Recruitment service resource recommendation method, device and equipment based on recruitment scenario

Abstract

The application discloses a recruitment service resource recommendation method, device and equipment based on a recruitment scene. The method comprises the following steps: obtaining recruitment data of each position published in history; training a pre-training model based on the recruitment data under a single evaluation dimension to obtain a recommendation model associated with the single evaluation dimension; inputting position description information of a current position into N recommendation models respectively, so that each recommendation model matches a first resource ID for the current position from each resource ID under the single evaluation dimension based on the position description information, evaluates a service effect of a recruitment service resource of the first resource ID on the current position, and obtains a service effect index value; taking the first resource ID and the service effect index value thereof as a recommendation result to obtain N recommendation results; screening a second resource ID in combination with the service effect index values in the N recommendation results, and determining a recruitment service resource corresponding to the second resource ID as a recommended recruitment service resource of the current position.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet, in particular to the technical field of Internet data processing, and particularly relates to a recruitment service resource recommendation method, device and equipment based on a recruitment scenario. BACKGROUND

[0002] In the current job recruitment field, various job recruitment platforms have become an important channel for enterprises to find suitable talents. In order to improve the recruitment efficiency, these platforms have launched various recruitment service products, aiming to help recruiters more effectively lock target candidates, shorten the recruitment cycle, and meet the needs of recruiters at different stages.

[0003] However, in actual use, recruiters often face the problem of choice difficulty. When in urgent need of recruitment personnel, they hope to accelerate the recruitment process by subscribing to recruitment service products, but when faced with a variety of recruitment service products on the platform, it is difficult to judge which products are most suitable for their needs. This uncertainty not only increases the decision-making cost of recruiters, but also may affect the recruitment effect, leading to an extended recruitment cycle or recruitment of unsuitable talents. SUMMARY

[0004] Therefore, the embodiments of the present application provide a recruitment service resource recommendation method, device, equipment, medium and product based on a recruitment scenario, which can recommend recruitment service products meeting the position requirements for each job position.

[0005] In a first aspect, the embodiments of the present application provide a recruitment service resource recommendation method based on a recruitment scenario, which comprises: obtaining recruitment data of each position published historically, wherein the recruitment data comprises position description information and recruitment service resource data corresponding to each position, the recruitment service resource data comprises a resource ID associated with the position, and service effect data and service evaluation data after subscribing to the recruitment service resource corresponding to the resource ID for the position; dividing the recruitment data of each position into N evaluation dimensions, and training a pre-trained model based on the recruitment data under a single evaluation dimension to obtain a recommendation model associated with the single evaluation dimension, wherein the recommendation model is used to match recruitment service resources for the position under the single evaluation dimension; in the case of obtaining the position description information of the current position, inputting the position description information of the current position into the N recommendation models respectively, so that each recommendation model matches a first resource ID for the current position from each resource ID based on the position description information under the single evaluation dimension, and evaluates the service effect of the recruitment service resource of the first resource ID on the current position to obtain a service effect index value; taking the first resource ID and the service effect index value output by each recommendation model as a recommendation result to obtain N recommendation results output by the N recommendation models; combining the service effect index values in the N recommendation results, screening a second resource ID from the N first resource IDs, and determining the recruitment service resource corresponding to the second resource ID as the recommended recruitment service resource of the current position.

[0006] In a second aspect, an embodiment of the present application provides a recruitment service resource recommendation device based on a recruitment scenario, which comprises: an acquisition module configured to acquire recruitment data of each position published historically, wherein the recruitment data comprises position description information corresponding to each position and recruitment service resource data, the recruitment service resource data comprises a resource ID associated with the position, and service effect data and service evaluation data after the resource ID corresponding to the recruitment service resource is subscribed for the position; a training module configured to divide the recruitment data of each position into N evaluation dimensions, and train a pre-trained model based on the recruitment data under a single evaluation dimension to obtain a recommendation model associated with the single evaluation dimension, wherein the recommendation model is configured to match the recruitment service resource for the position under the single evaluation dimension; an evaluation module configured to, when the position description information of a current position is acquired, input the position description information of the current position into the N recommendation models respectively, so that each recommendation model matches a first resource ID for the current position from each resource ID based on the position description information under the single evaluation dimension, and evaluates a service effect of the recruitment service resource of the first resource ID on the current position to obtain a service effect index value; an output module configured to take the first resource ID and the service effect index value output by each recommendation model as a recommendation result to obtain N recommendation results output by the N recommendation models; and a recommendation module configured to combine the service effect index values in the N recommendation results, filter a second resource ID from the N first resource IDs, and determine a recruitment service resource corresponding to the second resource ID as a recommended recruitment service resource for the current position.

[0007] In a third aspect, an embodiment of the present application provides an electronic device, which comprises a processor and a memory storing computer program instructions; and the processor implements the steps of the recruitment service resource recommendation method based on a recruitment scenario according to the first aspect when executing the computer program instructions.

[0008] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, which stores computer program instructions; and the computer program instructions are executed by a processor to implement the steps of the recruitment service resource recommendation method based on a recruitment scenario according to the first aspect.

[0009] In a fifth aspect, an embodiment of the present application provides a computer program product stored in a nonvolatile storage medium, which is executed by a processor to implement the steps of the recruitment service resource recommendation method based on a recruitment scenario according to the first aspect.

[0010] In a sixth aspect, an embodiment of the present application provides a chip, which comprises a processor and a communication interface, the communication interface and the processor are coupled, and the processor is configured to run a program or instructions to implement the steps of the recruitment service resource recommendation method based on a recruitment scenario according to the first aspect.

[0011] The application provides a recruitment service resource recommendation method, device, equipment, medium and product based on a recruitment scenario. Considering that different positions are suitable for subscribing to different recruitment service resources, the application obtains recruitment data of each position published in the past, which includes position description information and recruitment service resource data corresponding to each position. The recruitment service resource data includes a resource ID associated with the position, service effect data and service evaluation data after the position subscribes to the recruitment service resource corresponding to the resource ID. The recruitment data of each position is divided into N evaluation dimensions, and a pre-trained model is trained based on the recruitment data under a single evaluation dimension to obtain a recommendation model associated with the single evaluation dimension. Since the recruitment data under a single evaluation dimension can reflect the preferences and service effects of different types of positions for subscribing to recruitment service resources under the same evaluation dimension, after the recommendation model is obtained by training the model based on the recruitment data under a single evaluation dimension, the position description information of the current position is input into the recommendation model. The recommendation model analyzes historical subscription behavior, use effect and feedback information of the position and recruitment service resources, and mines the correlation between different recruitment service resources and positions, thereby providing personalized recommendation solutions for positions to be subscribed to services. Specifically, by analyzing the position description information of the current position, positions similar to the position characteristics of the current position are obtained, and the subscription of recruitment service resources by similar positions is combined to match a first resource ID for the current position from each resource ID. In this way, a matching recruitment service resource is selected for the current position for each evaluation dimension, completing the preliminary screening under a single evaluation dimension. On this basis, in combination with N service effect index values corresponding to N first resource IDs, the service effects of the recruitment service resources corresponding to the N first resource IDs on the current position can be accurately predicted, and a second resource ID is further selected in combination with the predicted service effects, and the recruitment service resource corresponding to the second resource ID is used as a recommended recruitment service resource, which can accurately select a recruitment service resource with better service effect for the current position for recommendation, thereby assisting recruiters to quickly locate and select recruitment service resources with higher adaptability and better expected effects, providing more accurate and personalized recruitment service recommendation solutions for recruiters, shortening the recruitment period, and improving recruitment efficiency and effectiveness. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following briefly introduces the drawings in the embodiments of the application.

[0013] Figure 1 is a flowchart of a recruitment service resource recommendation method based on a recruitment scenario provided by an embodiment of the application;

[0014] Figure 2 is a flowchart of a recruitment service resource recommendation method based on a recruitment scenario provided by another embodiment of the application;

[0015] Figure 3 FIG. 1 is a schematic diagram of a hardware structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0016] The principles and spirits of the present application will be described below with reference to a number of exemplary embodiments. It should be understood that the purpose of providing these embodiments is to make the principles and spirits of the present application clearer and more thorough, and to enable those skilled in the art to better understand and implement the principles and spirits of the present application. The exemplary embodiments provided herein are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments herein, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present application.

[0017] In this document, terms such as first, second, third, etc. are used to distinguish one entity (or operation) from another, and do not require or imply any order or association between the entities (or operations).

[0018] The services provided by the existing recruitment platform for job seekers are called C-end services, which include but are not limited to: a job seeker can create his / her own resume on the platform, including personal information, education experience, skills and work experience, etc. in the resume; the job seeker can also search and view the positions according to various set conditions through the search engine provided by the platform; the job seeker can deliver the resume in the position information page and communicate online with the recruiter when the recruiter is online, etc. The services provided by the existing recruitment platform for recruiters are called B-end services. B-end services include but are not limited to: a recruiter can create and publish position information on the platform, including enterprise information, job content, various requirements or conditions for job seekers, salary and welfare, etc.; the recruiter can also search, view and download resumes through the search engine provided by the platform according to various set conditions; the recruiter can respond to the online communication request of the job seeker and communicate online with the job seeker when the job seeker views the position information page published by the recruiter, etc.

[0019] In the related art, in order to improve the recruitment efficiency, various recruitment service products have been launched by recruitment platforms, aiming to help recruiters more effectively lock target candidates, shorten the recruitment cycle and meet the needs of recruiters at different stages.

[0020] The recruitment service resource recommendation method based on the recruitment scenario provided by the embodiments of the present application will be described in detail below in combination with the accompanying drawings, specific embodiments and application scenarios.

[0021] Figure 1This is a flow chart of a recruitment service resource recommendation method based on a recruitment scenario provided by an embodiment of the present application. The execution entity of the recruitment service resource recommendation method based on a recruitment scenario may be a recruitment platform.

[0022] The following uses the recruitment platform as an example to illustrate the recruitment service resource recommendation method based on recruitment scenarios. It should be noted that the above-mentioned execution entities and application scenarios do not constitute a limitation of this application.

[0023] like Figure 1 As shown, the recruitment service resource recommendation method based on the recruitment scenario provided by the embodiment of the present application may include steps 110 to 150.

[0024] Step 110: Obtain recruitment data for each position published in the past. The recruitment data includes job description information and recruitment service resource data corresponding to each position. The recruitment service resource data includes the resource ID associated with the position, as well as service effect data and service evaluation data after subscribing to the recruitment service resource corresponding to the resource ID for the position.

[0025] Step 120: Divide the recruitment data for each position into N evaluation dimensions, and train the pre-trained model based on the recruitment data under a single evaluation dimension to obtain a recommendation model associated with the single evaluation dimension;

[0026] Step 130: When the job description information of the current job is obtained, the job description information of the current job is input into each of the N recommendation models, so that each recommendation model matches a first resource ID for the current job from each resource ID based on the job description information under a single evaluation dimension, and evaluates the service effect of the recruitment service resource of the first resource ID on the current job, thereby obtaining a service effect index value.

[0027] Step 140: Taking the first resource ID and service effect index value output by each recommendation model as a recommendation result, and obtaining N recommendation results output by N recommendation models;

[0028] Step 150 , combining the service effect index values ​​in the N recommendation results, screening the second resource ID from the N first resource IDs, and determining the recruitment service resource corresponding to the second resource ID as the recommended recruitment service resource for the current position.

[0029] The method for recommending recruitment service resources based on a recruitment scenario provided in the embodiments of the present application can reflect the preferences and service effects of different types of positions for subscribing to recruitment service resources under the same evaluation dimension due to the recruitment data under a single evaluation dimension. Therefore, after a recommendation model is obtained by training a model based on the recruitment data under a single evaluation dimension, the position description information of a current position is input into the recommendation model. The recommendation model analyzes historical subscription behaviors, use effects, and feedback information of the position and the recruitment service resources, and mines the correlation between different recruitment service resources and positions, so as to provide a personalized recommendation scheme for the position to be subscribed to. Specifically, the position description information of the current position is analyzed to obtain positions similar to the position characteristics of the current position, and the recruitment service resources subscribed to by the similar positions are combined to match a first resource ID for the current position from each resource ID. In this way, a matching recruitment service resource is screened for the current position for each evaluation dimension, and the preliminary screening under a single evaluation dimension is completed. On this basis, the N service effect index values corresponding to the N first resource IDs are combined to accurately predict the service effects of the recruitment service resources corresponding to the N first resource IDs on the current position, and a second resource ID is further screened based on the predicted service effects, and the recruitment service resource corresponding to the second resource ID is taken as a recommended recruitment service resource, so that the recruitment service resource with better service effects can be accurately screened for the current position for recommendation, thereby assisting recruiters to quickly locate and select recruitment service resources with higher adaptability and better expected effects, providing more accurate and personalized recruitment service recommendation schemes for recruiters, shortening the recruitment cycle, and improving the recruitment efficiency and effect.

[0030] The specific implementation of the above steps will be described in detail below in combination with specific embodiments.

[0031] Step 110 is related to obtaining the recruitment data of each position published in the history.

[0032] In step 110, the recruitment data includes the position description information and the recruitment service resource data corresponding to each position. The position description information is used to describe the responsibilities, requirements, working environment, salary, and other key information of the position, aiming to help potential candidates understand the overall situation of the position and determine whether they are suitable for the position. The position description information can include dimension information such as function dimension, industry dimension, region dimension, enterprise size dimension, salary range dimension, and work experience dimension.

[0033] The recruitment service resource data includes a resource ID associated with a position, and service effect data and service evaluation data of the position after subscribing to the recruitment service resource corresponding to the resource ID. The resource ID is a product identifier for uniquely identifying the recruitment service resource. The recruitment service resource is a service resource provided for recruiters. The service effect data is used to represent the service effect of the position after subscribing to the recruitment service within a subscription period. The service evaluation data is evaluation data input by the recruiter for the service effect after the subscription period ends.

[0034] Exemplarily, the recruitment service resource can include an exposure promotion service resource for promoting the exposure of the position, and a delivery number promotion service resource for promoting the number of (received) resume deliveries. The exposure promotion service resource may, for example, be a "super exposure", "guaranteed exposure", "position top", "brand top", and the like, which are related to promoting the exposure of the position. The delivery number promotion service resource may, for example, be an "intelligent invitation", a "guaranteed invitation", and the like, which are related to promoting the number of resume deliveries.

[0035] It should be noted that the above recruitment service resource can be a value-added service product subscribed by the recruiter.

[0036] Involving step 120, the recruitment data of each position is divided into N evaluation dimensions, and a pre-training model is trained based on the recruitment data under a single evaluation dimension to obtain a recommendation model associated with the single evaluation dimension.

[0037] In step 120, a deep learning framework and a pre-training model can be built using the Transformers open source library. The Transformers library is an open source deep learning model library that provides various pre-training models based on the Transformer model structure and supports multiple tasks such as text classification, information extraction, question and answer systems, text summarization, machine translation, and text generation. The pre-training model can also be a neural network model. N is a positive integer. Each position description information can be composed of multiple field data. Each field is an evaluation dimension, such as a function dimension, a recruitment enterprise industry dimension, etc. The recruitment data under a single evaluation dimension can include: recruitment service resource data of each position, and part of the data in the position description information of each position associated with the single evaluation dimension. N pre-training models are trained using recruitment data under N evaluation dimensions, and finally N recommendation models are obtained. The recommendation model is used to match the recruitment service resource for the position under a single evaluation dimension.

[0038] In some embodiments of the present application, the N evaluation dimensions can include a function dimension, an industry dimension, a region dimension, an enterprise size dimension, a salary range dimension, and a work experience dimension. The service effect data and the service effect index value both include the exposure of the position and / or the number of resume deliveries.

[0039] Specifically, for the same recruitment service resource, the fields of the associated service effect data and service effect index values are consistent, for example, for the resource of "super exposure", the field of "exposure degree" is contained in the associated service effect data and service effect index values; for the resource of "intelligent invitation", the field of "resume delivery quantity" is contained in the associated service effect data and service effect index values; for the resource of "job position top", the fields of "exposure degree" and "resume delivery quantity" are contained in the associated service effect data and service effect index values.

[0040] It should be noted that the exposure degree and the resume delivery quantity of the above-mentioned position are all used to represent the exposure degree and the resume delivery quantity generated within the subscription period (or service period) of the recruitment service resource. The N evaluation dimensions can be increased and reduced according to specific needs, for example, the evaluation dimensions of "working time" and "working mode" are added, and the present application does not make specific limitations. The same recruitment service resource is associated with at least one service effect index.

[0041] In some embodiments of the present application, in step 120, the pre-trained model can be trained based on the recruitment data under a single evaluation dimension by using a cross-validation method to obtain a recommendation model associated with the single evaluation dimension.

[0042] Specifically, the recruitment data under a single evaluation dimension is divided into multiple groups to obtain multiple sets of sample data sets, wherein 80% of the multiple sets of sample data sets are used as training data sets to train the pre-trained model, and the remaining 20% are used as verification data sets to evaluate the trained recommendation model. The verification data set is input into the trained recommendation model for prediction, and the predicted result is compared with the actual data to obtain a matching degree. If the matching degree does not meet the preset expected value, the model parameters of the trained recommendation model are fine-tuned until the trained recommendation model under each evaluation dimension is obtained.

[0043] In some embodiments of the present application, in step 120, the Transformers open source library can be used to build a deep learning framework and a pre-trained model, the training data set is trained on the pre-trained model through the training task, and the parameters of the model are optimized by using the multi-card training and asynchronous stochastic gradient descent algorithm. The training data set is trained on the pre-trained model through the training task, the training data set is divided into a plurality of sub-training sample sets, the sub-training sample sets are assigned to different GPU (Graphics Processing Unit) devices for calculation, each GPU device calculates the gradient of the pre-trained model, and immediately updates the parameters of the pre-trained model. At the end of each iteration period, the updated parameters of the pre-trained model of all GPU devices are aggregated on a central node, and the aggregated parameters of the pre-trained model are updated on the central node.

[0044] It should be noted that the recommendation model under each evaluation dimension in the present application can be repeatedly executed step 120 for iterative updating based on the latest obtained recruitment data of the position to optimize the model performance and obtain more accurate recommendation results.

[0045] In some embodiments of the present application, in order to match accurate recruitment service resources for positions under a single evaluation dimension, the recruitment data of each position in the above step 120 can be divided into N evaluation dimensions, which can specifically include:

[0046] The position description information of each position is split into N evaluation dimensions to obtain N position tag information of each position information under N evaluation dimensions.

[0047] For each evaluation dimension, the recruitment service resource data of each position and the position tag information under the corresponding evaluation dimension are divided into the same group to obtain recruitment data under a single evaluation dimension.

[0048] Specifically, the position tag information of each position under a single evaluation dimension is combined with the recruitment service resource data corresponding to the position to obtain a recruitment data under the single evaluation dimension. The position description information is composed of position tag information under a plurality of evaluation dimensions, and the position tag information can be position element information related to the evaluation dimension in the position description information.

[0049] Exemplarily, for the position description information of "Position 1: Sales (function), computer software (industry), 3000 people (enterprise size)", the recruitment service resource data associated with Position 1 is "super exposure, exposure 3W+", and "sales + super exposure, exposure 3W+" can be taken as the recruitment data under the function dimension, "computer software + super exposure, exposure 3W+" can be taken as the recruitment data under the industry dimension, and "3000 people + super exposure, exposure 3W+" can be taken as the recruitment data under the enterprise size dimension.

[0050] In the embodiments of the present application, the recruitment data related to each evaluation dimension can be extracted from all recruitment data, and then the pre-training model can learn the positions with different characteristics under the same evaluation dimension, such as positions with different functions, through the recruitment data related to a single evaluation dimension, for example, which recruitment service resource is subscribed by positions with different functions, or what is the difference in service effect when positions with different functions subscribe to the same recruitment service resource. Based on this, the trained recommendation model can have the recommendation ability of the recruitment service resource, predict the recruitment service resource that the recruiter wants to subscribe for the current position based on the position characteristics of the current position, and recommend a recruitment service resource under each evaluation dimension, so that the recommended recruitment service resource can meet the needs of the recruiter as much as possible.

[0051] In some embodiments of the present application, before the step 120 of dividing the recruitment data of each position into N evaluation dimensions, the following steps can also be included:

[0052] From each position, a target position associated with a third resource ID is screened, wherein the resource type of the recruitment service resource corresponding to the third resource ID is a task completion type;

[0053] By analyzing the service effect data of the target position, it is determined whether the exposure and / or the number of resumes delivered of the target position in the service period of subscribing to the recruitment service resource corresponding to the third resource ID reaches the task completion amount;

[0054] In the case where the task completion amount is not reached, the recruitment data of the target position under a single evaluation dimension is taken as a negative sample to train the pre-training model.

[0055] Specifically, the target position is the position associated with the third resource ID in each position, and the recruitment service resource of the task completion type usually corresponds to the task completion amount, which can include exposure and / or number of resumes delivered, for example, "guaranteed exposure" with exposure 3W+ and "guaranteed invitation" with exposure 3W+.

[0056] In the embodiments of the present application, for a specific type of recruitment service resource that has a task completion amount, if the service effect thereof does not meet the task completion amount, the relevant recruitment data can be determined as a negative sample for training. In this way, when recommending a recruitment service resource for the current position, recruitment service resources whose future service effect may not meet the requirements can be avoided, so that the service effect of the recommended recruitment service resource meets the recruiter's demand as much as possible.

[0057] In some embodiments of the present application, in the case where the exposure and / or the number of resumes delivered of the target position within the service period of the recruitment service resource corresponding to the third resource ID do not reach the task completion amount, the recruitment data of the target position can also be selected to be deleted from the recruitment data of each position. In this way, by screening and filtering the recruitment data, the recruitment data that does not reach the task completion amount, i.e., the service effect does not meet the requirements, is filtered out.

[0058] Regarding step 130, in the case where the position description information of the current position is obtained, the position description information of the current position is input into the N recommendation models respectively, so that each recommendation model matches a first resource ID for the current position from each resource ID based on the position description information under a single evaluation dimension, and evaluates the service effect of the recruitment service resource of the first resource ID on the current position to obtain a service effect index value.

[0059] In step 130, the current position can be a position selected by a recruiter through user input for a recruitment service resource to be subscribed. By inputting the position description information of the current position into the recommendation model, the recommendation model can obtain positions similar to the current position by analyzing the position description information of the current position, and match a suitable first resource ID for the current position from each resource ID by analyzing the recruitment data of the similar positions and combining the subscription situation of the recruitment service resources of the similar positions. Under a single evaluation dimension, the current position and each resource ID correspond to a matching degree, and the first resource ID can be the resource ID with the highest matching degree.

[0060] For example, for the current position JD1: sales (function), fast consumption (industry), A city (work location), 10,000 people (enterprise size), no requirement (work experience), involving 5 evaluation dimensions, a first resource ID can be matched for it under each evaluation dimension.

[0061] Regarding step 140, the first resource ID and the service effect index value output by each recommendation model are taken as a recommendation result, and N recommendation results output by the N recommendation models are obtained.

[0062] In step 140, the service effect index value is used to represent the service effect of the first resource ID corresponding to the recruitment service resource on the current position in the service period, for example, the exposure size and / or the number of resumes delivered to the current position.

[0063] Referring to the above example, the five recommended results are: according to the function, the subscription product "position top" may be recommended, the exposure is 8W+, and the resume delivery is 8+; according to the industry, the subscription product "super exposure" may be recommended, the exposure is 3W+; according to the work location, the subscription product "brand top" may be recommended, the exposure is 10W+; according to the enterprise size, the subscription product "brand top" may be recommended, the exposure is 10W+; according to the work experience, the subscription product "intelligent invitation" may be recommended, and the resume delivery is 10+.

[0064] Involving step 150, the second resource ID is screened from the N first resource IDs in combination with the service effect index values in the N recommended results, and the recruitment service resource corresponding to the second resource ID is determined as the recommended recruitment service resource of the current position.

[0065] In step 150, the recommended recruitment service resource refers to the recruitment service resource recommended for the current position. After determining the recommended recruitment service resource, the recommended result corresponding to the second resource ID can be displayed in a prominent position on the recommendation page, facilitating the recruiter to view. The application can determine the first resource ID with the highest service index value as the second resource ID.

[0066] In some embodiments of the application, the recommended result further includes the matching degree between the current position and the first resource ID. The step 150 of screening the second resource ID from the N first resource IDs in combination with the service effect index values in the N recommended results can specifically include:

[0067] In the case that the N service effect index values in the N recommended results correspond to the same service index, the first resource ID with the highest service effect index value is determined as the second resource ID;

[0068] In the case that the N service effect index values in the N recommended results do not correspond to the same service index, the first resource ID with the highest matching degree is determined as the second resource ID.

[0069] Specifically, the first resource ID is the resource ID with the highest matching degree with the current position in a single evaluation dimension. The higher the matching degree, the more positions in the similar positions of the current position subscribe to the recruitment service resource corresponding to the resource ID in the evaluation dimension.

[0070] Exemplarily, for the current position JD1: sales (function), fast consumption (industry), A city (work location), 10000 people (enterprise size), no requirement (work experience), the corresponding 5 recommended results are respectively: according to the function, the subscription product "job position top" can be recommended, the exposure is 8W+, the resume delivery is 8+, and the matching degree is 68%; according to the industry, the subscription product "super exposure" can be recommended, the exposure is 3W+, and the matching degree is 61%; according to the work location, the subscription product "brand top" can be recommended, the exposure is 10W+, and the matching degree is 27%; according to the enterprise size, the subscription product "brand top" can be recommended, the exposure is 10W+, and the matching degree is 83%; according to the work experience, the subscription product "smart invitation" can be recommended, the resume delivery is 10+, and the matching degree is 74%.

[0071] Since the above-mentioned 5 service effect indicator values contain both "exposure" and "resume delivery", the second resource ID cannot be filtered according to the service effect indicator value. The "brand top" with a matching degree of 83% can be filtered as the final recommended recruitment service resource according to the matching degree.

[0072] In some embodiments of the present application, the above-mentioned step 150 filters the second resource ID from the N first resource IDs in combination with the service effect indicator values in the N recommended results, which can include:

[0073] Obtaining a preset correlation degree associated with each evaluation dimension;

[0074] In the case that the N service effect indicator values in the N recommended results correspond to the same service indicator, multiplying the preset correlation degree by the service effect indicator value in the recommended result under the same evaluation dimension to obtain a first comprehensive indicator value of the N first resource IDs;

[0075] Determining the first resource ID with the highest first comprehensive indicator value as the second resource ID.

[0076] Specifically, the correlation degree associated with each evaluation dimension can be set in advance, which is used to represent the importance of the evaluation dimension when recommending the recruitment service resource. The correlation degree associated with each evaluation dimension can be set by the recruiter corresponding to the current position or uniformly set by the platform.

[0077] Exemplarily, the current position JD2 has 4 recommendation results, respectively: according to the function, the product "position top" is recommended to be subscribed, the exposure is 8W+, and the resume delivery is 8+; according to the industry, the product "super exposure" is recommended to be subscribed, the exposure is 3W+; according to the work location, the product "brand top" is recommended to be subscribed, the exposure is 10W+; according to the enterprise scale, the product "brand top" is recommended to be subscribed, the exposure is 10W+. The relevance degrees of the function dimension, the industry dimension, the work location dimension, and the enterprise scale dimension are 25%, 20%, 5%, and 30%, respectively. Then, the first comprehensive index values of the 4 are calculated as 8W*25%, 3W*20%, 10W*5%, and 10W*30%, and the "brand top" under the enterprise scale dimension is selected as the final recommended recruitment service resource.

[0078] In the embodiments of the present application, by setting the corresponding relevance degree for each evaluation dimension, a higher relevance degree can be set for a more important evaluation dimension, so that the recruitment service resource under the important evaluation dimension can obtain a higher weight value. If the relevance degree is uniformly set by the platform, the corresponding relevance degree of the data of the specified evaluation dimension can be improved during the activity, so as to improve the recommendation rate of the screened recruitment service resource under the evaluation dimension, and further improve the exposure rate of the main promotion recruitment service resource product. If the relevance degree is set by the recruiter corresponding to the current position, a higher weight value can be set for the recruitment service resource under the preferred evaluation dimension according to the preference of the recruiter, so as to better meet the evaluation needs of the recruiter.

[0079] In some embodiments of the present application, each recommendation model matches a first resource ID for the current position from each resource ID based on the position description information under a single evaluation dimension, which can specifically include:

[0080] Each recommendation model determines the matching degree between the current position and each resource ID based on the position description information under a single evaluation dimension, and screens the resource ID with the highest matching degree as the first resource ID.

[0081] In this way, the recommendation model analyzes the similar positions of the current position, the historical subscription behavior, the use effect, and the feedback information of the recruitment service resource, and screens the recruitment service resource with the highest adaptation degree to the current position, so as to meet the recruitment needs of the position as much as possible.

[0082] In some embodiments of the present application, the matching degree between the current position and the first resource ID is also included in the recommendation result, and the above step 150 screens a second resource ID from the N first resource IDs in combination with the service effect index value in the N recommendation results, which can specifically include:

[0083] A preset relevance degree associated with each evaluation dimension is obtained;

[0084] In the case that the N service effect index values in the N recommended results do not correspond to the same service index, the preset correlation degree is multiplied by the matching degree of the recommended result in the same evaluation dimension to obtain the second comprehensive index value of the N first resource IDs;

[0085] The first resource ID with the highest second comprehensive index value is determined as the second resource ID.

[0086] Referring to the above example, the 5 recommended results of the current position JD1 are: according to the function, the subscription product "position top" is recommended, the exposure is 8W+, the resume delivery is 8+, and the matching degree is 68%; according to the industry, the subscription product "super exposure" is recommended, the exposure is 3W+, and the matching degree is 61%; according to the work location, the subscription product "brand top" is recommended, the exposure is 10W+, the matching degree is 27%; according to the enterprise scale, the subscription product "brand top" is recommended, the exposure is 10W+, the matching degree is 83%; according to the work experience, the subscription product "intelligent invitation" is recommended, the resume delivery is 10+, and the matching degree is 74%.

[0087] Since the above-mentioned 5 service effect index values contain both "exposure" and "resume delivery", they do not correspond to the same service index, so the second resource ID cannot be screened according to the service effect index value. The preset correlation degree can be multiplied by the matching degree of the recommended result in the same evaluation dimension according to the matching degree.

[0088] The correlation degrees associated with the function dimension, the industry dimension, the work location dimension, the enterprise scale dimension, and the work experience dimension are 25%, 20%, 5%, 30%, and 10%, respectively. After calculating the 5 second comprehensive index values 68%*25%, 61%*20%, 27%*5%, 83%*30%, and 74%*10%, the "brand top" with the second comprehensive index value of 83%*30% is selected as the final recommended recruitment service resource.

[0089] In the embodiments of the present application, when the N service effect indicator values of the N recommendation results do not correspond to the same service indicator, the service effects generated by different recruitment service resources cannot be objectively measured between different service indicators, so the matching degree can be used as an objective measurement indicator to ensure the objectivity and accuracy of the measurement, thereby improving the accuracy of the recruitment service resource recommendation. By multiplying the preset correlation degree and the matching degree, a higher correlation degree can be set for more important evaluation dimensions, so that the recruitment service resources under the important evaluation dimensions can obtain higher weight values. If the correlation degree is set uniformly by the platform, the corresponding correlation degree of the data of the specified evaluation dimension can be improved during the activity to improve the recommendation rate of the recruitment service resources screened under the evaluation dimension, thereby improving the exposure rate of the main recruitment service resource product. If the correlation degree is set by the recruiter corresponding to the current position, a higher weight value can be set for the recruitment service resources under the preferred evaluation dimension according to the recruiter's preferences, so as to better meet the evaluation needs of the recruiter.

[0090] As a specific example, in the present application, when the recruiter wants to subscribe to the recruitment service resource for the current position, the external interface is responsible for receiving the resource recommendation request sent by the recruiter from the app end or the pc end, and transmitting the resource recommendation request to the service end logic processing module. The service end logic processing module transmits the position description information of the current position to the large model data, and then obtains the recommendation data through the large model data, and returns the data to the external interface after formatting and standardizing. Then provided to the app end or pc end by the external interface, and displayed to the recruiter.

[0091] The recruiter subscription data, the use effect data after subscribing to the service, and the evaluation and feedback data of the recruiter are cleaned and then model training is performed.

[0092] 1. Collect positions that have subscribed to recruitment service resources, and obtain the exposure, the number of obtained deliveries, and the evaluation and feedback of the recruiter after using the recruitment service resource through data link tracking;

[0093] 2. The model training adopts multi-card training, and the optimization algorithm adopts an asynchronous stochastic gradient descent algorithm;

[0094] 3. The model data is manually adjusted and evaluated;

[0095] 4. Deployed to the actual environment for application, and the subsequent data is continuously tracked and then re-input into the model for continuous optimization;

[0096] 5. After collecting the evaluation and feedback data of the recruiter, the model is manually intervened and adjusted according to the actual situation.

[0097] Corresponding to the method embodiments of the present application, the present application also provides a recruitment service resource recommendation device based on a recruitment scenario.

[0098] Figure 2 is a structural schematic diagram of a recruitment service resource recommendation device based on a recruitment scenario provided by an embodiment of the present application. As shown in the figure, the recruitment service resource recommendation device 200 based on the recruitment scenario can include an acquisition module 210, a training module 220, an evaluation module 230, an output module 240, and a recommendation module 250. Figure 2

[0099] The acquisition module 210 is configured to acquire recruitment data of each position published historically, wherein the recruitment data includes position description information and recruitment service resource data corresponding to each position, the recruitment service resource data includes a resource ID associated with the position, and service effect data and service evaluation data after subscribing to a recruitment service resource corresponding to the resource ID for the position; the training module 220 is configured to divide the recruitment data of each position into N evaluation dimensions, and train a pre-trained model based on the recruitment data under a single evaluation dimension to obtain a recommendation model associated with the single evaluation dimension, wherein the recommendation model is used to match recruitment service resources for the position under the single evaluation dimension; the evaluation module 230 is configured to, in a case where the position description information of a current position is acquired, respectively input the position description information of the current position into the N recommendation models, so that each recommendation model matches a first resource ID for the current position from each resource ID based on the position description information under the single evaluation dimension, and evaluates a service effect of a recruitment service resource of the first resource ID on the current position to obtain a service effect index value; the output module 240 is configured to take the first resource ID and the service effect index value output by each recommendation model as a recommendation result to obtain N recommendation results output by the N recommendation models; and the recommendation module 250 is configured to combine the service effect index values in the N recommendation results, filter a second resource ID from the N first resource IDs, and determine a recruitment service resource corresponding to the second resource ID as a recommended recruitment service resource for the current position.

[0100] ​The device for recommending recruitment service resources based on a recruitment scenario provided in the embodiments of the present application can reflect the preferences and service effects of different types of positions for subscribing to recruitment service resources under the same evaluation dimension due to the recruitment data under a single evaluation dimension. Therefore, after a recommendation model is obtained by training a model based on the recruitment data under a single evaluation dimension, the position description information of a current position is input into the recommendation model. The recommendation model analyzes historical subscription behaviors, use effects, and feedback information of the position and the recruitment service resources, and mines the correlation between different recruitment service resources and positions, so as to provide a personalized recommendation scheme for the position to be subscribed to. Specifically, the position description information of the current position is analyzed to obtain positions similar to the position characteristics of the current position, and the recruitment service resources subscribed to by the similar positions are combined to match a first resource ID for the current position from each resource ID. In this way, a matching recruitment service resource is screened for the current position for each evaluation dimension, and preliminary screening under a single evaluation dimension is completed. On this basis, the N service effect index values corresponding to the N first resource IDs are combined to accurately predict the service effects of the recruitment service resources corresponding to the N first resource IDs on the current position, and a second resource ID is further screened based on the predicted service effects, and the recruitment service resource corresponding to the second resource ID is taken as a recommended recruitment service resource, so that a recruitment service resource with better service effect can be accurately screened for the current position for recommendation, thereby assisting recruiters to quickly locate and select recruitment service resources with higher adaptability and better expected effects, providing more accurate and personalized recruitment service recommendation schemes for recruiters, shortening the recruitment cycle, and improving recruitment efficiency and effect.

[0101] In some embodiments of the present application, the recommendation module is specifically configured to: obtain a preset correlation degree associated with each evaluation dimension; in the case that the N service effect index values in the N recommendation results all correspond to the same service index, multiply the preset correlation degree by the service effect index value in the recommendation result under the same evaluation dimension to obtain a first comprehensive index value of the N first resource IDs; and determine the first resource ID with the highest first comprehensive index value as the second resource ID.

[0102] In some embodiments of the present application, the evaluation module is specifically configured to: based on the position description information, determine the matching degree between the current position and each resource ID under a single evaluation dimension for each recommendation model, and screen the resource ID with the highest matching degree as the first resource ID.

[0103] In some embodiments of the present application, the matching degree between the current position and the first resource ID is also included in the recommendation result, and the recommendation module is specifically configured to: obtain a preset correlation degree associated with each evaluation dimension; in a case where the N service effect indicator values in the N recommendation results do not correspond to the same service indicator, multiply the preset correlation degree by the matching degree in the recommendation result under the same evaluation dimension to obtain a second comprehensive indicator value of the N first resource IDs; and determine the first resource ID with the highest second comprehensive indicator value as the second resource ID.

[0104] In some embodiments of the present application, the N evaluation dimensions include a function dimension, an industry dimension, a region dimension, an enterprise size dimension, a salary range dimension, and a work experience dimension, and the service effect data and the service effect indicator value both include exposure of the position and / or the number of resume deliveries.

[0105] In some embodiments of the present application, the training module is specifically configured to: split the position description information of each position into N evaluation dimensions to obtain N position label information of each position information under the N evaluation dimensions; and for each evaluation dimension, divide the recruitment service resource data of each position and the position label information under the corresponding evaluation dimension into the same group to obtain recruitment data under a single evaluation dimension.

[0106] In some embodiments of the present application, further comprising: a screening module configured to, before dividing the recruitment data of each position into N evaluation dimensions, screen a target position associated with a third resource ID from each position, wherein the third resource ID corresponds to a resource type of the recruitment service resource being a task-fulfilling type; a determination module configured to determine, by analyzing service effect data of the target position, whether exposure of the target position and / or the number of resume deliveries in a service period of subscribing to the recruitment service resource corresponding to the third resource ID reaches a task fulfillment amount; and the training module is specifically configured to: in a case where the task fulfillment amount is not reached, train the pre-trained model by taking the recruitment data of the target position under a single evaluation dimension as a negative sample.

[0107] In some embodiments of the present application, the matching degree between the current position and the first resource ID is also included in the recommendation result, and the recommendation module is specifically configured to: in a case where the N service effect indicator values in the N recommendation results correspond to the same service indicator, determine the first resource ID with the highest service effect indicator value as the second resource ID; and in a case where the N service effect indicator values in the N recommendation results do not correspond to the same service indicator, determine the first resource ID with the highest matching degree as the second resource ID.

[0108] The recruitment service resource recommendation device based on a recruitment scenario provided by the embodiments of the present application can implement Figure 1 the processes implemented by the service platform in the method embodiments, and can achieve the same technical effects. To avoid repetition, details are not repeated here.

[0109] Figure 3 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application.

[0110] like Figure 3 As shown, the electronic device 300 includes a memory 301 , a processor 302 , and a computer program stored in the memory 301 and executable on the processor 302 .

[0111] In an example, the processor 302 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.

[0112] The memory 301 may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical or other physical / tangible memory storage device. Therefore, generally, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., a memory device) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method for recommending recruitment service resources based on a recruitment scenario in the embodiment of the first aspect of the present application.

[0113] The processor 302 runs a computer program corresponding to the executable program code by reading the executable program code stored in the memory 301, so as to implement the recruitment service resource recommendation method based on the recruitment scenario in the embodiment of the first aspect.

[0114] In some examples, the electronic device 300 may further include a communication interface 303 and a bus 310. Figure 3 As shown, the memory 301 , the processor 302 , and the communication interface 303 are connected via a bus 310 and communicate with each other.

[0115] The communication interface 303 is mainly used to implement communication between the modules, devices, units and / or equipment in the embodiment of the present application. Input devices and / or output devices can also be connected through the communication interface 303.

[0116] The bus 310 includes hardware, software, or both that couples the components of the electronic device 300 to each other. By way of example and not limitation, the bus 310 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand 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-E) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, the bus 310 may include one or more buses. Although embodiments herein describe and illustrate a particular bus, this application contemplates any suitable bus or interconnect.

[0117] The electronic device provided in the embodiment of the present application can realize Figure 1 The various processes implemented by the electronic device in the method embodiment can achieve the same technical effect, and to avoid repetition, they will not be described here.

[0118] In conjunction with the recruitment service resource recommendation method based on recruitment scenarios in the above embodiments, embodiments of the present application may provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when executed by a processor, the computer program instructions implement the steps of any of the recruitment service resource recommendation methods based on recruitment scenarios in the above embodiments.

[0119] In conjunction with the recruitment service resource recommendation method based on recruitment scenarios in the above embodiments, embodiments of the present application may provide a computer program product for implementation. This (computer) program product is stored in a non-volatile storage medium and, when executed by at least one processor, implements the steps of any of the recruitment service resource recommendation methods based on recruitment scenarios in the above embodiments.

[0120] The chip provided by the embodiment of the present application comprises a processor and a communication interface, the communication interface is coupled with the processor, the processor is used for running programs or instructions to realize the processes of the method for recommending a recruitment service resource based on a recruitment scenario and achieve the same technical effects. To avoid repetition, details are not described here.

[0121] It should be understood that the chip mentioned in the embodiment of the present application can also be referred to as a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0122] It should be noted that the present application is not limited to the specific configurations and processes described above and shown in the drawings. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present application are not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between steps, after understanding the spirit of the present application.

[0123] The functional blocks shown in the structural block diagram described above 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 functional card, etc. When implemented in software, the elements of the present application are program or code segments used to perform the required tasks. The program or code segments can be stored in a machine-readable medium or transmitted through a data signal carried in a carrier wave over a transmission medium or communication link. The "machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segments can be downloaded via a computer network such as the Internet, an intranet, etc.

[0124] It should also be noted that the exemplary embodiments mentioned in the present application describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiments, or in an order different from the embodiments, or several steps can be performed simultaneously.

[0125] The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0126] The above only specifically describes the embodiments of the present application. For the convenience and brevity of description, the specific working processes of the above-described system, module and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described herein. It should be understood that the protection scope of the present application is not limited in this way. Any person skilled in the art can easily think of various equivalent modifications or replacements within the technical range disclosed in the present application, and these modifications or replacements should be covered within the protection scope of the present application.

Claims

1. A recruitment service resource recommendation method based on recruitment scenarios, characterized in that: include: Obtaining recruitment data for each historically published position, wherein the recruitment data includes position description information and recruitment service resource data corresponding to each position, wherein the recruitment service resource data includes the resource ID associated with the position, as well as service performance data and service evaluation data after subscribing to the recruitment service resource corresponding to the resource ID for the position; Dividing the recruitment data for each position into N evaluation dimensions, and training a pre-trained model based on the recruitment data under a single evaluation dimension to obtain a recommendation model associated with the single evaluation dimension, wherein the recommendation model is used to match recruitment service resources for the position under the single evaluation dimension; When the job description information of the current job is obtained, the job description information of the current job is input into each of N recommendation models, so that each recommendation model matches a first resource ID for the current job from various resource IDs based on the job description information under a single evaluation dimension, and evaluates the service effect of the recruitment service resource of the first resource ID on the current job, thereby obtaining a service effect index value; The first resource ID and service effect index value output by each recommendation model are used as the recommendation result, and N recommendation results output by N recommendation models are obtained; In combination with the service effect index values ​​in the N recommendation results, a second resource ID is screened from the N first resource IDs, and the recruitment service resource corresponding to the second resource ID is determined as the recommended recruitment service resource for the current position.

2. The method according to claim 1, characterized in that In combination with the service effect indicator values ​​in the N recommendation results, the second resource ID is selected from the N first resource IDs, including: Obtain the preset relevance associated with each evaluation dimension; When the N service effect index values ​​in the N recommendation results all correspond to the same service index, multiplying the preset correlation by the service effect index value in the recommendation results under the same evaluation dimension to obtain the first comprehensive index value of the N first resource IDs; The first resource ID with the highest first comprehensive index value is determined as the second resource ID.

3. The method according to claim 1, characterized in that Each recommendation model matches a first resource ID for the current position from various resource IDs based on the position description information under a single evaluation dimension, including: Each recommendation model determines the matching degree between the current position and each resource ID based on the position description information under a single evaluation dimension, and selects the resource ID with the highest matching degree as the first resource ID.

4. The method according to claim 3, characterized in that The recommendation result also includes a matching degree between the current position and the first resource ID. In combination with the service effect index values ​​in the N recommendation results, a second resource ID is selected from the N first resource IDs, including: Obtain the preset relevance associated with each evaluation dimension; If the N service effect index values ​​in the N recommendation results do not correspond to the same service index, multiplying the preset relevance by the matching degree in the recommendation results under the same evaluation dimension to obtain the second comprehensive index value of the N first resource IDs; The first resource ID with the highest second comprehensive index value is determined as the second resource ID.

5. The method according to claim 1, wherein The N evaluation dimensions include functional dimension, industry dimension, regional dimension, enterprise size dimension, salary range dimension, and work experience dimension. The service effect data and the service effect index value both include the exposure of the position and / or the number of resumes submitted.

6. The method according to claim 1, characterized in that Divide the recruitment data of each position into N evaluation dimensions, including: Split the job description information of each position into N evaluation dimensions, and obtain N job label information of each position information under N evaluation dimensions; For each evaluation dimension, the recruitment service resource data of each position and the position label information under the corresponding evaluation dimension are divided into the same group to obtain the recruitment data under a single evaluation dimension.

7. The method according to claim 1, characterized in that The pre-trained model is trained based on recruitment data under a single evaluation dimension, including: Filtering target positions associated with a third resource ID from the positions, wherein the resource type of the recruitment service resource corresponding to the third resource ID is a standard-achieving task type; Determine, by analyzing the service performance data of the target position, whether the exposure and / or the number of resume submissions of the target position within the service period of the recruitment service resource corresponding to the third resource ID has reached the task completion amount; In the case that the task target is not achieved, the recruitment data of the target position under a single evaluation dimension is used as a negative sample to train the pre-training model.

8. The method according to claim 1, characterized in that The recommendation result also includes a matching degree between the current position and the first resource ID. In combination with the service effect index values ​​in the N recommendation results, a second resource ID is selected from the N first resource IDs, including: When the N service effect index values ​​in the N recommendation results all correspond to the same service index, determining the first resource ID with the highest service effect index value as the second resource ID; In a case where the N service effect indicator values ​​in the N recommendation results do not correspond to the same service indicator, the first resource ID with the highest matching degree is determined as the second resource ID.

9. A recruitment service resource recommendation device based on recruitment scenarios, characterized in that: include: An acquisition module is used to obtain recruitment data for each historically published position, wherein the recruitment data includes the position description information and recruitment service resource data corresponding to each position, and the recruitment service resource data includes the resource ID associated with the position, as well as service effect data and service evaluation data after subscribing to the recruitment service resource corresponding to the resource ID for the position; a training module, configured to divide the recruitment data for each position into N evaluation dimensions, and train a pre-trained model based on the recruitment data under a single evaluation dimension to obtain a recommendation model associated with the single evaluation dimension, wherein the recommendation model is used to match recruitment service resources to the position under the single evaluation dimension; An evaluation module is configured to, upon obtaining job description information for a current job, input the job description information for the current job into each of N recommendation models, so that each recommendation model matches a first resource ID for the current job from various resource IDs based on the job description information under a single evaluation dimension, and evaluate the service effect of the recruitment service resource of the first resource ID on the current job, thereby obtaining a service effect index value; An output module, configured to use the first resource ID and service effect indicator value output by each recommendation model as a recommendation result, thereby obtaining N recommendation results output by the N recommendation models; The recommendation module is used to combine the service effect index values ​​in the N recommendation results, screen the second resource ID from the N first resource IDs, and determine the recruitment service resource corresponding to the second resource ID as the recommended recruitment service resource for the current position.

10. An electronic device, characterized in that: The electronic device comprises: a processor and a memory storing computer program instructions; and when the electronic device executes the computer program instructions, the method according to any one of claims 1 to 8 is implemented.

Images