online data traffic billing device to each other. For example, the bus may include at least one of the following: an accelerated graphics port (AGP) or other graphics bus, an enhanced industrial standard architecture (EISA) bus, a front-end bus (FSB), a hypertransport (HT) interconnect, an industrial standard architecture (ISA) bus, an infinite bandwidth interconnect, a low pin count (LPC) bus, a memory bus, a micro channel architecture (MCA) bus, a peripheral component interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a serial advanced technology attachment (SATA) bus, a video electronics standard association local (VLB) bus or other suitable buses. The bus 710 may include one or more buses. Although the embodiments of the present application describe or illustrate a specific bus, the embodiments of the present application may consider any suitable bus or interconnection method.

[0145] In combination with the method in the above embodiments, an embodiment of the present application also provides a computer-readable storage medium, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the process or function of any method in the above embodiments is implemented.

[0146] In addition, an embodiment of the present application further provides a computer program product, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the process or function of any one of the methods in the above embodiments is implemented.

[0147] The above exemplarily describes the flowcharts and / or block diagrams of the methods, devices, systems and computer program products of the embodiments of the present application, and describes the relevant various aspects. It should be understood that each box or combination thereof in the flowchart and / or block diagram can be implemented by computer program instructions, or by dedicated hardware that performs a specified function or action, or by a combination of dedicated hardware and computer instructions. For example, these computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to form a machine that enables these instructions executed by such a processor to enable the implementation of the functions / actions specified in each box or combination thereof in the flowchart and / or block diagram. Such a processor can be a general-purpose processor, a special-purpose processor, a special application processor or a field programmable logic circuit.

[0148] The functional blocks shown in the structured block diagram of the embodiment of the present application can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application-specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc.; when implemented in software, it is a program or code segment used to perform the required task. The program or code segment can be stored in a memory, or transmitted on a transmission medium or a communication link via a data signal carried in a carrier. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.

[0149] It should be noted that the present application is not limited to the specific configurations and processes described above or shown in the figures. The above is only a specific implementation mode of the present application. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working process of the described system, device, module or unit can refer to the corresponding process in the method embodiment without further description. It should be understood that the scope of protection of the present application is not limited to this. Any technician familiar with the technical field can think of various equivalent modifications or substitutions within the technical scope disclosed in the present application, and these modifications or substitutions should be included in the scope of protection of the present application.

Claims

1. A training method for an information fusion model, characterized in that: The method comprises: Acquire first user information corresponding to a plurality of job seekers and second user information corresponding to a plurality of recruiters, wherein the first user information includes job search behavior and resume submission behavior, and the second user information includes resume feedback behavior; Determine the job search relationship between each job seeker and each position, and the feedback relationship between the recruiting user corresponding to each position and each job seeker based on the plurality of first user information and the plurality of second user information; Generate a training sample set based on a plurality of first user information, a plurality of second user information, a job-seeking relationship between each job-seeking user and a position, and a feedback relationship between a recruiting user corresponding to each position and each job-seeking user, wherein the training sample set includes a plurality of training samples and a sample label corresponding to each training sample, and the sample label includes a job-seeking relationship label and a feedback label, wherein the job-seeking relationship label is a submitted label corresponding to a job-seeking user who has submitted a resume, or a non-submitted label corresponding to a job-seeking user who has not submitted a resume, and the feedback label is a positive feedback label or a negative feedback label, wherein the positive feedback label indicates that the recruiting user corresponding to the position has marked the resume as suitable, and the negative feedback label indicates that the recruiting user corresponding to the position has not marked the resume as suitable; The pre-constructed information fusion model is trained according to the training sample set, and the output result of the information fusion model is trained until the preset training conditions are met to obtain a target information fusion model, wherein the output result of the information fusion model includes the delivery rate between the job seekers corresponding to the resumes and the positions and the positive feedback rate between the job seekers corresponding to the positions and the resumes, and the preset training conditions are that the error of the delivery rate output by the model is less than a first threshold, and the positive feedback rate output by the model is less than a second threshold.

2. The method according to claim 1, characterized in that The determining, based on the plurality of first user information and the plurality of second user information, the job search relationship between each job seeker and each position, and the feedback relationship between the recruiting user corresponding to each position and each job seeker, includes: Determine a set of job seekers whose relationships are to be confirmed and a set of positions whose relationships are to be confirmed based on the plurality of first user information and the plurality of second user information, wherein the set of job seekers includes a plurality of job seekers and resumes corresponding to the job seekers one by one, and the set of positions includes a plurality of recruiters and positions corresponding to the recruiters one by one; For each job seeker, according to the first user information, determine a position having a first relationship with the resume corresponding to the job seeker, a position having a second relationship with the resume corresponding to the job seeker, and a position having a third relationship with the resume corresponding to the job seeker; For each recruiting user, determining, based on the second user information, a resume having a fourth relationship with the position corresponding to the recruiting user, and a resume having a fifth relationship with the position corresponding to the recruiting user; The first relationship is used to indicate that the relationship between the resume of the job seeker and the position is a viewing undelivered relationship; The second relationship is used to indicate that the relationship between the resume of the job seeker and the position is a view and submit relationship; The third relationship is used to indicate that the relationship between the corresponding resume of the job seeker and the position is not reviewed and not submitted; The fourth relationship is used to indicate that there is a positive feedback relationship between the corresponding position of the recruiting user and the resume; The fifth relationship is used to indicate that there is a negative feedback relationship between the corresponding position of the recruiting user and the resume.

3. The method according to claim 1, characterized in that Each first user information also includes: basic user information, access behavior information, and context information of the access behavior information; Each second user information also includes: basic information of the position, behavior information of the recruiting user corresponding to the position, and context information of the behavior information of the recruiting user corresponding to the position.

4. A method for matching positions with resumes, characterized in that: The method comprises: Acquire multiple job seekers to be matched and user information of each job seeker, as well as position information corresponding to multiple positions, wherein the user information includes basic user information and resume information; Inputting the user information corresponding to the plurality of job seekers and the position information corresponding to the plurality of positions into an information fusion model, and outputting the delivery rate between each job seeker and each position, and the positive feedback rate between the recruiting user corresponding to each position and the job seeker through the information fusion model, wherein the target information fusion model is trained based on the information fusion model training method according to any one of claims 1 to 3; The matching degree between each job seeker and each position is determined based on the submission rate between each resume and each position, and the positive feedback rate between each position and the recruiting user and the resume.

5. The method according to claim 4, characterized in that The first weight parameter corresponding to the delivery rate, and the second weight parameter corresponding to the positive feedback rate; Determining the matching degree between each job seeker and each position according to the delivery rate between each resume and each position, and the positive feedback rate between each position and the recruiting user and the resume, includes: The matching degree between each job seeker and each position is determined based on the submission rate between each resume and each position and the first weight parameter, as well as the positive feedback rate between each position and the recruiting user and the resume and the second weight parameter.

6. The method according to claim 4, characterized in that The method further comprises: In response to a target object visiting a recruitment platform, determining recommendation information to the target object according to a matching degree; Wherein, when the target object is a job seeker, the recommended information includes the top M positions with the highest matching degree with the job seeker; When the target object is a recruiting user, the recommendation information includes the top N job seekers with the highest matching degree with the corresponding position of the recruitment.

7. A training device for an information fusion model, characterized in that: include: A first acquisition module is used to acquire first user information corresponding to a plurality of job seekers and second user information corresponding to a plurality of recruiters, wherein the first user information includes job search behavior and resume submission behavior, and the second user information includes resume feedback behavior; A first processing module, configured to determine, based on the plurality of first user information and the plurality of second user information, a job search relationship between each job seeker and each position, and a feedback relationship between a recruiting user corresponding to each position and each job seeker; The first processing module is further used to generate a training sample set based on the first user information, the second user information, the job-seeking relationship between each job-seeking user and the position, and the feedback relationship between the recruiting user corresponding to each position and each job-seeking user, wherein the training sample set includes multiple training samples and sample labels corresponding to each training sample, and the sample labels include job-seeking relationship labels and feedback labels, wherein the job-seeking relationship label is a delivery label corresponding to a job-seeking user who has delivered a resume, or a non-delivery label corresponding to a job-seeking user who has not delivered a resume, and the feedback label is a positive feedback label or a negative feedback label, wherein the positive feedback label indicates that the recruiting user corresponding to the position has marked the resume as suitable, and the negative feedback label indicates that the recruiting user corresponding to the position has not marked the resume as suitable; The first processing module is also used to train a pre-constructed information fusion model based on the training sample set, and the output result of the information fusion model is obtained until a preset training condition is met to obtain a target information fusion model, wherein the output result of the information fusion model includes a delivery rate between job seekers corresponding to resumes and positions and a positive feedback rate between job seekers corresponding to positions and resumes, and the preset training condition is that the error of the delivery rate output by the model is less than a first threshold, and the positive feedback rate output by the model is less than a second threshold.

8. A device for matching positions with resumes, characterized in that: include: A second acquisition module is used to acquire multiple job seekers to be matched and user information of each job seeker, as well as position information corresponding to multiple positions, wherein the user information includes basic user information and resume information; a second processing module, for inputting the user information corresponding to the plurality of job seekers and the position information corresponding to the plurality of positions into an information fusion model, and outputting the delivery rate between each job seeker and each position, and the positive feedback rate between the recruiting user corresponding to each position and the job seeker through the information fusion model, wherein the target information fusion model is trained based on the information fusion model training method according to any one of claims 1 to 3; The second processing module is further used to determine the matching degree between each job seeker and each position based on the delivery rate between each resume corresponding to the job seeker and each position, and the positive feedback rate between each position corresponding to the recruiting user and the resume.

9. An electronic device, characterized in that: The electronic device comprises: a processor and a memory storing computer program instructions; when the electronic device executes the computer program instructions, the method according to any one of claims 1 to 6 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.

11. A computer program product, characterized in that It comprises computer program instructions, which implement the method according to any one of claims 1 to 6 when executed by a processor.

Images