Precise job recommendation system based on big data dimension

Abstract

The present invention discloses a multi-dimensional job precision recommendation system based on big data, and relates to the technical field of job recommendation. The activity analysis module analyzes the activity of users to determine whether the users need to be matched. If it is determined that matching is necessary, the matching module first matches the users with enterprises and generates matching scores for the users. The threshold adjustment module dynamically adjusts the matching score threshold in combination with the daily active information of the users. The recommendation judgment module compares the matching score of the users with the dynamically adjusted matching score threshold. If the matching score is greater than or equal to the matching score threshold, the users are recommended to the enterprises. If the matching score is less than the matching score threshold, no recommendation is made. The recommendation system analyzes the user activity, dynamically adjusts the matching score threshold, effectively improves the matching efficiency and enhances the user participation, generates user portrait information, and displays it to the enterprises, so as to help the enterprises better understand the user's activity and matching degree.

Description

Technical Field

[0001] The present invention relates to the technical field of job recommendation, and in particular to a multi-dimensional accurate job recommendation system based on big data. Background Art

[0002] The job recommendation system is a technology based on data analysis and algorithms, which aims to help job seekers and employers match suitable jobs and talents more efficiently. With the development of the Internet and the increase in job demand, the traditional human resource matching method can no longer cope with the huge recruitment market, so the intelligent job recommendation system came into being; The prior art has the following defects: 1. The existing recommendation system adopts a fixed threshold strategy. Regardless of the user's activity level, it will recommend positions or companies that meet a certain matching score. This will lead to a waste of resources, especially for users with low activity levels, who may not be interested in the recommendation, resulting in invalid recommendations in the system. 2. The recommendation system only makes recommendations based on static user characteristics or behavioral data, ignoring the user's real-time status changes and behavioral trends, resulting in low recommendation accuracy and low user acceptance of recommendation results. In addition, the recommendation system often focuses on the matching of users and positions, but ignores the deep integration of user portraits and corporate needs. User portraits may not be fully displayed to companies, resulting in companies being unable to obtain sufficient information in recruitment decisions.

[0003] Based on this, the present invention proposes a multi-dimensional accurate job recommendation system based on big data. By analyzing user activity and dynamically adjusting the matching score threshold, the matching efficiency is effectively improved and user participation is enhanced, user portrait information is generated, and displayed to enterprises, helping enterprises to better understand user activity and matching. Summary of the invention

[0004] The purpose of the present invention is to provide a multi-dimensional accurate job recommendation system based on big data to address the deficiencies in the background technology.

[0005] In order to achieve the above-mentioned purpose, the present invention provides the following technical solutions: a multi-dimensional accurate job recommendation system based on big data, including an activity analysis module, a matching module, a threshold adjustment module, a recommendation judgment module, and a display module; Activity analysis module: analyzes user activity and determines whether users need to be matched; Matching module: If it is determined that a match is necessary, the user is matched with the enterprise first, and a matching score is generated for the user; Threshold adjustment module: dynamically adjusts the matching score threshold based on the user's daily activity information; Recommendation judgment module: compares the user's matching score with the dynamically adjusted matching score threshold. If the matching score is greater than or equal to the matching score threshold, the user is recommended to the enterprise. If the matching score is less than the matching score threshold, no recommendation is made. Display module: After integrating the user's activity information and maximum matching score into a data set, user portrait information is generated based on the data set and displayed to the enterprise.

[0006] The activity index is obtained by comprehensively analyzing the online time, operation fluctuation assignment, and unresponsive frequency. The expression is: Where, At level is the activity index, OL durate is the online time, OA agm Assign a value to the operating fluctuation, rpd not is the unresponsive frequency, σ1, σ2, and σ3 are the adjustment coefficients of online time, operation fluctuation value, and unresponsive frequency, respectively, and all are greater than 0.

[0007] In a preferred embodiment, the activity analysis module compares the acquired activity index with a preset first index threshold. The first index threshold is used to determine whether it is necessary to match the user. If the activity index is greater than or equal to the first index threshold, it is determined that the user needs to be matched. If the activity index is less than the first index threshold, it is determined that the user does not need to be matched.

[0008] In a preferred embodiment, the matching module obtains the user's skill-position matching index and salary expectation matching index, normalizes the skill-position matching index and the salary expectation matching index, maps the value range of the skill-position matching index and the salary expectation matching index to between [0,1], obtains the normalized value of the skill-position matching index and the normalized value of the salary expectation matching index, and sums the normalized value of the skill-position matching index and the normalized value of the salary expectation matching index to obtain a matching score.

[0009] In a preferred embodiment, after the threshold adjustment module obtains the activity index of the user on the previous day, since the activity index is greater than or equal to the first index threshold when the user is matched, the activity index is compared with the second index threshold, the second index threshold is greater than the first index threshold, and the second index threshold is used to analyze whether the user activity is high or average; If the activity index is less than the second index threshold, the user activity is average and the matching score threshold needs to be increased. If the activity index is greater than or equal to the second index threshold, the user activity is high and the matching score threshold needs to be reduced. The adjustment algorithm is: In the formula, Rt new is the adjusted matching score threshold, Rt oldis the matching score threshold before adjustment, At level is the activity index, and A2 is the second index threshold.

[0010] In a preferred embodiment, the display module obtains the user's activity index and maximum matching score, and weights the activity index and the maximum matching score to obtain the user's portrait coefficient, which is expressed as: portrait = ω1·At level +ω2·max(mat i ), where is the image coefficient, At level is the activity index, mat i represents the matching score of the user for the i-th enterprise, which means selecting the maximum value, ω1 and ω2 are weights, and ω1+ω2=1; Sort all users on the recruitment platform according to the portrait coefficient from large to small, generate a user list, and display the user list in a table.

[0011] In a preferred embodiment, the acquisition logic of the operation fluctuation assignment is: obtaining the number of times the user used the recruitment platform the previous day, and obtaining the number of operations each time the recruitment platform is used, wherein the number of operations is obtained by summing the number of page views, the number of job views, the number of job searches, the number of resume updates, and the number of interactions; Calculate the average number of operations and the standard deviation of the number of operations. The expression is: Where, OPR avg is the average number of operations, OPR q is the standard deviation of the number of operations, n represents the number of times the user used the recruitment platform the day before, and OPR i represents the number of operations when the recruitment platform is used for the i-th time, and generates an operation fluctuation value based on the average number of operations and the standard deviation of the number of operations; If the average number of operations is greater than or equal to the number threshold, and the standard deviation of the number of operations is less than or equal to the standard deviation threshold, the operation fluctuation is assigned OA agm =2.0; if the average number of operations is greater than or equal to the number threshold, and the standard deviation of the number of operations is greater than the standard deviation threshold, the operation fluctuation is assigned the value OA agm =1.8; if the average number of operations is less than the number threshold, and the standard deviation of the number of operations is greater than the standard deviation threshold, the operation fluctuation is assigned OA agm =1.4; if the average number of operations is less than the number threshold, and the standard deviation of the number of operations is less than or equal to the standard deviation threshold, the operation fluctuation is assigned OA agm =1.0.

[0012] In a preferred embodiment, the logic for obtaining the online duration is as follows: obtaining the number of times the user used the recruitment platform the previous day, wherein the time from when the user opens the recruitment platform to when the user closes the recruitment platform is considered as one use, recording the duration of each use of the recruitment platform, and summing up the duration of each use of the recruitment platform to obtain the online duration; The calculation logic of the non-response frequency is: obtain the number of times the user did not respond to the interview invitation, the number of times the user did not respond to the conversation, and the number of times the user did not respond to the notification on the previous day, sum the number of times the user did not respond to the interview invitation, the number of times the user did not respond to the conversation, and the number of times the user did not respond to the notification to obtain the total number of times the user did not respond, and divide the total number of times the user did not respond by 24 to obtain the non-response frequency.

[0013] In the above technical solution, the technical effects and advantages provided by the present invention are: The present invention analyzes the activity of users through an activity analysis module to determine whether the user needs to be matched. If it is determined that matching is necessary, the matching module first matches the user with the enterprise and generates a matching score for the user. The threshold adjustment module dynamically adjusts the matching score threshold in combination with the user's daily activity information. The recommendation judgment module compares the user's matching score with the dynamically adjusted matching score threshold. If the matching score is greater than or equal to the matching score threshold, the user is recommended to the enterprise. If the matching score is less than the matching score threshold, no recommendation is made. The recommendation system analyzes the user's activity, dynamically adjusts the matching score threshold, effectively improves matching efficiency and enhances user participation, generates user portrait information, and displays it to the enterprise to help the enterprise better understand the user's activity and matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0015] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0017] Example 1: Please refer to Figure 1As shown, the multi-dimensional job precision recommendation system based on big data described in this embodiment includes an activity analysis module, a matching module, a threshold adjustment module, a recommendation judgment module, and a display module; Activity analysis module: analyzes the activity of users and determines whether users need to be matched. The judgment result is sent to the matching module, and the activity information is sent to the threshold adjustment module and the display module. Matching module: If it is determined that matching is necessary, the user is first matched with the enterprise, and a matching score is generated for the user. The larger the matching score, the more likely it is that the user needs to be recommended to the enterprise. The matching score is sent to the threshold adjustment module, the recommendation judgment module, and the display module; Threshold adjustment module: dynamically adjusts the matching score threshold based on the user's daily activity information (usually, the higher the activity, the lower the matching score threshold, thereby relaxing the matching authority of the user; the lower the activity, the higher the matching score threshold, thereby avoiding waste of matching resources). The dynamically adjusted matching score threshold is sent to the recommendation judgment module; Recommendation judgment module: compares the user's matching score with the dynamically adjusted matching score threshold. If the matching score is greater than or equal to the matching score threshold, the user is recommended to the enterprise. If the matching score is less than the matching score threshold, no recommendation is made. Display module: After integrating the user's activity information and maximum matching score into a data set, user portrait information is generated based on the data set and displayed to the enterprise.

[0018] The specific workflow of the recommendation system is as follows: The recommendation system analyzes the user's activity to determine whether the user needs to be matched. If it is determined that matching is necessary, the user is first matched with the enterprise to generate a matching score for the user. The larger the matching score, the more necessary it is to recommend the user to the enterprise. Then, the matching score threshold is dynamically adjusted based on the user's daily activity information (usually, the higher the activity, the lower the matching score threshold, thereby relaxing the matching authority for the user; the lower the activity, the higher the matching score threshold, to avoid wasting matching resources). The user's matching score is compared with the dynamically adjusted matching score threshold. If the matching score is greater than or equal to the matching score threshold, the user is recommended to the enterprise. If the matching score is less than the matching score threshold, no recommendation is made. After the user's activity information and the maximum matching score are integrated into a data set, user portrait information is generated based on the data set and displayed to the enterprise.

[0019] This application uses the activity analysis module to analyze the user's activity and determine whether the user needs to be matched. If it is determined that matching is necessary, the matching module first matches the user with the enterprise and generates a matching score for the user. The threshold adjustment module dynamically adjusts the matching score threshold based on the user's daily activity information. The recommendation judgment module compares the user's matching score with the dynamically adjusted matching score threshold. If the matching score is greater than or equal to the matching score threshold, the user is recommended to the enterprise. If the matching score is less than the matching score threshold, no recommendation is made. The recommendation system analyzes the user's activity and dynamically adjusts the matching score threshold, effectively improving matching efficiency and enhancing user participation, generating user portrait information, and displaying it to the enterprise to help the enterprise better understand the user's activity and matching.

[0020] Embodiment 2: The activity analysis module performs activity analysis on the user to determine whether the user needs to be matched; The activity analysis module obtains the user's participation data on the recruitment platform the previous day. The participation data includes online time, operation fluctuation value, and non-response frequency; The activity analysis module comprehensively analyzes the online time, operation fluctuation value, and unresponsive frequency to obtain the activity index, which is expressed as: Where, At level is the activity index, OL durate is the online time, OA agm Assign a value to the operating fluctuation, rpd not is the unresponsive frequency, σ1, σ2, and σ3 are the adjustment coefficients of online duration, operation fluctuation assignment, and unresponsive frequency, respectively, and all are greater than 0; The larger the activity index, the more active the user is. The obtained activity index is compared with the preset first index threshold. The first index threshold is used to determine whether the user needs to be matched. If the activity index is greater than or equal to the first index threshold, it is determined that the user needs to be matched. If the activity index is less than the first index threshold, it is determined that the user does not need to be matched.

[0021] The logic for obtaining the online time is as follows: obtain the number of times the user used the recruitment platform the day before. The time from when the user opened the recruitment platform to when the user closed the recruitment platform is considered as one use. The duration of each use of the recruitment platform is recorded, and the online time is obtained by summing up the duration of each use of the recruitment platform. The longer the online time, the more active the user was on the recruitment platform the day before. Specifically: The longer the time, the more active it is: Online time reflects the time users spend on the platform. The longer the time, the more time users spend on the platform, and the more operations or interactions they may have performed. This usually indicates that users are more active, perhaps because they are actively searching for jobs, updating their resumes, or interacting with recruiters.

[0022] High online time means more frequent operations: When users spend a lot of time on the platform, it usually indicates that they have performed a series of job-seeking related operations, such as viewing positions, searching for new positions, applying for positions, updating resumes, etc. A large number of operations and operation duration also reflect the user's dependence on and participation in the platform.

[0023] The logic for obtaining the operation fluctuation value is as follows: obtain the number of times the user used the recruitment platform the previous day, and obtain the number of operations each time the recruitment platform is used, where the number of operations is obtained by summing the number of page views, job views, job searches, resume updates, and interactions; Calculate the average number of operations and the standard deviation of the number of operations. The expression is: Where, OPR avg is the average number of operations, OPR q is the standard deviation of the number of operations, n represents the number of times the user used the recruitment platform the day before, and OPR i represents the number of operations when using the recruitment platform for the i-th time; Generate an operation fluctuation value based on the average number of operations and the standard deviation of the number of operations; If the average number of operations is greater than or equal to the number threshold, and the standard deviation of the number of operations is less than or equal to the standard deviation threshold, the user's overall number of operations on the recruitment platform the day before was high, and the operation fluctuation is assigned OA agm =2.0; If the average number of operations is greater than or equal to the number threshold, and the standard deviation of the number of operations is greater than the standard deviation threshold, the user's number of operations on the recruitment platform the day before was relatively large (but there were fewer operations when the recruitment platform was used less frequently), and the operation fluctuation is assigned an OA value. agm =1.8; If the average number of operations is less than the number threshold, and the standard deviation of the number of operations is greater than the standard deviation threshold, the user's number of operations on the recruitment platform the day before was small (but there were more operations when the recruitment platform was used less frequently), and the operation fluctuation is assigned OA agm =1.4; If the average number of operations is less than the number threshold, and the standard deviation of the number of operations is less than or equal to the standard deviation threshold, the user's overall number of operations on the recruitment platform the day before was small, and the operation fluctuation is assigned OA agm =1.0; In summary, the larger the value of the operation fluctuation assignment is, the more active the user is.

[0024] The calculation logic of the non-response frequency is as follows: obtain the number of times the user did not respond to interview invitations, the number of times they did not respond to sessions, and the number of times they did not respond to notifications the day before, sum up the number of times they did not respond to interview invitations, the number of times they did not respond to sessions, and the number of times they did not respond to notifications to get the total number of times they did not respond, and divide the total number of times they did not respond by 24 to get the non-response frequency, where each day is calculated as 24 hours, so the monitoring time is 24 hours.

[0025] The smaller the non-response frequency, the more active the user is; The user's response to the system (such as clicking to view positions, updating resumes, applying for positions, etc.) is an important indicator of user activity. When a user frequently responds to the system's recommendations and prompts, it indicates that he is interested in the platform's content or services and is willing to participate in the interaction. If a user does not respond to the system's recommendations or operation prompts, it indicates that he has a low level of attention or participation in the platform, or does not currently need these services; When the frequency of non-response is low, it means that the user is responding to the system's recommendations or functions most of the time. This shows that the user is very interested in the activity content on the platform and participates in it frequently. For example, the user may check new job recommendations in a timely manner, or actively respond to recruitment information, message notifications, etc. Such behavior is a characteristic of highly active users.

[0026] Reflection of high engagement: Low frequency of non-response indicates high user engagement with the platform. This not only means that users spend more time and energy on the platform, but also indicates that users have a high demand for the services provided by the platform and are willing to use the platform to find suitable job opportunities or recruitment information.

[0027] If it is determined that a match is needed, the matching module will first match the user with the enterprise and generate a matching score for the user. The higher the matching score, the more likely it is that the user needs to be recommended to the enterprise. The matching module obtains the user's skill-position matching index and salary expectation matching index, normalizes the skill-position matching index and the salary expectation matching index, maps the value range of the skill-position matching index and the salary expectation matching index to [0,1], and obtains the normalized value of the skill-position matching index and the normalized value of the salary expectation matching index; The matching score is obtained by summing the normalized value of the skill-position matching index and the normalized value of the salary expectation matching index. The larger the matching score, the more matched the user is with the enterprise, and the more necessary it is to recommend the user to the enterprise.

[0028] The calculation expression of skill-job matching index is: skills =Sim(U skills ,J skills ), where f skills is the skill-job matching index, U skills is the user skill set, Jskills is the enterprise position skill set, Sim(*) is the skill similarity function, In the formula, ∩ represents the intersection of two sets (skills possessed by both users and positions), and ∪ represents the union of two sets (skills possessed by either users or positions). The larger the skill-position matching index, the more matching the enterprise position is with the user. Specifically: High skill match: When the skill-job match index is high, it means that the user's skills are very close to the requirements of the company's job. The user's skills can well meet the requirements of the job, so he is more suitable for the job.

[0029] Skill fit: The more the user's skills overlap with the core skills required for the position, or the higher the match between the key skills, the greater the match index will be, indicating that the user has the ability to perform the position.

[0030] Wider skill coverage: The larger the matching index, the more likely it is that the user's skills can cover the core requirements of the position, and even exceed the basic skill requirements of the position in some aspects. This means that the user is not only competent for the basic work of the position, but is also likely to perform well in higher-level tasks.

[0031] Accurately meet job requirements: When companies post jobs, they usually clearly require certain specific skills (such as programming languages, tool usage, industry experience, etc.). If the user's skills are highly consistent with these requirements, the matching index will be high, indicating that the user can get started on the job faster and meet the company's expectations.

[0032] Highly matched key skills: There may be some particularly important key skills for a position. If the user has a high match in these key skills, the match index will increase accordingly. This usually means that the user not only meets the basic requirements of the position, but also has the core competitiveness required for the position.

[0033] Enterprise perspective: reducing training and adaptation costs Reduce training costs: Companies are more inclined to choose candidates with high skill matching because these candidates can quickly adapt to job requirements, reducing the company's investment in skill training and job adaptation.

[0034] Improve work efficiency: When the skill-job matching index is high, users can directly apply their existing skills, quickly get started and improve work efficiency, which is an ideal situation for enterprises.

[0035] Reduce recruitment risks: Highly matched candidates are not only suitable for the position, but also reduce the risk of recruitment failure, ensuring that new employees can smoothly integrate into the team and achieve work goals.

[0036] 6. User perspective: better career development opportunities Higher job fit: Positions with a high matching index usually mean that users can give full play to their skills and experience, not only being competent in the position but also being able to develop.

[0037] Potential for career development: When a user's skills are highly matched to the position, it means that they are not only competent for the current job, but also likely to have more opportunities for future career development. For users, such highly matched positions are more in line with their long-term career plans.

[0038] The calculation expression of salary expectation matching index is: In the formula, f salary is the salary expectation matching index, U saiary is the user's expected salary, J salary Provide salary for corporate positions. The larger the expected salary matching index is, the more the corporate position matches the user. Specifically: Salary is a key factor for job seekers: Salary is one of the most important factors for users to choose a job. If the salary offered by the company is highly consistent with the user's expected salary, the user will think that the position is in line with his or her career goals and living standards, and is more likely to be willing to accept the position.

[0039] High matching degree reduces the salary negotiation process: When salary expectations and corporate salaries are highly matched, the time and energy spent by both parties on salary negotiations will be greatly reduced, increasing the probability of successful job matching. This means that both parties have reached a consensus on this important condition.

[0040] Enhance job-seeking attractiveness: The higher the salary matching index, the more attractive the position is to users. Users will not only pay more attention to other details of the position (such as job content, company culture, etc.), but will also be more inclined to further contact the recruiting company.

[0041] The significance of salary expectation matching to enterprises Improve recruitment success rate: When the salary offered by the company is highly matched with user expectations, the recruitment success rate will be greatly improved. Candidates are more likely to accept the offer because salary is a key decision factor.

[0042] Reduce the risk of employee turnover: Salaries that match user expectations mean that users are more satisfied with their salary and benefits, which will reduce employee dissatisfaction after joining the company and reduce the risk of leaving the company due to salary issues.

[0043] Recruitment cost optimization: By improving the salary matching index, companies can more effectively attract suitable candidates, reduce unnecessary salary negotiation processes, shorten the recruitment cycle, and optimize the recruitment process.

[0044] User perspective: Career stability brought by salary matching Improved salary satisfaction: For users, the higher the salary expectation matching index, the more they do not need to compromise on salary and can obtain income that meets their expectations, which makes them more satisfied with their positions and companies.

[0045] Potential for long-term development: When users’ salary expectations are met, they will be more inclined to develop in the company for a long time and reduce the possibility of job-hopping due to salary issues. Therefore, a highly matched salary index not only affects users’ willingness to join the company, but also has a positive impact on their subsequent career development.

[0046] The threshold adjustment module dynamically adjusts the matching score threshold based on the user's daily activity information (usually, the higher the activity, the lower the matching score threshold, thereby relaxing the matching authority of the user, and the lower the activity, the higher the matching score threshold, to avoid wasting matching resources); After the threshold adjustment module obtains the user's activity index of the previous day, since the activity index is greater than or equal to the first index threshold when the user is matched, we need to compare the activity index with the second index threshold. The second index threshold is greater than the first index threshold. The second index threshold is used to analyze whether the user's activity is high or average; If the activity index is less than the second index threshold, the user activity is average and the matching score threshold needs to be increased. If the activity index is greater than or equal to the second index threshold, the user activity is high and the matching score threshold needs to be reduced. The adjustment algorithm is: In the formula, Rt new is the adjusted matching score threshold, Rt old is the matching score threshold before adjustment, At level is the activity index, and A2 is the second index threshold.

[0047] The recommendation judgment module compares the user's matching score with the dynamically adjusted matching score threshold. If the matching score is greater than or equal to the matching score threshold, the user is recommended to the enterprise. If the matching score is less than the matching score threshold, no recommendation is made.

[0048] The display module integrates the user's activity information and the maximum matching score into a data set, generates user portrait information based on the data set and displays it to the enterprise; The display module obtains the user's activity index and maximum matching score (since the user will generate several matching scores for multiple companies, the maximum matching score is selected as the maximum matching score), and calculates the user's portrait coefficient by weighted calculation of the activity index and the maximum matching score. The expression is: portrait = ω1·At level +ω2·max(mat i ), where is the image coefficient, At levelis the activity index, mat i represents the matching score of the user for the i-th enterprise, which means selecting the maximum value, ω1 and ω2 are weights, and ω1+ω2=1; All users in the recruitment platform are sorted from large to small according to the portrait coefficient to generate a user list. Each item in the user list contains the user's basic information, such as name, skills, work experience, recommendation priority, etc. The user list is displayed in a table, which contains various key information of the user (such as portrait coefficient, skills, experience, etc.).

[0049] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0050] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0051] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A multi-dimensional job recommendation system based on big data, characterized by: It includes activity analysis module, matching module, threshold adjustment module, recommendation judgment module and display module; Activity analysis module: analyzes user activity and determines whether users need to be matched; Matching module: If it is determined that a match is necessary, the user is matched with the enterprise first, and a matching score is generated for the user; Threshold adjustment module: dynamically adjusts the matching score threshold based on the user's daily activity information; Recommendation judgment module: compares the user's matching score with the dynamically adjusted matching score threshold. If the matching score is greater than or equal to the matching score threshold, the user is recommended to the enterprise. If the matching score is less than the matching score threshold, no recommendation is made. Display module: After integrating the user's activity information and maximum matching score into a data set, user portrait information is generated based on the data set and displayed to the enterprise.

2. The multi-dimensional job recommendation system based on big data according to claim 1 is characterized by: The activity analysis module obtains the user's participation data on the recruitment platform the previous day, the participation data including online time, operation fluctuation value and non-response frequency; The activity index is obtained by comprehensively analyzing the online time, operation fluctuation assignment, and unresponsive frequency. The expression is: , where is the activity index, is the online time, Assign a value to the operating fluctuation, is the unresponse frequency, , , They are the adjustment coefficients of online time, operation fluctuation value and unresponsive frequency, and all are greater than 0.

3. The multi-dimensional job recommendation system based on big data according to claim 2 is characterized by: The activity analysis module compares the acquired activity index with a preset first index threshold. The first index threshold is used to determine whether it is necessary to match the user. If the activity index is greater than or equal to the first index threshold, it is determined that the user needs to be matched. If the activity index is less than the first index threshold, it is determined that it is not necessary to match the user.

4. The multi-dimensional job recommendation system based on big data according to claim 3 is characterized by: The matching module obtains the user's skill-position matching index and salary expectation matching index, normalizes the skill-position matching index and the salary expectation matching index, maps the value range of the skill-position matching index and the salary expectation matching index to between [0,1], obtains the normalized value of the skill-position matching index and the normalized value of the salary expectation matching index, and sums the normalized value of the skill-position matching index and the normalized value of the salary expectation matching index to obtain a matching score.

5. The multi-dimensional job recommendation system based on big data according to claim 4 is characterized by: After the threshold adjustment module obtains the activity index of the user on the previous day, since the activity index of the user is greater than or equal to the first index threshold when the user is matched, the activity index is compared with the second index threshold, and the second index threshold is greater than the first index threshold. The second index threshold is used to analyze whether the user's activity is high or average; If the activity index is less than the second index threshold, the user activity is average and the matching score threshold needs to be increased. If the activity index is greater than or equal to the second index threshold, the user activity is high and the matching score threshold needs to be reduced. The adjustment algorithm is: , where is the adjusted matching score threshold, is the matching score threshold before adjustment, is the activity index, is the second index threshold.

6. The multi-dimensional job recommendation system based on big data according to claim 5 is characterized by: The display module obtains the user's activity index and maximum matching score, and performs weighted calculation on the activity index and maximum matching score to obtain the user's portrait coefficient, which is expressed as: , where is the image coefficient, is the activity index, Indicates that the user is facing The matching scores of enterprises are selected, indicating that the maximum value is selected. , is the weight, and ; Sort all users on the recruitment platform according to the portrait coefficient from large to small, generate a user list, and display the user list in a table.

7. The multi-dimensional job recommendation system based on big data according to claim 2 is characterized by: The logic for obtaining the operation fluctuation value is as follows: obtaining the number of times the user used the recruitment platform the previous day, and obtaining the number of operations each time the recruitment platform is used, wherein the number of operations is obtained by summing the number of page views, the number of job views, the number of job searches, the number of resume updates, and the number of interactions; Calculate the average number of operations and the standard deviation of the number of operations. The expression is: , where is the average number of operations, is the standard deviation of the number of operations, Indicates the number of times the user used the recruitment platform the day before. Indicates The number of operations when using the recruitment platform each time, and the operation fluctuation value is generated based on the average number of operations and the standard deviation of the number of operations; If the average number of operations is greater than or equal to the number threshold, and the standard deviation of the number of operations is less than or equal to the standard deviation threshold, the operation fluctuation is assigned ; If the average number of operations is greater than or equal to the number threshold, and the standard deviation of the number of operations is greater than the standard deviation threshold, the operation fluctuation is assigned ; If the average number of operations is less than the number threshold, and the standard deviation of the number of operations is greater than the standard deviation threshold, the operation fluctuation is assigned ; If the average number of operations is less than the number threshold, and the standard deviation of the number of operations is less than or equal to the standard deviation threshold, the operation fluctuation is assigned .

8. The multi-dimensional job recommendation system based on big data according to claim 2 is characterized by: The logic for obtaining the online duration is as follows: obtaining the number of times the user used the recruitment platform the previous day, wherein the time from when the user opens the recruitment platform to when the user closes the recruitment platform is considered as one use, recording the duration of each use of the recruitment platform, and summing up the duration of each use of the recruitment platform to obtain the online duration; The calculation logic of the non-response frequency is: obtain the number of times the user did not respond to the interview invitation, the number of times the user did not respond to the conversation, and the number of times the user did not respond to the notification on the previous day, sum the number of times the user did not respond to the interview invitation, the number of times the user did not respond to the conversation, and the number of times the user did not respond to the notification to obtain the total number of times the user did not respond, and divide the total number of times the user did not respond by 24 to obtain the non-response frequency.

Images