Position exposure smoothing method and device based on proportional integral control and electronic equipment

Abstract

The invention discloses a position exposure smoothing method and device based on proportional integral control and electronic equipment, and the method comprises the steps: determining the target exposure of a current time period according to the total exposure budget of a target recruitment position in one period and the exposure proportion of each time period, and in one period, determining the target exposure of the current time period according to the total exposure budget of the target recruitment position; the exposure ratio of at least two non-adjacent time periods is higher than that of other time periods; at the ith moment of the current time period, obtaining the actual exposure of the current time period, and calculating an exposure difference value; based on a proportional integral control algorithm, according to the exposure difference value and the target exposure, calculating a dynamic adjustment error corresponding to the ith moment; adjusting the initial issuing threshold value based on the learning step length and the dynamic adjustment error to obtain a dynamic issuing threshold value; and when the matching degree pre-estimated score between the target job seeker and the target recruitment position is greater than or equal to the dynamic issuing threshold value, pushing the target recruitment position to the target job seeker. According to the invention, the click rate and the exhibition and investment ratio of the target recruitment position can be improved.

Description

Technical Field

[0001] The present application relates to the field of big data service technology, and in particular to a method, device, electronic device, computer-readable storage medium and computer program product for position exposure smoothing based on proportional integral control. Background Art

[0002] In the highly competitive recruitment market, recruiting companies hope to increase the visibility of their positions so that more job seekers can see and apply for the positions they post. By purchasing job exposure, recruiting companies can give priority to the recruitment information they post on the recruitment platform, attract more suitable job seekers, and thus improve recruitment efficiency. In addition, increasing job exposure helps to quickly form a candidate pool, thereby accelerating the interview and recruitment process, helping recruiting companies find suitable candidates faster, and shortening recruitment time.

[0003] In the recruitment industry, the current method is to increase the exposure of job openings to increase the number of clicks and submissions from users. Although this has increased the number of clicks and submissions from job openings to a certain extent, the cost of investment will also increase significantly. It is possible that multiple exposures will be required to exchange for one user click or submission. The ratio of increased exposures effectively converted into user clicks and submissions is low, that is, the click-through rate, exposure-to-subscription ratio and return rate of job openings are still low.

[0004] Some of the current exposure budget smoothing algorithms generally directly control the exposure, thereby indirectly increasing the click-through rate of recruitment positions. However, this method still has a low effect on improving the click-through rate, exposure ratio and return on investment of recruitment positions, which is difficult to meet the expectations of recruitment companies, resulting in a low user experience for recruitment companies. Summary of the invention

[0005] In view of this, the embodiments of the present application provide a position exposure smoothing method, device, electronic device, computer-readable storage medium and computer program product based on proportional integral control, which are used to solve at least one of the above-mentioned technical problems.

[0006] In a first aspect, an embodiment of the present application provides a method for smoothing job exposure based on proportional integral control, including: determining a target exposure of a target job in a current time period according to a total exposure budget of the target job in a cycle and an exposure ratio of each time period in a preset cycle, wherein, in a cycle, the exposure ratio of at least two non-adjacent time periods is higher than the exposure ratio of other time periods; at the i-th moment in the current time period, obtaining the actual exposure of the target job in the current time period, and calculating the exposure difference between the target exposure and the actual exposure; wherein, the current time period includes n moments, and the interval between any two adjacent moments is a preset time length, 1≤i≤n, and n is a large wherein the dynamic adjustment error corresponding to the ith moment is calculated based on the proportional-integral control algorithm and according to the exposure difference and the target exposure; the preset initial delivery threshold is adjusted based on the preset learning step and the dynamic adjustment error corresponding to the ith moment to obtain the dynamic delivery threshold; when the estimated match score between the target job seeker and the target recruitment position is greater than or equal to the dynamic delivery threshold, the target recruitment position is pushed to the target job seeker in the current time period; when the estimated match score is less than the dynamic delivery threshold, the target recruitment position is not pushed to the target job seeker in the current time period or the target recruitment position is pushed to the target job seeker according to the preset delivery probability, wherein the preset delivery probability is less than 100%.

[0007] According to some embodiments of the present application, optionally, at least two non-adjacent time periods include a first time period and a second time period, the first time period is a time period in the morning, and the second time period is a time period in the afternoon or evening.

[0008] According to some embodiments of the present application, optionally, before adjusting the preset initial delivery threshold based on the preset learning step size and the dynamic adjustment error corresponding to the i-th moment to obtain the dynamic delivery threshold, the position exposure smoothing method based on proportional integral control also includes: obtaining the historical exposure ratio of the target grid under different historical matching degree estimation scores; wherein different historical matching degree estimation scores correspond to different historical exposure ratios, the target grid is a grid composed of the workplace where the target recruitment position is located and the function corresponding to the target recruitment position, the historical matching degree estimation is divided into the matching degree estimation score between the job seekers in the historical period and the recruitment positions under the target grid, and the historical exposure ratio is the exposure ratio of multiple recruitment positions of the target grid in the historical period; the historical matching degree estimation score corresponding to the target historical exposure ratio is used as the initial delivery threshold.

[0009] According to some embodiments of the present application, optionally, based on a proportional-integral control algorithm and according to the exposure difference and the target exposure, a dynamic adjustment error corresponding to the i-th moment is calculated, including:

[0010] According to the following expression, the dynamic adjustment error corresponding to the i-th moment is calculated:

[0011]

[0012] Among them, δ i Indicates the exposure difference, E i represents the target exposure, μ represents the preset smoothing coefficient, α represents the preset reference error rate, and e i-1 represents the dynamic adjustment error corresponding to the i-1th moment in the current time period, where when i=1, e i-1 is 0.

[0013] According to some embodiments of the present application, optionally, a preset initial sending threshold is adjusted based on a preset learning step and a dynamic adjustment error corresponding to the i-th moment to obtain a dynamic sending threshold, including: calculating the product of the learning step and the dynamic adjustment error corresponding to the i-th moment; calculating the difference between the initial sending threshold and the product to obtain the dynamic sending threshold.

[0014] According to some embodiments of the present application, optionally, before determining whether the estimated score of the match between the target job seeker and the target recruitment position is greater than or equal to the dynamically issued threshold, the position exposure smoothing method based on proportional integral control also includes: in response to the position recommendation request of the target job seeker, inputting the portrait features of the target job seeker and the position features of the target recruitment position into the trained position recommendation model to obtain the estimated score of the match between the target job seeker and the target recruitment position predicted by the position recommendation model; wherein, the portrait features of the target job seeker include at least one of the following: the education, major, industry, age, gender, identity, work experience, expected job function, expected work area, expected salary and expected job type of the target job seeker; the position features of the target recruitment position include at least one of the following: the function, industry, education requirements, major requirements, work experience requirements, salary, work area, job type, age requirements and gender requirements of the target recruitment position.

[0015] In a second aspect, an embodiment of the present application provides a position exposure smoothing device based on proportional integral control, including: a determination module, used to determine the target exposure of the target recruitment position in the current time period according to the total exposure budget of the target recruitment position in a cycle and the exposure ratio of each time period in a preset cycle, wherein in a cycle, the exposure ratio of at least two non-adjacent time periods is higher than the exposure ratio of other time periods; a first calculation module, used to obtain the actual exposure of the target recruitment position in the current time period at the i-th moment in the current time period, and calculate the exposure difference between the target exposure and the actual exposure; wherein the current time period includes n moments, and the interval between any two adjacent moments is a preset time length, 1≤i≤n, n is an integer greater than 1. number, i is a positive integer; a second calculation module, used to calculate the dynamic adjustment error corresponding to the i-th moment based on the proportional-integral control algorithm and according to the exposure difference and the target exposure; an adjustment module, used to adjust the preset initial sending threshold based on the preset learning step and the dynamic adjustment error corresponding to the i-th moment to obtain the dynamic sending threshold; a sending module, used to push the target recruitment position to the target job seeker in the current time period when the estimated match score between the target job seeker and the target recruitment position is greater than or equal to the dynamic sending threshold; when the estimated match score is less than the dynamic sending threshold, not push the target recruitment position to the target job seeker in the current time period or push the target recruitment position to the target job seeker according to the preset sending probability, wherein the preset sending probability is less than 100%.

[0016] In a third aspect, an embodiment of the present application provides an electronic device, comprising a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the steps of the position exposure smoothing method based on proportional integral control as described above are implemented.

[0017] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having computer program instructions stored thereon, and when the computer program instructions are executed by a processor, the steps of the position exposure smoothing method based on proportional integral control as described above are implemented.

[0018] In a fifth aspect, 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 position exposure smoothing method based on proportional integral control as described above are implemented.

[0019] The method, device, electronic device, computer-readable storage medium and computer program product for job exposure smoothing based on proportional integral control provided by the embodiments of the present application, on the one hand, by setting the exposure ratio of the target recruitment position in each time period in a cycle, in one cycle, the exposure ratio of at least two non-adjacent time periods is higher than the exposure ratio of other time periods, which can effectively prevent the exposure budget of the target recruitment position from being consumed too quickly, and effectively avoid the concentrated release of the exposure budget in a single time period and missing the subsequent peak high-quality traffic, so that more job seekers in different login time periods can see the target recruitment position, which is conducive to improving the click-through rate and exposure-to-investment ratio of the target recruitment position; on the other hand, the proportional integral control can adjust the dynamic issuance threshold according to the preset duration or preset frequency, and then adjust the exposure of the target recruitment position in time. Corresponding delivery strategies are given to target job openings that are over-exposed or under-exposed, thereby reducing the loss of exposure. On the other hand, by setting a comparison mechanism between the estimated match score between the target job seeker and the target job opening and the dynamic delivery threshold, when the estimated match score is greater than or equal to the dynamic delivery threshold, the target job opening is pushed to the target job seeker in the current time period; when the estimated match score is less than the dynamic delivery threshold, the target job opening is not pushed to the target job seeker in the current time period or is pushed to the target job seeker according to the preset delivery probability, so that the target job opening can be given priority to job seekers with higher matching degrees. The higher the matching degree, the more willing the job seeker is to click and submit the target job opening. Therefore, the click-through rate, display-to-investment ratio and return on investment of the target job opening can be improved, thereby improving the user experience and satisfaction of the recruiting company. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] 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.

[0021] Figure 1 A schematic flow chart of a method for smoothing job exposure based on proportional-integral control provided in an embodiment of the present application.

[0022] Figure 2 The curve schematically shows the exposure ratio of the target recruitment position in each time period in a cycle.

[0023] Figure 3 A schematic flow chart of determining a dynamic sending threshold value for a position exposure smoothing method based on proportional integral control provided in an embodiment of the present application.

[0024] Figure 4 A first exposure curve that has not been processed for exposure balancing and a second exposure curve that has been processed for exposure balancing are schematically shown.

[0025] Figure 5A schematic diagram of the structure of a position exposure smoothing device based on proportional-integral control provided in an embodiment of the present application.

[0026] Figure 6 A schematic diagram of the hardware structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.

[0028] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0029] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0030] It is obvious to those skilled in the art that various modifications and changes can be made in the present application without departing from the spirit or scope of the present application. Therefore, the present application is intended to cover modifications and changes of the present application that fall within the scope of the corresponding claims (technical solutions for protection) and their equivalents. It should be noted that the implementation methods provided in the embodiments of the present application can be combined with each other without contradiction.

[0031] Before describing the technical solutions provided by the embodiments of the present application, in order to facilitate the understanding of the embodiments of the present application, the present application first specifically describes the problems existing in the related art:

[0032] In the highly competitive recruitment market, recruiting companies hope to increase the visibility of their positions so that more job seekers can see and apply for the positions they post. By purchasing job exposure, recruiting companies can give priority to the recruitment information they post on the recruitment platform, attract more suitable job seekers, and thus improve recruitment efficiency. In addition, increasing job exposure helps to quickly form a candidate pool, thereby accelerating the interview and recruitment process, helping recruiting companies find suitable candidates faster, and shortening recruitment time.

[0033] In the recruitment industry, the current method is to increase the exposure of job openings to increase the number of clicks and submissions from users. Although this has increased the number of clicks and submissions from job openings to a certain extent, the cost of investment will also increase significantly. It is possible that multiple exposures will be required to exchange for one user click or submission. The ratio of increased exposures effectively converted into user clicks and submissions is low, that is, the click-through rate, exposure-to-subscription ratio and return rate of job openings are still low.

[0034] Some of the current exposure budget smoothing algorithms generally directly control the exposure, thereby indirectly increasing the click-through rate of recruitment positions. However, this method still has a low effect on improving the click-through rate, exposure ratio and return on investment of recruitment positions, which is difficult to meet the expectations of recruitment companies, resulting in a low user experience for recruitment companies.

[0035] In view of the above research findings of the inventors, the embodiments of the present application provide a position exposure smoothing method, device, electronic device, computer-readable storage medium and computer program product based on proportional integral control, which can solve at least one of the above technical problems existing in the related art.

[0036] The following first introduces the position exposure smoothing method based on proportional integral control provided in an embodiment of the present application.

[0037] Figure 1 A flow chart of a method for smoothing job exposure based on proportional integral control provided in an embodiment of the present application. Figure 1 As shown, the position exposure smoothing method based on proportional integral control provided in the embodiment of the present application may include the following steps S101 to S105.

[0038] S101: Determine the target exposure of the target job in the current time period based on the total exposure budget of the target job in a cycle and the exposure ratio of each time period in a preset cycle, wherein in a cycle, the exposure ratio of at least two non-adjacent time periods is higher than the exposure ratio of other time periods.

[0039] A cycle can be one day, three days, one week or one month, etc., and can be flexibly adjusted according to actual conditions, and this application does not limit this. Unless otherwise specified, this application will take a cycle of one day as an example to illustrate the position exposure smoothing method based on proportional integral control provided in the embodiment of this application. The target recruitment position can be any recruitment position. The total exposure budget of the target recruitment position in a cycle can be flexibly adjusted according to actual conditions, and this application does not limit this. For example, the total exposure budget of the target recruitment position in a cycle is 1000, that is, the target recruitment position is exposed 1000 times in one cycle, such as pushing the target recruitment position to 1000 job seekers in one cycle.

[0040] The exposure ratio of the target job position in each time period in a cycle can be preset and can be flexibly adjusted according to actual conditions, and this application does not limit this. Taking a cycle of one day as an example, the exposure ratio of the target job position in each time period in a cycle can be the exposure ratio of the target job position in each hour of the day, such as 5% from 8 am to 9 am, 10% from 9 am to 10 am, ..., 10% from 13 pm to 14 pm, etc.

[0041] In one cycle, the exposure ratio of at least two non-adjacent time periods is higher than the exposure ratio of other time periods. Non-adjacent time periods are time periods that are not adjacent to each other, such as 10:00-11:00 a.m. and 15:00-16:00 p.m. There are other time periods between these two time periods, so these two time periods are non-adjacent time periods. For example, the exposure ratios of the two time periods of 10:00-11:00 a.m. and 15:00-16:00 p.m. are both 20%, which is higher than the exposure ratios of other time periods (such as 1%, 5%, 10%, etc.). It should be noted that the above exposure ratios are for example only and do not constitute a limitation of the present application.

[0042] In this way, the exposure ratio of the target recruitment position in each time period within a cycle can have at least two "peaks", which is conducive to preventing the exposure budget of the target recruitment position from being consumed too quickly, and effectively avoiding the concentrated release of the exposure budget of the target recruitment position in a single time period and missing out on subsequent peak high-quality traffic, so that more job seekers in different login time periods can see the target recruitment position, which is conducive to improving the exposure ratio of the target recruitment position and achieving smooth exposure of the target recruitment position.

[0043] For example, a high proportion of job seekers log in to the recruitment platform to search for job openings during the two time periods of 10am-11am and 3pm-4pm. By setting two "peaks" of exposure ratios during the two time periods of 10am-11am and 3pm-4pm, rather than concentrating the exposure budget of the target job openings in a single time period of 10am-11am, job seekers who log in to the recruitment platform during 3pm-4pm are more likely to receive recommended target job openings, which is conducive to improving the click-through rate and exposure-to-investment ratio of the target job openings.

[0044] The click-through rate of a job post can be the ratio of the number of times the job post is clicked to the number of times it is exposed. For example, if a job post is exposed 1,000 times and receives 50 clicks, then the click-through rate of the job post is 5%. The exposure-to-investment ratio of a job post can be the ratio of the number of times the job post is submitted to the number of times it is exposed. For example, if a job post is exposed 1,000 times and receives 50 clicks, then the exposure-to-investment ratio of the job post is 5%.

[0045] In S101, based on the total exposure budget of the target job position in a cycle and the exposure ratio of each time period in a preset cycle, the exposure budget of the target job position in each time period in a cycle can be determined. This includes the exposure budget of the target job position in the current time period. The current time period can be any time period in a cycle, such as 10:00-11:00 a.m. or 15:00-16:00 p.m. For ease of explanation, the exposure budget of the target job position in the current time period is referred to as the target exposure.

[0046] S102: At the i-th moment in the current time period, obtain the actual exposure amount of the target job position in the current time period, and calculate the exposure amount difference between the target exposure amount and the actual exposure amount.

[0047] The current time period includes n moments, and the interval between any two adjacent moments is a preset time length. The i-th moment can be any moment among the n moments, 1≤i≤n, n is an integer greater than 1, and i is a positive integer. The size of the preset time length can be flexibly adjusted according to actual conditions, and this application does not limit this. For example, the preset time length can be 1 minute, 5 minutes, 10 minutes, or 30 minutes, etc.

[0048] For example, if the current time period is 10:00-11:00 a.m. and the preset duration is 10 minutes, the n moments in the current time period may include 10:00, 10:10, 10:20, 10:30, 10:40 and 10:50. The i-th moment may be any of the above moments.

[0049] At the i-th moment in the current time period, the actual exposure of the target job position in the current time period can be obtained, that is, the actual exposure of the target job position from the start moment of the current time period to the i-th moment. Then, the exposure difference between the target exposure and the actual exposure is calculated.

[0050] S103: Based on the proportional-integral control algorithm and according to the exposure difference and the target exposure, the dynamic adjustment error corresponding to the i-th moment is calculated.

[0051] Proportional-Integral Control (PI control) combines the characteristics of proportional (P) control and integral (I) control, and can eliminate steady-state errors while responding quickly. In the embodiment of the present application, based on the proportional-integral control algorithm, the dynamic adjustment error corresponding to the i-th moment is calculated according to the difference between the actual exposure and the target exposure, so as to adjust the dynamic delivery threshold, and then adjust the exposure from the i-th moment to the i+1th moment, effectively solving the problem of overexposure or underexposure of the target recruitment position.

[0052] S104: adjusting the preset initial sending threshold based on the preset learning step size and the dynamic adjustment error corresponding to the i-th moment to obtain a dynamic sending threshold.

[0053] In S104, the preset initial delivery threshold can be adjusted based on the preset learning step size and the dynamic adjustment error corresponding to the i-th moment to obtain a dynamic delivery threshold. The initial delivery threshold and the learning step size can be flexibly adjusted according to actual conditions, and this application does not limit this. The initial delivery threshold and the learning step size can both be greater than 0 and less than 1.

[0054] S105: When the estimated match score between the target job seeker and the target job position is greater than or equal to the dynamic delivery threshold, the target job position is pushed to the target job seeker in the current time period; when the estimated match score is less than the dynamic delivery threshold, the target job position is not pushed to the target job seeker in the current time period or the target job position is pushed to the target job seeker according to the preset delivery probability.

[0055] The target job seeker may be any job seeker, and the estimated score of the matching degree between the target job seeker and the target recruitment position is a score reflecting the matching degree between the target job seeker and the target recruitment position.

[0056] In the recruitment field, when recommending a target job position to a job seeker, the profile features of the target job seeker and the position features of the target job position can be input into a trained job recommendation model. The position recommendation model can obtain an estimated score of the degree of match between the target job seeker and the target job position predicted by the position recommendation model based on the profile features of the target job seeker and the position features of the target job position.

[0057] For example, in some embodiments, the target job applicant's profile features may include at least one of the following: the target job applicant's education, major, industry, age, gender, identity, work experience, desired job function, desired work area, expected salary, and desired job type. The target job post's job features may include at least one of the following: the target job post's function, industry, education requirements, major requirements, work experience requirements, salary, work area, job type, age requirements, and gender requirements. The job type may include full-time, part-time, or internship.

[0058] The job recommendation model can set multiple scoring items and basic scores corresponding to each scoring item. For example, multiple scoring items may include educational requirements, professional requirements, work experience requirements, age requirements, gender requirements and / or salary, etc. If a certain item of the target job seeker meets the requirements of the corresponding scoring item, the basic score of the item is given, and if it does not meet the requirements, the basic score of the item is not given, and then the basic scores of the multiple scoring items obtained by the target job seeker are calculated, such as adding or multiplying, to obtain the estimated score of the match between the target job seeker and the target job position. In some embodiments, the weights corresponding to each scoring item can also be set, and the estimated score of the match between the target job seeker and the target job position is obtained by weighted calculation of the basic scores of the multiple scoring items obtained by the target job seeker.

[0059] It should be noted that the calculation process of the above-mentioned job recommendation model is only an example, and this application does not limit this. In other embodiments, in order to improve the prediction accuracy, in addition to the image features of the target job seeker and the job features of the target job recruitment, the job recommendation model can also input the recent behavior features of the target job seeker and / or the recent behavior features of the publisher of the target job recruitment.

[0060] The match estimation score between the target job seeker and the target job position can also be called the ranking score. In the recruitment field, when recommending multiple job positions including the target job position to job seekers, the multiple job positions can be ranked in descending order according to the match estimation score between the target job seeker and each job position. The higher the match estimation score, the higher the position is, and the easier it is for the target job seeker to view the position, and the easier it is for the target job seeker to click on it and submit it.

[0061] In S105, when the match estimation score between the target job seeker and the target job position is greater than or equal to the dynamic delivery threshold, it indicates that the match estimation score between the target job seeker and the target job position is high, and the target job position can be pushed to the target job seeker in the current time period. When the match estimation score between the target job seeker and the target job position is less than the dynamic delivery threshold, it indicates that the match estimation score between the target job seeker and the target job position is low, and the target job position can not be pushed to the target job seeker in the current time period, or the target job position can be pushed to the target job seeker according to the preset delivery probability. The preset delivery probability can be less than 100%. That is, there is a certain probability (i.e., the preset delivery probability x%) to push the target job position to the target job seeker, and there is also a certain probability (such as 1-(x%)) to filter out the target job position and not push the target job position to the target job seeker. The size of the preset delivery probability can be flexibly adjusted according to the actual situation, and this application does not limit this.

[0062] The embodiment of the present application provides a method for smoothing job exposure based on proportional integral control. On the one hand, by setting the exposure ratio of the target recruitment position in each time period in a cycle, in one cycle, the exposure ratio of at least two non-adjacent time periods is higher than the exposure ratio of other time periods, which can effectively prevent the exposure budget of the target recruitment position from being consumed too quickly, and effectively avoid the concentrated release of the exposure budget in a single time period and missing the subsequent peak high-quality traffic, so that more job seekers in different login time periods can see the target recruitment position, which is conducive to improving the click-through rate and exposure-to-investment ratio of the target recruitment position; on the other hand, the proportional integral control can adjust the dynamic delivery threshold according to the preset duration or preset frequency, and then adjust the exposure of the target recruitment position, and timely correct the target recruitment position that is over-exposed or under-exposed. The corresponding delivery strategy is given to the position, which reduces the loss of exposure. On the other hand, by setting a comparison mechanism between the estimated match score between the target job seeker and the target recruitment position and the dynamic delivery threshold, when the estimated match score is greater than or equal to the dynamic delivery threshold, the target recruitment position is pushed to the target job seeker in the current time period; when the estimated match score is less than the dynamic delivery threshold, the target recruitment position is not pushed to the target job seeker in the current time period or the target recruitment position is pushed to the target job seeker according to the preset delivery probability. This allows the target recruitment position to be given priority to job seekers with higher matching degrees. The higher the matching degree, the more willing the job seeker is to click and submit the target recruitment position. Therefore, the click-through rate, display-to-investment ratio and return on investment of the target recruitment position can be improved, thereby improving the user experience and satisfaction of the recruiting company.

[0063] Figure 2 The following diagram schematically shows the curve of the exposure ratio of the target job position in each time period in a cycle. Figure 2As shown, taking a cycle of one day as an example, the curve can also be called the hourly exposure ratio curve of the market position, with the horizontal axis being hours and the vertical axis being exposure ratio. In a cycle, at least two non-adjacent time periods may include a first time period t1 and a second time period t2, the first time period t1 may be a time period in the morning, such as 11 am to 12 pm, and the second time period t2 may be a time period in the afternoon or evening, such as 3 pm to 4 pm. Figure 2 The second time period t2 is taken as an example of a time period in the afternoon.

[0064] In this way, by setting the exposure ratio of the target recruitment position in each time period in a cycle, in one cycle, the exposure ratio of at least two non-adjacent time periods is higher than the exposure ratio of other time periods. This can effectively prevent the exposure budget of the target recruitment position from being consumed too quickly, and effectively avoid the exposure budget being concentrated in a single time period and missing out on subsequent peak high-quality traffic. This allows more job seekers in different login time periods to see the target recruitment position, which is conducive to improving the click-through rate and exposure-to-investment ratio of the target recruitment position.

[0065] According to some embodiments of the present application, optionally, for example, the initial sending threshold can be determined by referring to the historical match degree estimation score of the target grid. The target grid can be a grid composed of the workplace where the target recruitment position is located and the function corresponding to the target recruitment position, and the historical match degree estimation score of the target grid can be the match degree estimation score between the job seekers in the historical period and the recruitment position under the target grid.

[0066] Specifically, in some specific embodiments, optionally, in S104: adjusting the preset initial sending threshold based on the preset learning step size and the dynamic adjustment error corresponding to the i-th moment, before obtaining the dynamic sending threshold, the position exposure smoothing method based on proportional integral control may also include the following steps one and two.

[0067] Step 1: Obtain the historical exposure ratio of the target grid under different historical matching degree estimation scores.

[0068] In some embodiments, for example, the work place-function is a grid, and the work place can be a city (such as city A). The target grid can be a grid composed of the work place where the target job is located and the function corresponding to the target job. Under the target grid, there are multiple job positions with the same work place and the same function as the target job. The total exposure and historical exposure ratio of the multiple job positions in the historical period are known. The historical period can be one or more periods in the past, such as yesterday. For example, the total exposure of the multiple job positions in the historical period is 1000. The historical exposure ratios corresponding to different historical matching degree estimates are different. For example, the historical matching degree estimate score is positively correlated with the historical exposure ratio. That is, as the historical matching degree estimate score is higher, the historical exposure ratio is also higher. For any job position, in the historical period, the matching degree estimate scores between multiple different job seekers and the job position can all be in the range of 0 to 1. For example, in the historical period, the matching degree estimate scores between multiple different job seekers and job position a can all be in the range of 0 to 1, and the matching degree estimate scores between multiple different job seekers and job position b can all be in the range of 0 to 1. Recruitment position a and recruitment position b are different recruitment positions among the above multiple recruitment positions.

[0069] Then, for the multiple job openings, for example, the historical exposure ratio (such as 20%) under the statistical matching estimated score of 0.2 (≤0.2) is calculated. For example, taking the total exposure of the multiple job openings in the historical period as 1000, for example, the historical exposure of the multiple job openings under the matching estimated score of 0.2 is 200, and the historical exposure ratio under the matching estimated score of 0.2 is 20%. Similarly, for example, the historical exposure ratio (such as 50%) under the statistical matching estimated score of 0.5 (≤0.5) is calculated, and so on. The historical exposure ratio (such as 85%) under the statistical matching estimated score of 0.8 (≤0.8) is calculated, and so on. In this way, the historical exposure ratio of the target grid under different historical matching estimated scores can be obtained.

[0070] Step 2: Use the estimated historical matching score corresponding to the target historical exposure ratio as the initial delivery threshold.

[0071] According to the actual exposure demand, one of the historical exposure ratios can be selected as the target historical exposure ratio, and the historical matching score corresponding to the target historical exposure ratio can be used as the initial delivery threshold. For example, in some examples, for example, 50% of the historical exposure ratio is selected as the target historical exposure ratio, and the historical matching score corresponding to 50% of the historical exposure ratio is 0.5 as the initial delivery threshold.

[0072] In this way, determining the initial delivery threshold based on the historical match estimation score of the target grid can solve the problem that it is difficult to determine the initial delivery threshold due to too little data on the target job position in the historical period. On the other hand, determining the initial delivery threshold based on the historical match estimation score of the target grid is conducive to improving the accuracy of the delivery strategy, such as increasing the delivery ratio for job seekers with a higher match estimation score with the target job position, thereby improving the click-through rate, exhibition-to-investment ratio and return rate of the target job position.

[0073] Figure 3 A schematic diagram of a flow chart for determining a dynamic threshold for sending a job exposure smoothing method based on proportional integral control provided in an embodiment of the present application. Figure 1 and Figure 3 As shown, in S101, the target exposure Ei of the target job position in the current time period can be determined according to the total exposure budget of the target job position in a cycle and the exposure ratio of each time period in a preset cycle. In S102, at the i-th moment in the current time period, the actual exposure of the target job position in the current time period can be obtained, and the exposure difference δ between the target exposure and the actual exposure can be calculated. i In S103, based on the proportional-integral control algorithm, the exposure difference δ i and target exposure Ei, calculate the dynamic adjustment error e corresponding to the i-th moment i .

[0074] Specifically, in some specific embodiments, optionally, S103: calculating the dynamic adjustment error corresponding to the i-th moment based on the proportional integral control algorithm and according to the exposure difference and the target exposure, may include the following steps:

[0075] According to the following expression, calculate the dynamic adjustment error e corresponding to the i-th moment i :

[0076]

[0077] Among them, δ i Indicates the exposure difference, E i represents the target exposure, μ represents the preset smoothing coefficient, α represents the preset reference error rate, and e i-1 represents the dynamic adjustment error corresponding to the i-1th moment in the current time period, where when i=1, e i-1 The smoothing coefficient and the reference error rate can both be greater than 0 and less than 1. The smoothing coefficient and the reference error rate can be flexibly adjusted according to actual conditions, and this application does not limit this.

[0078] Combination Figure 1 and Figure 3As shown, according to some embodiments of the present application, optionally, S104: adjusting the preset initial delivery threshold based on the preset learning step size and the dynamic adjustment error corresponding to the i-th moment to obtain the dynamic delivery threshold may include the following steps:

[0079] Calculate the product of the learning step size and the dynamic adjustment error corresponding to the i-th moment;

[0080] The difference between the initial delivery threshold and the product is calculated to obtain the dynamic delivery threshold.

[0081] For example, the dynamic delivery threshold T can be calculated according to the following expression: i :

[0082] T i =T-step×e i (2)

[0083] Among them, T represents the initial sending threshold, step represents the learning step length, and e i Represents the dynamic adjustment error corresponding to the i-th moment.

[0084] In this way, the dynamic sending threshold is adjusted according to the preset duration or frequency based on proportional integral control, thereby adjusting the exposure of the target recruitment position, and timely giving the corresponding sending strategy to the target recruitment position with excessive or insufficient exposure, thereby reducing the loss of exposure. In addition, this application sets multi-parameter controls such as smoothing coefficient, baseline error rate and learning step length, which can flexibly and timely adjust the threshold starting range and change range, and can adapt to different sending probabilities according to scene requirements, so as to reduce the waste of exposure resources and ensure the synchronous completion of the entire cycle and the exposure budget targets of each time period.

[0085] As mentioned above, the estimated score of the match between the target job seeker and the target recruitment position can be predicted based on the trained position recommendation model.

[0086] Specifically, in some specific embodiments, optionally, before determining whether the estimated score of the matching degree between the target job seeker and the target job position is greater than or equal to the dynamically issued threshold, the job exposure smoothing method based on proportional integral control may further include the following steps:

[0087] In response to a job recommendation request from a target job seeker, the profile features of the target job seeker and the job features of the target job are input into a trained job recommendation model to obtain an estimated score of the matching degree between the target job seeker and the target job, as predicted by the job recommendation model.

[0088] Specifically, after the target job seeker logs in to the recruitment platform, the target job seeker's job recommendation request can be responded to by inputting the target job seeker's profile features and the job features of the target job recruitment position into the trained job recommendation model, thereby obtaining the match degree estimation score between the target job seeker and the target job recruitment position predicted by the job recommendation model. The specific process has been described above and will not be repeated here.

[0089] The effect of the position exposure smoothing method based on proportional integral control provided in the embodiment of the present application is verified in combination with a specific application embodiment.

[0090] In a specific application embodiment, the baseline error rate α is set to 0.15, the smoothing coefficient μ is set to 0.4, the learning step is set to 0.05, and the preset sending probability is set to 0% (that is, when the matching estimate score is less than the dynamic sending threshold, the target recruitment position will not be pushed to the target job seeker in the current time period).

[0091] Figure 4 A first exposure curve that has not been processed for exposure balancing and a second exposure curve that has been processed for exposure balancing are schematically shown. Figure 4 Taking a cycle of one day as an example, the horizontal axis is hours and the vertical axis is exposure ratio. Figure 4 As shown, the first exposure curve q1 without exposure balancing has only one single peak, and the exposure will be concentrated in a single time period, thus missing the subsequent peak quality traffic. The second exposure curve q2 is the exposure curve of the target recruitment position obtained by using the position exposure smoothing method based on proportional integral control provided in the embodiment of the present application. It can be seen that the second exposure curve q2 becomes smoother, and the exposure is more evenly filled in each time period. For example, the morning exposure peak moves from after nine o'clock to around eleven o'clock, and the exposure throughout the day changes from a single peak to a double peak. Overall, it is closer to the hourly exposure ratio curve of the market position. For example, after one day of data statistics, using the position exposure smoothing method based on proportional integral control provided in the embodiment of the present application, the exhibition-to-investment ratio of the target recruitment position increased by 50%, and the proportion of zero-delivery orders decreased by 8%, which did not have a significant negative impact on the market.

[0092] Based on the same technical concept as the position exposure smoothing method based on proportional integral control provided in the above embodiment, the present application also provides a position exposure smoothing device based on proportional integral control. Please refer to the following embodiments for details.

[0093] Figure 5 A schematic diagram of the structure of a position exposure smoothing device based on proportional integral control provided in an embodiment of the present application. Figure 5 As shown, the position exposure smoothing device 50 based on proportional integral control provided in the embodiment of the present application may include the following modules:

[0094] The determination module 501 is used to determine the target exposure of the target job position in the current time period according to the total exposure budget of the target job position in a cycle and the exposure ratio of each time period in a preset cycle, wherein in a cycle, the exposure ratio of at least two non-adjacent time periods is higher than the exposure ratio of other time periods;

[0095] The first calculation module 502 is used to obtain the actual exposure of the target job position in the current time period at the i-th moment in the current time period, and calculate the exposure difference between the target exposure and the actual exposure; wherein the current time period includes n moments, and the interval between any two adjacent moments is a preset time length, 1≤i≤n, n is an integer greater than 1, and i is a positive integer;

[0096] The second calculation module 503 is used to calculate the dynamic adjustment error corresponding to the i-th moment based on the proportional integral control algorithm and according to the exposure difference and the target exposure;

[0097] An adjustment module 504 is used to adjust the preset initial sending threshold based on the preset learning step size and the dynamic adjustment error corresponding to the i-th moment to obtain a dynamic sending threshold;

[0098] The sending module 505 is used to push the target recruitment position to the target job seeker in the current time period when the estimated matching score between the target job seeker and the target recruitment position is greater than or equal to the dynamic sending threshold; when the estimated matching score is less than the dynamic sending threshold, the target recruitment position is not pushed to the target job seeker in the current time period or the target recruitment position is pushed to the target job seeker according to a preset sending probability, wherein the preset sending probability is less than 100%.

[0099] The embodiment of the present application provides a position exposure smoothing device based on proportional integral control. On the one hand, by setting the exposure ratio of the target recruitment position in each time period in a cycle, in one cycle, the exposure ratio of at least two non-adjacent time periods is higher than the exposure ratio of other time periods, which can effectively prevent the exposure budget of the target recruitment position from being consumed too quickly, and effectively avoid the concentrated release of the exposure budget in a single time period and missing the subsequent peak high-quality traffic, so that more job seekers in different login time periods can see the target recruitment position, which is conducive to improving the click-through rate and exposure-to-investment ratio of the target recruitment position; on the other hand, the proportional integral control can adjust the dynamic delivery threshold according to the preset duration or preset frequency, and then adjust the exposure of the target recruitment position, and timely correct the target recruitment position that is over-exposed or under-exposed. The corresponding delivery strategy is given to the position, which reduces the loss of exposure. On the other hand, by setting a comparison mechanism between the estimated match score between the target job seeker and the target recruitment position and the dynamic delivery threshold, when the estimated match score is greater than or equal to the dynamic delivery threshold, the target recruitment position is pushed to the target job seeker in the current time period; when the estimated match score is less than the dynamic delivery threshold, the target recruitment position is not pushed to the target job seeker in the current time period or the target recruitment position is pushed to the target job seeker according to the preset delivery probability. This allows the target recruitment position to be given priority to job seekers with higher matching degrees. The higher the matching degree, the more willing the job seeker is to click and submit the target recruitment position. Therefore, the click-through rate, display-to-investment ratio and return on investment of the target recruitment position can be improved, thereby improving the user experience and satisfaction of the recruiting company.

[0100] According to some embodiments of the present application, optionally, at least two non-adjacent time periods include a first time period and a second time period, the first time period is a time period in the morning, and the second time period is a time period in the afternoon or evening.

[0101] According to some embodiments of the present application, optionally, the position exposure smoothing device 50 based on proportional integral control provided in the embodiments of the present application may also include a setting module for obtaining the historical exposure ratio of the target grid under different historical matching degree estimation scores; wherein different historical matching degree estimation scores correspond to different historical exposure ratios, the target grid is a grid composed of the workplace where the target recruitment position is located and the function corresponding to the target recruitment position, the historical matching degree estimation is divided into the matching degree estimation score between the job seekers in the historical period and the recruitment positions under the target grid, and the historical exposure ratio is the exposure ratio of multiple recruitment positions in the target grid in the historical period; the historical matching degree estimation score corresponding to the target historical exposure ratio is used as the initial sending threshold.

[0102] According to some embodiments of the present application, optionally, the second calculation module 503 may be specifically configured to calculate the dynamic adjustment error corresponding to the i-th moment according to the following expression:

[0103]

[0104] Among them, δ i Indicates the exposure difference, E i represents the target exposure, μ represents the preset smoothing coefficient, α represents the preset reference error rate, and e i-1 represents the dynamic adjustment error corresponding to the i-1th moment in the current time period, where when i=1, e i-1 is 0.

[0105] According to some embodiments of the present application, optionally, the adjustment module 504 can be specifically used to calculate the product of the learning step and the dynamic adjustment error corresponding to the i-th moment; calculate the difference between the initial delivery threshold and the product to obtain the dynamic delivery threshold.

[0106] According to some embodiments of the present application, optionally, the position exposure smoothing device 50 based on proportional integral control provided in the embodiments of the present application may also include a prediction module for responding to a position recommendation request of a target job seeker, inputting the portrait features of the target job seeker and the position features of the target recruitment position into a trained position recommendation model, and obtaining an estimated score of the degree of match between the target job seeker and the target recruitment position predicted by the position recommendation model.

[0107] Among them, the profile characteristics of the target job seeker include at least one of the following: the target job seeker's education, major, industry, age, gender, identity, work experience, expected job function, expected work area, expected salary and expected job type; the position characteristics of the target recruitment position include at least one of the following: the function, industry, education requirements, major requirements, work experience requirements, salary, work area, job type, age requirements and gender requirements of the target recruitment position.

[0108] Figure 5 Each module / unit in the device shown has the function of implementing each step in the position exposure smoothing method based on proportional integral control provided by the above method embodiment, and can achieve its corresponding technical effect. For the sake of concise description, it will not be repeated here.

[0109] 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 6 This is a hardware structure diagram of an electronic device of an embodiment of the present application. The electronic device may include a processor 601 and a memory 602 storing computer program instructions. When the processor 601 executes the computer program instructions, it implements the process or function of any of the above-mentioned embodiment methods.

[0110] Specifically, the processor 601 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 602 may include a large-capacity memory for data or instructions. For example, the memory 602 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 602 may include a removable or non-removable (or fixed) medium. For another example, the memory 602 may be inside or outside the integrated gateway disaster recovery device. The memory 602 may be a non-volatile solid-state memory. In other words, the memory 602 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 operation described in the method of the embodiment of the present application can be performed. The processor 601 implements the process or function of any method in the above embodiments by reading and executing the computer program instructions stored in the memory 602.

[0111] In one example, Figure 6 The electronic device shown may also include a communication interface 603 and a bus 610. Among them, the processor 601, the memory 602, and the communication interface 603 are connected through the bus 610 and complete the communication between each other. The communication interface 603 is mainly used to realize the communication between each module, device, unit and / or device in the embodiment of the present application. The bus 610 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 610 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.

[0112] 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.

[0113] 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.

[0114] 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.

[0115] 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.

[0116] 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 position exposure smoothing method based on proportional integral control, characterized in that: include: Determine the target exposure of the target job position in the current time period according to the total exposure budget of the target job position in a cycle and the exposure ratio of each time period in a preset cycle, wherein in a cycle, the exposure ratio of at least two non-adjacent time periods is higher than the exposure ratio of other time periods; At the i-th moment in the current time period, obtain the actual exposure of the target job position in the current time period, and calculate the exposure difference between the target exposure and the actual exposure; wherein the current time period includes n moments, and the interval between any two adjacent moments is a preset time length, 1≤i≤n, n is an integer greater than 1, and i is a positive integer; Based on the proportional-integral control algorithm, and according to the exposure difference and the target exposure, the dynamic adjustment error corresponding to the i-th moment is calculated; The preset initial sending threshold is adjusted based on the preset learning step size and the dynamic adjustment error corresponding to the i-th moment to obtain the dynamic sending threshold; When the estimated match score between the target job seeker and the target recruitment position is greater than or equal to the dynamic delivery threshold, the target recruitment position is pushed to the target job seeker in the current time period; when the estimated match score is less than the dynamic delivery threshold, the target recruitment position is not pushed to the target job seeker in the current time period or the target recruitment position is pushed to the target job seeker according to the preset delivery probability, where the preset delivery probability is less than 100%.

2. The method according to claim 1, characterized in that The at least two non-adjacent time periods include a first time period and a second time period, the first time period is a time period in the morning, and the second time period is a time period in the afternoon or evening.

3. The method according to claim 1, characterized in that Before adjusting the preset initial delivery threshold based on the preset learning step size and the dynamic adjustment error corresponding to the i-th moment to obtain the dynamic delivery threshold, the method further includes: Obtain the historical exposure ratio of the target grid under different historical matching degree estimation scores; different historical matching degree estimation scores correspond to different historical exposure ratios, the target grid is a grid composed of the work location of the target recruitment position and the function corresponding to the target recruitment position, the historical matching degree estimation is divided into the matching degree estimation score between the job seekers in the historical period and the recruitment positions under the target grid, and the historical exposure ratio is the exposure ratio of multiple recruitment positions in the target grid in the historical period; The estimated historical matching score corresponding to the target historical exposure ratio is used as the initial delivery threshold.

4. The method according to claim 1, characterized in that: Based on the proportional integral control algorithm, and according to the exposure difference and the target exposure, the dynamic adjustment error corresponding to the i-th moment is calculated, including: According to the following expression, the dynamic adjustment error corresponding to the i-th moment is calculated: Among them, δ i Indicates the exposure difference, E i represents the target exposure, μ represents the preset smoothing coefficient, α represents the preset reference error rate, and e i-1 represents the dynamic adjustment error corresponding to the i-1th moment in the current time period, where when i=1, e i-1 is 0.

5. The method according to claim 1, characterized in that The preset initial delivery threshold is adjusted based on the preset learning step size and the dynamic adjustment error corresponding to the i-th moment to obtain the dynamic delivery threshold, including: Calculate the product of the learning step size and the dynamic adjustment error corresponding to the i-th moment; The difference between the initial sending threshold and the product is calculated to obtain the dynamic sending threshold.

6. The method according to claim 1, characterized in that Before determining whether the estimated score of the matching degree between the target job seeker and the target job position is greater than or equal to the dynamic delivery threshold, the method further includes: In response to a job recommendation request from a target job seeker, input the target job seeker's profile features and the target job features into a trained job recommendation model to obtain an estimated score of the matching degree between the target job seeker and the target job predicted by the job recommendation model; The target job seeker's profile features include at least one of the following: the target job seeker's education, major, industry, age, gender, identity, work experience, expected job function, expected work area, expected salary and expected job type; The job characteristics of the target recruitment position include at least one of the following: the function, industry, educational requirements, professional requirements, work experience requirements, salary, work area, job type, age requirements and gender requirements of the target recruitment position.

7. A position exposure smoothing device based on proportional integral control, characterized in that: include: A determination module, used to determine the target exposure of the target job position in the current time period according to the total exposure budget of the target job position in a cycle and the exposure ratio of each time period in a preset cycle, wherein in a cycle, the exposure ratio of at least two non-adjacent time periods is higher than the exposure ratio of other time periods; The first calculation module is used to obtain the actual exposure of the target job position in the current time period at the i-th moment in the current time period, and calculate the exposure difference between the target exposure and the actual exposure; wherein the current time period includes n moments, and the interval between any two adjacent moments is a preset time length, 1≤i≤n, n is an integer greater than 1, and i is a positive integer; A second calculation module is used to calculate the dynamic adjustment error corresponding to the i-th moment based on the proportional integral control algorithm and according to the exposure difference and the target exposure; An adjustment module, used to adjust a preset initial sending threshold based on a preset learning step size and a dynamic adjustment error corresponding to the i-th moment to obtain a dynamic sending threshold; The sending module is used to push the target recruitment position to the target job seeker in the current time period when the estimated matching score between the target job seeker and the target recruitment position is greater than or equal to the dynamic sending threshold; when the estimated matching score is less than the dynamic sending threshold, the target recruitment position is not pushed to the target job seeker in the current time period or the target recruitment position is pushed to the target job seeker according to a preset sending probability, wherein the preset sending probability is less than 100%.

8. 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 position exposure smoothing method based on proportional integral control as described in any one of claims 1 to 6 is implemented.

9. 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 position exposure smoothing method based on proportional integral control as described in any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that It includes computer program instructions, which, when executed by a processor, implement the position exposure smoothing method based on proportional integral control as described in any one of claims 1 to 6.

Images