A method, device and storage medium for determining a recall post

Abstract

This invention discloses a method, apparatus, device, and storage medium for determining recall positions, relating to the field of computer technology. The method includes: determining the user's corresponding user profile bucket from a pre-divided set of user profile buckets based on the user's basic information; searching a job category migration probability table based on the user's initial job category and the user profile bucket to determine the user's corresponding candidate job categories and the migration probability of each candidate job category; filtering the candidate job categories based on probability thresholds and the migration probabilities of each candidate job category to obtain migration job categories; and determining the positions corresponding to the initial job category and the migration job category as the recall positions corresponding to the user. This technical solution, by expanding the user's effective job categories, more accurately filters target job categories that the user is truly likely to accept, determines recall positions that are more suitable for the user, and reduces the recall of invalid positions.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method, apparatus, device and storage medium for determining recall positions. Background Technology

[0002] The recruitment industry is currently undergoing rapid development and transformation. With the rise of the internet and the wave of entrepreneurship, various industries are facing industrial upgrading, and the key to competition among companies lies in the competition for talent. Therefore, talent recruitment has become a top priority for all companies. Recall strategies are a crucial part of job recommendations; subsequent job filtering, sorting, and diversified display all rely on the jobs recalled in the early stages.

[0003] In existing technologies, rule-based matching recall is commonly used. One approach is to expand the recall based on a manually configured job category similarity table. For example, "Java Developer" and "Architect" are labeled as similar job categories. When a user's desired job category is "Java Developer," "Architect" related positions are also recalled. However, this static similarity table-based approach cannot reflect the differentiated migration intentions of different user groups. For instance, junior developers with less work experience may have significantly different migration intentions for the "Architect" position compared to senior developers, but a static similarity table cannot distinguish these differences. Another approach is recall methods based on collaborative filtering or vector similarity, which recall positions by calculating the similarity between user vectors and job vectors. However, this approach is a black-box recall method, with poor interpretability of the recall results, making it difficult for business-side strategy adjustment and threshold configuration.

[0004] Therefore, this application proposes a method for determining recall positions, which identifies recall positions with a higher degree of matching with users. Summary of the Invention

[0005] This invention provides a method, apparatus, device, and storage medium for determining recall positions, in order to identify recall positions that are more suitable for users.

[0006] In a first aspect, embodiments of the present invention provide a method for determining recall positions, including: Based on the user's basic information, the user profile bucket corresponding to the user is determined from multiple pre-divided profile buckets, wherein the profile bucket is determined based on a combination of multiple target user features that have an impact on job category migration; Based on the user's initial job category and the user profile bucketing, the candidate job categories corresponding to the user and the migration probability of each candidate job category are determined by searching the job category migration probability table. The candidate job categories are filtered based on probability thresholds and the migration probability of each candidate job category to obtain migration job categories; The positions corresponding to the initial job category and the migration job category are determined as the recall positions corresponding to the user.

[0007] The technical solution of this invention provides a method for determining recall positions, comprising: determining a user profile bucket corresponding to the user in a pre-divided plurality of profile buckets based on the user's basic information, wherein the profile bucket is determined based on a combination of multiple target user features that may affect job category migration; searching a job category migration probability table based on the user's initial job category and the user profile bucket to determine the candidate job categories corresponding to the user and the migration probability of each candidate job category; filtering the candidate job categories based on probability thresholds and the migration probabilities of each candidate job category to obtain migration job categories; and determining the positions corresponding to the initial job category and the migration job categories as the recall positions corresponding to the user. The above technical solution first determines the user's corresponding user profile bucket based on the user's basic information and multiple pre-divided profile buckets, thus achieving user profile bucketing. Second, it searches a job category migration probability table based on the user's initial job category and user profile bucket to determine the user's corresponding candidate job categories and the migration probability of each candidate job category. This table identifies the candidate job categories to which the user has potential choices and the migration probabilities between the user's initial job and each candidate category. Then, it determines the migration job category from the candidate job categories, thus expanding the user's effective job categories and more accurately filtering the target job categories that the user is truly likely to accept, thereby reducing invalid recall. Finally, it identifies the job corresponding to the user's initial job category and migration job category as the corresponding recall job, thus achieving recall job determination. Furthermore, by moving the complex calculation process offline, with only indexing and filtering operations performed online, the job recall response time is significantly reduced, making it suitable for handling high-concurrency requests in large-scale recruitment systems.

[0008] Furthermore, before determining the user profile bucket corresponding to the user among the pre-divided multiple profile buckets based on the user's basic information, the process also includes: Filter target user characteristics that have an impact on job category migration from user characteristics; The feature dimensions of each target user feature are determined, and the feature dimensions of each target user feature are combined to obtain multiple profile buckets.

[0009] Further, based on the user's basic information, the user profile bucket corresponding to the user is determined from multiple pre-divided profile buckets, including: Determine user characteristics based on the user's basic information; The user profile bucket corresponding to the user is determined by comparing the feature dimensions of the target user features in the user features with the feature dimensions of the target user features that make up each profile bucket.

[0010] Furthermore, it also includes: For each of the aforementioned profile buckets, determine the migration probability of users in each job category under that profile bucket; Based on each of the aforementioned profile buckets and the migration probability of users within each of the aforementioned profile buckets across different job categories, a job category migration probability table is constructed.

[0011] Further, determining the migration probability of users across different job categories under the aforementioned user profile bucketing includes: Each job category in the job category set is combined with other job categories to obtain multiple sets of starting job categories and target job categories; For each initial job category and target job category, the feature dimensions of the target user features corresponding to the profile bucket, as well as the initial job category and target job category, are input into a pre-trained migration probability model so that the migration probability model can determine the migration probability of a user in the profile bucket from the initial job category to the target job category.

[0012] Furthermore, when the target job category is the job category corresponding to a less popular job, determining the migration probability of users between job categories under the user profile bucketing includes: The migration probability determined by the migration probability model is amplified by an amplification factor to obtain the target migration probability.

[0013] Furthermore, when the candidate job category corresponds to a less popular job, the candidate job categories are filtered based on a probability threshold and the migration probability of each candidate job category to obtain migration job categories, including: The probability threshold is reduced by a reduction factor to obtain the target probability threshold; If the migration probability of the candidate job category is greater than the target probability threshold, then the candidate job category is determined as the migration job category.

[0014] Secondly, embodiments of the present invention also provide a recall job determination device, comprising: The determination module is used to determine the user profile bucket corresponding to the user from multiple pre-divided profile buckets based on the user's basic information, wherein the profile bucket is determined based on a combination of multiple target user features that have an impact on job category migration; The search module is used to search in the job category migration probability table according to the user's initial job category and the user profile binning, to determine the candidate job categories corresponding to the user and the migration probability of each candidate job category; A filtering module is used to filter the candidate job categories based on a probability threshold and the migration probability of each candidate job category to obtain migration job categories; The execution module is used to determine the positions corresponding to the starting job category and the migration job category as the recall positions corresponding to the user.

[0015] Thirdly, embodiments of the present invention also provide an electronic device, the electronic device comprising: At least one processor; and a memory communicatively connected to said at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the recall job determination method as described in any of the first aspects.

[0016] Fourthly, embodiments of the present invention also provide a storage medium containing computer-executable instructions, characterized in that the computer-executable instructions, when executed by a computer processor, are used to perform the recall job determination method as described in any of the first aspects.

[0017] Fifthly, this application provides a computer program product including computer instructions that, when executed on a computer, cause the computer to perform the recall job determination method provided in the first aspect.

[0018] It should be noted that the aforementioned computer instructions may be stored, in whole or in part, on a computer-readable storage medium. This computer-readable storage medium may be packaged together with the processor of the recall post determination device, or it may be packaged separately from the processor of the recall post determination device; this application does not impose any limitations on this.

[0019] The descriptions of the second, third, fourth, and fifth aspects in this application can be referred to the detailed description of the first aspect; and the beneficial effects of the descriptions of the second, third, fourth, and fifth aspects can be referred to the analysis of the beneficial effects of the first aspect, which will not be repeated here.

[0020] In this application, the name of the aforementioned recall job determination device does not limit the equipment or functional module itself. In actual implementation, these devices or functional modules may appear under other names. As long as the function of each device or functional module is similar to that of this application, it falls within the scope of the claims of this application and its equivalents.

[0021] These or other aspects of this application will become more readily apparent in the following description. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 A flowchart of a method for determining recall positions provided in an embodiment of the present invention; Figure 2 A flowchart of another method for determining recall positions provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of a recall job determination device provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0024] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0025] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0026] The terms "first" and "second," etc., used in the specification and drawings of this application are used to distinguish different objects or to distinguish different treatments of the same object, rather than to describe a specific order of objects.

[0027] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus.

[0028] Before discussing the exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the figures. The process can correspond to a method, function, procedure, subroutine, subroutine, etc. Moreover, embodiments and features in the embodiments of the present invention can be combined with each other without conflict.

[0029] It should be noted that in the embodiments of this application, the words "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0030] In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0031] Figure 1 This is a flowchart of a method for determining a recall position provided by an embodiment of the present invention. This embodiment is applicable to situations where job recommendations need to be made to users. This method can be executed by a recall position determination device, such as... Figure 1 As shown, the specific steps include the following: Step 110: Determine the user profile bucket corresponding to the user from the pre-divided multiple profile buckets based on the user's basic information.

[0032] The user's basic information includes, but is not limited to, years of work experience, number of skills, frequency of job hopping, salary trends, and whether they are a recent graduate. User characteristics are determined by feature extraction from the user's basic information, and may include, for example, years of work experience, skill diversity, and salary level.

[0033] User profile binning is determined by combining multiple target user features that influence job category migration. Specifically, after identifying the target user features that influence job category migration, multiple feature dimensions are determined for each target user feature. Combining these target user features along these feature dimensions yields multiple user profile bins. Therefore, the feature dimensions of the target user features constituting each user profile bin are different, and the user profile bin corresponding to a user can be determined based on the user features they possess.

[0034] Specifically, the user characteristics can be determined based on the user's basic information. Specifically, the user's years of work experience, skill diversity, job-hopping frequency, salary level, and whether they are a recent graduate can be determined. Then, the feature dimensions of each target user feature in the user characteristics can be determined. Based on the feature dimensions of each target user feature in the user characteristics and the feature dimensions of the target user features that make up each profile bucket, the user profile bucket corresponding to the user can be determined, thus realizing the determination of the profile bucket to which the user belongs.

[0035] In this embodiment of the invention, the user's profile bucket is determined based on the user's basic information, thereby realizing the user's profile bucket division.

[0036] Step 120: Based on the user's initial job category and the user profile bucketing, search the job category migration probability table to determine the candidate job categories corresponding to the user and the migration probability of each candidate job category.

[0037] Job categories are classification tags for positions in a recruitment system, belonging to a hierarchical classification system (major category → intermediate category → minor category). For example, Java Development Engineer, Front-end Development Engineer, and Product Manager. Positions are specific job postings published by the recruiter. Each position belongs to a specific job category, and a job category can contain multiple specific positions. For example, "XX Company - Senior Java Development Engineer (Beijing)" is a specific position belonging to the Java Development Engineer job category. The starting job category can be understood as the user's current job category, specifically derived from the job category corresponding to the user's current or desired position in their resume. For example, if a user's current position in their resume is Mold Engineer, then their starting job category is Mold Engineer.

[0038] The job category migration probability table consists of the migration probabilities between job categories under each profile bucket. It can include the migration probabilities between job categories that serve as the starting job category and job categories that serve as the target job category within each profile bucket. Of course, to ensure the validity of the data in the job category migration probability table, data with migration probabilities lower than a basic threshold can be removed. That is, if the migration probability between any two job categories under any profile bucket is lower than the basic threshold, then that data entry is removed from the job category migration probability table.

[0039] Specifically, based on the user's initial job category and user profile bucketing, a search can be performed in the job category migration probability table to determine the corresponding candidate job categories and the migration probability of each candidate job category. The migration probability between the initial job category and candidate job categories under the user's profile bucket is not less than a basic threshold. Therefore, candidate job categories can be understood as the job categories to which the user has possible job choices.

[0040] In this embodiment of the invention, the candidate job categories to which a user has possible job choices and the migration probabilities between the user's initial job and each candidate job category are determined in the job category migration probability table.

[0041] Step 130: Filter the candidate job categories based on the probability threshold and the migration probability of each candidate job category to obtain the migration job categories.

[0042] The probability threshold is used to filter candidate job categories.

[0043] Specifically, the probability threshold can be compared with the migration probability of each candidate job category. If the migration probability of a candidate job category is greater than the probability threshold, then the candidate job category is determined as a migration job category. Since the probability between a user's starting job and the migration job category is greater than the probability threshold, the migration job category is understood as the job category to which the job can be recalled belongs.

[0044] In this embodiment of the invention, by determining the migration job category from the candidate job categories, the user's migration job category under the initial job category is determined, thereby realizing the effective job category expansion for the user.

[0045] Step 140: Determine the positions corresponding to the starting job category and the migration job category as the recall positions corresponding to the user.

[0046] Specifically, after determining the job category to be migrated, the user's initial job category and the corresponding positions for the job category to be migrated can be determined, and these positions can be identified as the corresponding recall positions for the user.

[0047] In this embodiment of the invention, the positions corresponding to the user's initial job category and the job category they migrate to are determined as the recall positions corresponding to the user, thereby realizing the determination of the recall positions.

[0048] The method for determining recall positions provided in this invention includes: determining a user profile bucket corresponding to the user from a pre-divided plurality of profile buckets based on the user's basic information, wherein the profile bucket is determined based on a combination of multiple target user features that may affect job category migration; searching a job category migration probability table based on the user's initial job category and the user profile bucket to determine the candidate job categories corresponding to the user and the migration probability of each candidate job category; filtering the candidate job categories based on probability thresholds and the migration probabilities of each candidate job category to obtain migration job categories; and determining the positions corresponding to the initial job category and the migration job category as the recall positions corresponding to the user. The above technical solution first determines the user's corresponding user profile bucket based on the user's basic information and multiple pre-divided profile buckets, thus achieving user profile bucketing. Second, it searches a job category migration probability table based on the user's initial job category and user profile bucket to determine the user's corresponding candidate job categories and the migration probability of each candidate job category. This table identifies the candidate job categories to which the user has potential choices and the migration probabilities between the user's initial job and each candidate category. Then, it determines the migration job category from the candidate job categories, thus expanding the user's effective job categories and more accurately filtering the target job categories that the user is truly likely to accept, thereby reducing invalid recall. Finally, it identifies the job corresponding to the user's initial job category and migration job category as the corresponding recall job, thus achieving recall job determination. Furthermore, by moving the complex migration probability calculation process offline, with only indexing and filtering operations performed online, the job recall response time is significantly reduced, making it suitable for high-concurrency request processing in large-scale recruitment systems.

[0049] Figure 2 This is a flowchart illustrating another method for determining recall positions provided by an embodiment of the present invention. This embodiment is a specific modification based on the above embodiments. Figure 2 As shown, in this embodiment, the method may further include: Step 210: Filter target user features that have an impact on job category migration from user features; determine the feature dimensions of each target user feature, and combine each feature dimension of each target user feature to obtain multiple profile buckets.

[0050] Specifically, user characteristics can include years of work experience, skill diversity, job-hopping frequency, salary level, and whether the user is a recent graduate. Target user characteristics are those that influence job category migration and can include years of work experience, skill diversity, and salary level. After identifying target user characteristics that influence job category migration from the user characteristics, the feature dimensions of each target user characteristic can be determined. This means dividing each target user characteristic into multiple segments. For example, years of work experience can be divided into four segments: 0-2 years, 3-5 years, 6-10 years, and over 10 years; skill diversity into three segments: low, medium, and high; and salary level into three segments: low, medium, and high. Then, the feature dimensions of each target user characteristic can be combined to obtain 4 × 3 × 3 = 36 user profile buckets. Users within the same user profile bucket share the same feature dimensions in terms of target user characteristics and share the same set of job category migration probabilities. This achieves a balance between personalization accuracy and storage / query efficiency, avoiding the need to use the same migration probabilities for all users and the need to store a separate migration probability table for each user.

[0051] In this embodiment of the invention, user profiles are divided into buckets for all users within the recruitment system.

[0052] Step 220: For each of the aforementioned profile buckets, determine the migration probability of users in each job category under the aforementioned profile bucket; construct the job category migration probability table based on each of the aforementioned profile buckets and the migration probability of users in each of the aforementioned profile buckets.

[0053] In one implementation, determining the migration probability of a user across different job categories under the user profile bucketing includes: Each job category in the job category set is combined with other job categories to obtain multiple sets of starting job categories and target job categories. For each set of starting job categories and target job categories, the feature dimensions of the target user features corresponding to the profile bucket, as well as the starting job category and target job category, are input into a pre-trained migration probability model so that the migration probability model can determine the migration probability of a user under the profile bucket from the starting job category to the target job category.

[0054] Specifically, the job category set can include all job categories within the recruitment system. Combining each job category with other job categories yields multiple sets of starting and target job categories. The feature dimensions of the target user characteristics corresponding to the user profile bucket, along with the starting and target job categories, are input into a pre-trained migration probability model. This model determines the migration probability of a user moving from the starting job category to the target job category under that user profile bucket. The migration probability between job categories reflects whether a user has the willingness to migrate across job categories, the migration differences under different user profiles, and the order of migration strength from the same job category to different target job categories.

[0055] A job category migration probability table can be constructed based on each user profile bucket and the migration probability of users within each user profile bucket across different job categories. This table can filter out job category combinations with migration probabilities below a basic threshold, retaining only meaningful migration paths. The job category migration probability table can be a structured Key-Value table, where the Key is (user profile bucket ID and starting job category), and the Value is the target job category and its corresponding migration probability, resulting in the job category migration probability table shown in Table 1.

[0056] Table 1 Among them, the profile bucket ID can be understood as the profile bucket identifier, and the job attribute can be used to identify unpopular jobs.

[0057] In one implementation, when the target job category is the job category corresponding to a less popular job, determining the migration probability of a user between job categories under the user profile bucketing includes: The migration probability determined by the migration probability model is amplified by an amplification factor to obtain the target migration probability.

[0058] In one scenario, to increase the exposure of less popular positions and alleviate the uneven exposure of positions within the recruitment system, a position attribute adjustment mechanism can be introduced. When a target job category is identified as belonging to a less popular job category, its corresponding migration probability can be amplified. Specifically, a less popular label can be set for low-exposure positions. When a target job category is identified as belonging to a less popular job category, the migration probability determined by the migration probability model corresponding to the target job category can be amplified. This can be achieved by multiplying the migration probability by the amplification factor β (β > 1), resulting in the amplified target migration probability P' = P × β, where P represents the migration probability and P' represents the target migration probability. This mechanism effectively increases the exposure of less popular positions by controllably amplifying the migration probability corresponding to the job category of less popular positions without compromising overall recall relevance.

[0059] For example, the migration probability of a less popular job category is P=0.07, the amplification factor is β=2.0, and the target migration probability is P'=0.07×2.0=0.14>0.10.

[0060] Of course, the migration probability model can be retrained and the migration probability table updated according to the latest user behavior data within a preset period (such as every week or every two weeks), so that the job category migration probability can be automatically updated with market changes and user behavior changes, avoiding long-term reliance on manual maintenance of the job category rule table.

[0061] The training process of the transfer probability model includes: 1) Obtaining training samples: Taking users as the main body, construct sample triplets (user identifier, starting job category, target job category). Within a preset time window, if the user generates effective behavior towards a job in the target job category, a positive sample is constructed. If a job in the target job category exposes the user but the user does not generate effective behavior, a negative sample is constructed. Jobs that do not expose the user are not considered negative samples to avoid introducing invalid noise. 2) Constructing features: For each training sample, construct the following feature set: User features, including but not limited to years of work experience, skill diversity, job-hopping frequency, salary level, and whether the user is a recent graduate; Target job category features, including the number of jobs, industry distribution, city coverage, and historical application conversion rate; Job category difference features, including job category level difference, skill overlap ratio, salary difference, and artificial job category similarity level; User behavior features, including exposure count, click count, application count, and behavior time distribution; Job attribute features, including whether the job promotes employment, whether it is a niche job, and job stability indicators. 3) Data preprocessing: Normalize or discretize continuous features (such as years of work experience, salary differences, etc.), encode categorical features (such as industry, city, etc.), and impute missing values ​​(e.g., impute based on the median or mode of user features). 4) Sample balancing: Since positive samples are usually far fewer than negative samples, resulting in severe sample imbalance, negative samples can be downsampled to bring the positive-to-negative sample ratio to a reasonable range (e.g., 1:3 to 1:5). 5) Model training: Use a gradient boosting decision tree (such as LightGBM) as the binary classification model, with job category migration as the supervision signal. The loss function is the binary cross-entropy, and the model output is the probability value of a user migrating from the initial job category to the target job category. Model hyperparameters (such as tree depth, learning rate, number of leaf nodes, etc.) are determined through five-fold cross-validation. 6) Model evaluation: Use AUC (Area Under the ROC Curve) as the primary evaluation metric, while also referencing precision and recall to evaluate recall performance at different probability thresholds.

[0062] It should be noted that effective behavior refers to proactive and positive user interactions with a job posting, specifically including the following types: Application submission: The user submits their resume for the job (strongest positive signal); Deep click behavior: The user clicks on the job details page and stays for more than a preset time (e.g., more than 10 seconds); Favoriting behavior: The user saves the job posting to their favorites; Proactive communication behavior: The user initiates online communication with the recruiter. Additionally, the following situations are not considered effective behavior: The job posting only appears in the user's search results list but the user does not click (exposure only); the user quickly clicks and immediately returns; the dwell time is less than a preset threshold (accidental click).

[0063] The hierarchical difference between job categories can be understood as the shortest path distance between the starting job category and the target job category in the corresponding job category classification tree of the recruitment system. For example, two different subcategories belonging to the same "IT / Technology" category have a hierarchical difference of 2 (same category, different subcategories); one belongs to the "IT / Technology" category and the other belongs to the "Marketing / Marketing" category, with a hierarchical difference of 4 (across categories). The smaller the hierarchical difference, the closer the two job categories are in the system's classification.

[0064] The skill overlap ratio between job categories can be calculated using the Jaccard similarity coefficient. For example, we can extract the typical skill set A associated with the starting job category and the typical skill set B associated with the target job category (the skill sets are extracted from job descriptions and user resumes under that job category), and determine the skill overlap ratio = |A∩B| / |A∪B|. For example, if the skill set of the starting job category "Java Development" is {Java, Spring, MySQL, Redis, Linux}, and the skill set of the target job category "Architect" is {Java, Spring, Distributed Systems, Microservices, MySQL}, then the intersection is {Java, Spring, MySQL}, and the union is {Java, Spring, MySQL, Redis, Linux, Distributed Systems, Microservices}, with a skill overlap ratio of 3 / 7.

[0065] Salary differences between job categories can be determined by comparing the median salary of the target job category with the median salary of the starting job category. The formula is: Salary difference = Median salary of the target job category / Median salary of the starting job category. A ratio greater than 1 indicates migration to a higher-paying job category, while a ratio less than 1 indicates the opposite.

[0066] The artificial similarity level between job categories can be determined by the artificially configured job category similarity relationship table maintained in the recruitment system. The job category similarity relationship table marks the similarity level between different job categories (e.g., level 1-5, with 5 being the most similar).

[0067] Additionally, the number of exposures in user behavior refers to the total number of times a job posting appears in various display scenarios such as search results lists and recommendation lists within a preset time window. Specifically, a job posting appearing on a user-visible interface (regardless of whether the user clicks on it) counts as one exposure. The number of exposures reflects the degree to which users are exposed to the job posting. If the number of exposures is high but the user does not take any effective action, it indicates that the user's willingness to switch to the job posting may be low; conversely, if the number of exposures is low, the user's lack of action may simply be due to not seeing the posting.

[0068] Job stability indicators reflect the continued activity and authenticity of a job on the recruitment system. Specifically, they can be measured by the following metrics: Continuous online days: the time span from job posting to delisting; Update frequency: whether the recruiter regularly refreshes the job posting; Recruiter's historical activity: the overall frequency of the recruiter's recruitment activities on the platform. Highly stable jobs typically represent genuine, long-term recruitment needs; low-stability jobs may be temporary, expired, or fraudulent. Job stability indicators help the model distinguish between genuinely valid and noisy jobs, avoiding the misinterpretation of user behavior on low-quality jobs as reliable migration signals.

[0069] In this embodiment of the invention, a job category migration probability table is constructed based on each user profile bucket and the migration probability of users within each user profile bucket across different job categories. By introducing user profile buckets to predict job category migration probabilities, personalized migration prediction is achieved. This allows the migration probability of the same pair of job categories to vary across different user profile buckets, unlike the traditional static similarity table between job categories. By pre-constructing the job category migration probability table, the calculation process of job category migration probabilities can be completed entirely offline, and the results can be stored in a structured table format, enabling quantitative modeling of job category migration probabilities.

[0070] Step 230: Determine the user profile bucket corresponding to the user from the pre-divided multiple profile buckets based on the user's basic information.

[0071] In one implementation, step 230 may specifically include: User characteristics are determined based on the user's basic information; the user profile bucket corresponding to the user is determined by comparing the feature dimensions of the target user characteristics in the user characteristics with the feature dimensions of the target user characteristics that make up each profile bucket.

[0072] As mentioned earlier, the user characteristics can first be determined based on the user's basic information. Specifically, this includes determining the user's work experience segments, skill diversity, job-hopping frequency, salary level, and whether they are a recent graduate. Then, the target user characteristics and feature dimensions of each target user characteristic can be identified. The feature dimensions of the target user characteristics in the user characteristics are compared with the feature dimensions of the target user characteristics in each user profile bucket. The user profile bucket that matches the feature dimensions of the target user characteristics in the user characteristics is determined as the user profile bucket to which the user belongs, thus identifying the user's user profile bucket.

[0073] In this embodiment of the invention, the user's profile bucket is determined based on the user's basic information, thereby realizing the user's profile bucket division.

[0074] Step 240: Based on the user's initial job category and the user profile bucketing, search the job category migration probability table to determine the candidate job categories corresponding to the user and the migration probability of each candidate job category.

[0075] Specifically, the user profile bucket ID corresponding to the user profile bucket can be determined. The user profile bucket ID and the starting job category are used as the key. The corresponding value is then searched in the job category migration probability table to obtain the user's candidate job categories and the migration probability of each candidate job category.

[0076] Compared to black-box recall methods based on vector similarity, the migration probability determined in this application has good interpretability, which facilitates strategy debugging, threshold configuration, and business intervention.

[0077] In this embodiment of the invention, the candidate job categories to which a user has possible job choices and the migration probabilities between the user's initial job and each candidate job category are determined in the job category migration probability table.

[0078] Step 250: Filter the candidate job categories based on the probability threshold and the migration probability of each candidate job category to obtain the migration job categories.

[0079] The probability threshold is a configurable parameter set according to business needs. A higher probability threshold results in more accurate recall but narrower coverage, while a lower probability threshold results in broader coverage but may introduce weakly related job categories.

[0080] Specifically, a probability threshold can be compared with the migration probability of each candidate job category. If the migration probability of a candidate job category is greater than the probability threshold, then the candidate job category is determined as a migration job category. Since the probability between a user's initial job and the migration job category is greater than the probability threshold, the migration job category is understood as the job category to which the recalled job belongs. If the migration probability of a candidate job category is not greater than the probability threshold, then the job corresponding to the candidate job category is determined not to participate in the recall.

[0081] In one implementation, when the candidate job category is the job category corresponding to a less popular job, step 250 may specifically include: The probability threshold is reduced by a reduction factor to obtain a target probability threshold; if the migration probability of the candidate job category is greater than the target probability threshold, then the candidate job category is determined as the migration job category.

[0082] In another scenario, to increase the exposure of less popular positions and alleviate the uneven exposure of positions within the recruitment system, a job attribute adjustment mechanism can be introduced. When the target job category is determined to be a less popular job category, the probability threshold used for screening can be reduced. Specifically, when the target job category is determined to be a less popular job category, the probability threshold can be reduced. This can be achieved by reducing the probability threshold using a reduction factor, resulting in a target probability threshold. Specifically, the probability threshold can be multiplied by the reduction factor α (α < 1) to obtain the reduced target probability threshold T' = T × α, where T represents the probability threshold and T' represents the target probability threshold. This mechanism effectively increases the exposure of less popular positions without compromising overall recall relevance.

[0083] For example, when the probability threshold T=0.10 and the target probability threshold T'=0.05, the migration probability of the target job category corresponding to the unpopular job is 0.07. If the target job category is filtered according to the probability threshold, it will be filtered out. If the target job category is filtered according to the target probability threshold, it can be retained.

[0084] In practical applications, the exposure opportunities for less popular jobs can be increased by either reducing the probability threshold or increasing the migration probability.

[0085] In this embodiment of the invention, a migration job category is determined from the candidate job categories, thereby determining the migration job category of the user under the initial job category and realizing the effective job category expansion for the user.

[0086] Step 260: Determine the positions corresponding to the starting job category and the migration job category as the recall positions corresponding to the user.

[0087] Specifically, after determining the job category to be migrated, the user's initial job category and the corresponding positions for the job category to be migrated can be determined, and these positions can be identified as the corresponding recall positions for the user.

[0088] Existing recruitment systems typically have multiple parallel recall channels, such as: rule-based recall (directly filtering positions based on user expectations (city, salary, education, etc.), similar job category recall (including positions in similar categories to the user's initial job category based on a manually configured job category similarity table), and collaborative filtering recall (recommending positions based on the behavior of similar users). The job determination method disclosed in this application can be integrated into the existing architecture as a new recall channel, working in parallel with existing recall channels. The results of all recall channels are ultimately aggregated, deduplicated, and sorted at the merging and sorting layer before being displayed to the user. This design allows for integration without significant modifications to the existing recruitment system, offering good engineering compatibility and low risk.

[0089] In this embodiment of the invention, the positions corresponding to the user's initial job category and the job category they migrate to are determined as the recall positions corresponding to the user, thereby realizing the determination of the recall positions.

[0090] The method for determining recall positions provided in this invention includes: screening target user features that have an impact on job category migration from user features; determining the feature dimensions of each target user feature, and combining each feature dimension of each target user feature to obtain multiple profile buckets; for each profile bucket, determining the migration probability of a user between job categories under the profile bucket; constructing a job category migration probability table based on each profile bucket and the migration probability of a user between job categories under each profile bucket; determining the user profile bucket corresponding to the user from the pre-divided multiple profile buckets based on the user's basic information; searching the job category migration probability table based on the user's initial job category and the user profile bucket to determine the candidate job categories corresponding to the user and the migration probability of each candidate job category; filtering the candidate job categories based on probability thresholds and the migration probabilities of each candidate job category to obtain migration job categories; and determining the positions corresponding to the initial job category and the migration job categories as the recall positions corresponding to the user. The above technical solution firstly involves creating user profiles and bucketing them across all users in the recruitment system. Based on each user profile bucket and the migration probability of users across different job categories within each bucket, a job category migration probability table is constructed. By introducing user profile bucketing, personalized migration prediction is achieved, allowing the migration probability of the same job category pair to vary across different user profile buckets. By pre-constructing the job category migration probability table, the calculation process is completed entirely offline, and the results are stored in a structured table, enabling quantitative modeling of job category migration probabilities. Based on the user's basic information and multiple pre-defined user profile buckets, the user's corresponding user profile bucket is determined, achieving user profile bucketing. The system then considers the user's initial job category and user profile bucket during job category migration. The algorithm searches a probability table to determine the user's candidate job categories and the migration probability of each category. This allows for the identification of the candidate job categories to which the user's potential choices belong, as well as the migration probabilities between the user's initial job and each candidate category. Compared to methods based on manual similarity tables or simple collaborative filtering, this approach more accurately filters candidate job categories that the user is likely to accept, thus reducing invalid recall. Within these candidate categories, the algorithm identifies the migration job category, thus expanding the user's effective job category pool and further accurately filtering the target job categories that the user is likely to accept, further reducing invalid recall. Finally, the job categories corresponding to the user's initial and migration job categories are identified as the corresponding recall job categories. Furthermore, by moving the complex calculation process offline, with only indexing and filtering operations performed online, the job recall response time is significantly reduced, making it suitable for handling high-concurrency requests in large-scale recruitment systems.

[0091] In addition, by introducing a job attribute weight adjustment mechanism, the exposure opportunities of less popular jobs can be increased while ensuring user relevance, thus alleviating the problem of uneven exposure.

[0092] Figure 3 This is a schematic diagram of a job selection device for recall provided by an embodiment of the present invention. This device can be applied to situations requiring job recommendations to users. The device can be implemented through software and / or hardware and is generally integrated into electronic devices, such as computer equipment.

[0093] like Figure 3 As shown, the device includes: The determination module 310 is used to determine the user profile bucket corresponding to the user in a pre-divided plurality of profile buckets based on the user's basic information, wherein the profile bucket is determined based on a combination of multiple target user features that have an impact on job category migration. The lookup module 320 is used to search in the job category migration probability table according to the user's initial job category and the user profile binning, to determine the candidate job categories corresponding to the user and the migration probability of each candidate job category; The filtering module 330 is used to filter the candidate job categories based on a probability threshold and the migration probability of each candidate job category to obtain the migration job categories; The execution module 340 is used to determine the positions corresponding to the starting job category and the migration job category as the recall positions corresponding to the user.

[0094] The recall job determination device provided in this embodiment determines the user profile bucket corresponding to the user from multiple pre-divided profile buckets based on the user's basic information. The profile buckets are determined based on a combination of target user features that may affect job category migration. The device then searches a job category migration probability table based on the user's initial job category and the user profile bucket to determine the candidate job categories corresponding to the user and the migration probability of each candidate job category. Based on probability thresholds and the migration probabilities of each candidate job category, the device filters the candidate job categories to obtain migration job categories. Finally, the job corresponding to the initial job category and the migration job category is determined as the recall job corresponding to the user. The above technical solution first determines the user's corresponding user profile bucket based on the user's basic information and multiple pre-divided profile buckets, thus achieving user profile bucketing. Second, it searches a job category migration probability table based on the user's initial job category and user profile bucket to determine the user's corresponding candidate job categories and the migration probability of each candidate job category. This table identifies the candidate job categories to which the user has potential choices and the migration probabilities between the user's initial job and each candidate category. Then, it determines the migration job category from the candidate job categories, thus expanding the user's effective job categories and more accurately filtering the target job categories that the user is truly likely to accept, thereby reducing invalid recall. Finally, it identifies the job corresponding to the user's initial job category and migration job category as the corresponding recall job, thus achieving recall job determination. Furthermore, by moving the complex calculation process offline, with only indexing and filtering operations performed online, the job recall response time is significantly reduced, making it suitable for handling high-concurrency requests in large-scale recruitment systems.

[0095] Based on the above embodiments, the device further includes: The combination module is used to filter target user features that have an impact on job category migration from user features; determine the feature dimensions of each target user feature; and combine the feature dimensions of each target user feature to obtain multiple profile buckets.

[0096] In one implementation, user characteristics are determined based on the user's basic information; The user profile bucket corresponding to the user is determined by comparing the feature dimensions of the target user features in the user features with the feature dimensions of the target user features that make up each profile bucket.

[0097] Based on the above embodiments, the device further includes: A construction module is used to determine the migration probability of users in each job category under each of the aforementioned profile buckets; and to construct the job category migration probability table based on each of the aforementioned profile buckets and the migration probability of users in each of the aforementioned profile buckets.

[0098] In one implementation, determining the migration probability of a user across different job categories under the user profile bucketing includes: Each job category in the job category set is combined with other job categories to obtain multiple sets of starting job categories and target job categories. For each set of starting job categories and target job categories, the feature dimensions of the target user features corresponding to the profile bucket, as well as the starting job category and target job category, are input into a pre-trained migration probability model so that the migration probability model can determine the migration probability of a user under the profile bucket from the starting job category to the target job category.

[0099] In one implementation, when the target job category is the job category corresponding to a less popular job, determining the migration probability of a user between job categories under the user profile bucketing includes: The migration probability determined by the migration probability model is amplified by an amplification factor to obtain the target migration probability.

[0100] Based on the above embodiments, when the candidate job category is a job category corresponding to a less popular job, the filtering module 330 is specifically used for: The probability threshold is reduced by a reduction factor to obtain a target probability threshold; if the migration probability of the candidate job category is greater than the target probability threshold, then the candidate job category is determined as the migration job category.

[0101] The recall job determination device provided in this embodiment of the invention can execute the recall job determination method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects for executing the recall job determination method.

[0102] It is worth noting that in the embodiments of the above-mentioned recall job determination device, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0103] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Figure 4 A block diagram of an exemplary electronic device 4 suitable for implementing embodiments of the present invention is shown. Figure 4 The electronic device 4 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0104] like Figure 4 As shown, electronic device 4 is represented in the form of a general-purpose computing electronic device. The components of electronic device 4 may include, but are not limited to: one or more processors or processing units 16, system memory 28, and bus 18 connecting different system components (including system memory 28 and processing unit 16).

[0105] Bus 18 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0106] Electronic device 4 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by electronic device 4, including volatile and non-volatile media, removable and non-removable media.

[0107] System memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. Electronic device 4 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 34 may be used to read and write non-removable, non-volatile magnetic media (… Figure 4 Not shown; usually referred to as a "hard drive"). Although Figure 4 As not shown, disk drives for reading and writing to removable non-volatile disks (e.g., "floppy disks") and optical disc drives for reading and writing to removable non-volatile optical discs (e.g., CD-ROMs, DVD-ROMs, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. System memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of the present invention.

[0108] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 42 typically perform the functions and / or methods described in the embodiments of the present invention.

[0109] Electronic device 4 can also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), and with one or more devices that enable a user to interact with electronic device 4, and / or with any device that enables electronic device 4 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed through input / output (I / O) interface 22. Furthermore, electronic device 4 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 20. Figure 4 As shown, network adapter 20 communicates with other modules of electronic device 4 via bus 18. It should be understood that, although... Figure 4 Not shown, it can be combined with electronic device 4 to use other hardware and / or software modules, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0110] Processing unit 16 executes various functional applications and page displays by running programs stored in system memory 28, such as implementing the recall job determination method provided in this embodiment of the invention, which includes: Based on the user's basic information, the user profile bucket corresponding to the user is determined from multiple pre-divided profile buckets, wherein the profile bucket is determined based on a combination of multiple target user features that have an impact on job category migration; Based on the user's initial job category and the user profile bucketing, the candidate job categories corresponding to the user and the migration probability of each candidate job category are determined by searching the job category migration probability table. The candidate job categories are filtered based on probability thresholds and the migration probability of each candidate job category to obtain migration job categories; The positions corresponding to the initial job category and the migration job category are determined as the recall positions corresponding to the user.

[0111] Of course, those skilled in the art will understand that the processor can also implement the technical solution of the recall job determination method provided in any embodiment of the present invention.

[0112] This invention provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the program implements, for example, the recall job determination method provided in this invention, which includes: Based on the user's basic information, the user profile bucket corresponding to the user is determined from multiple pre-divided profile buckets, wherein the profile bucket is determined based on a combination of multiple target user features that have an impact on job category migration; Based on the user's initial job category and the user profile bucketing, the candidate job categories corresponding to the user and the migration probability of each candidate job category are determined by searching the job category migration probability table. The candidate job categories are filtered based on probability thresholds and the migration probability of each candidate job category to obtain migration job categories; The positions corresponding to the initial job category and the migration job category are determined as the recall positions corresponding to the user.

[0113] The computer storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0114] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0115] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0116] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0117] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computing device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0118] Furthermore, the acquisition, storage, use, and processing of data in the technical solution of this invention all comply with relevant laws and regulations.

[0119] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A method for determining recall positions, characterized in that, include: Based on the user's basic information, the user profile bucket corresponding to the user is determined from multiple pre-divided profile buckets, wherein the profile bucket is determined based on a combination of multiple target user features that have an impact on job category migration; Based on the user's initial job category and the user profile bucketing, the candidate job categories corresponding to the user and the migration probability of each candidate job category are determined by searching the job category migration probability table. The candidate job categories are filtered based on probability thresholds and the migration probability of each candidate job category to obtain migration job categories; The positions corresponding to the initial job category and the migration job category are determined as the recall positions corresponding to the user.

2. The method for determining recall positions according to claim 1, characterized in that, Before determining the user profile bucket corresponding to the user among multiple pre-divided profile buckets based on the user's basic information, the process also includes: Filter target user characteristics that have an impact on job category migration from user characteristics; The feature dimensions of each target user feature are determined, and the feature dimensions of each target user feature are combined to obtain multiple profile buckets.

3. The method for determining recall positions according to claim 2, characterized in that, Based on the user's basic information, the user profile bucket corresponding to the user is determined from multiple pre-divided profile buckets, including: Determine user characteristics based on the user's basic information; The user profile bucket corresponding to the user is determined by comparing the feature dimensions of the target user features in the user features with the feature dimensions of the target user features that make up each profile bucket.

4. The method for determining recall positions according to claim 1, characterized in that, Also includes: For each of the aforementioned profile buckets, determine the migration probability of users in each job category under that profile bucket; Based on each of the aforementioned profile buckets and the migration probability of users within each of the aforementioned profile buckets across different job categories, a job category migration probability table is constructed.

5. The method for determining recall positions according to claim 4, characterized in that, Determining the migration probability of users across job categories under the aforementioned user profile bucketing includes: Each job category in the job category set is combined with other job categories to obtain multiple sets of starting job categories and target job categories; For each initial job category and target job category, the feature dimensions of the target user features corresponding to the profile bucket, as well as the initial job category and target job category, are input into a pre-trained migration probability model so that the migration probability model can determine the migration probability of a user in the profile bucket from the initial job category to the target job category.

6. The method for determining recall positions according to claim 5, characterized in that, When the target job category is the job category corresponding to a less popular job, the migration probability of the user between each job category under the profile bucket is determined, including: The migration probability determined by the migration probability model is amplified by an amplification factor to obtain the target migration probability.

7. The method for determining recall positions according to claim 1, characterized in that, When the candidate job category corresponds to a less popular job, the candidate job categories are filtered based on a probability threshold and the migration probability of each candidate job category to obtain migration job categories, including: The probability threshold is reduced by a reduction factor to obtain the target probability threshold; If the migration probability of the candidate job category is greater than the target probability threshold, then the candidate job category is determined as the migration job category.

8. A device for determining the recall position, characterized in that, include: The determination module is used to determine the user profile bucket corresponding to the user from multiple pre-divided profile buckets based on the user's basic information, wherein the profile bucket is determined based on a combination of multiple target user features that have an impact on job category migration; The search module is used to search in the job category migration probability table according to the user's initial job category and the user profile binning, to determine the candidate job categories corresponding to the user and the migration probability of each candidate job category; A filtering module is used to filter the candidate job categories based on a probability threshold and the migration probability of each candidate job category to obtain migration job categories; The execution module is used to determine the positions corresponding to the starting job category and the migration job category as the recall positions corresponding to the user.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to said at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the recall job determination method as described in any one of claims 1-7.

10. A storage medium containing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the recall job determination method as described in any one of claims 1-7.

Images