Job application and recruitment recommendation system and method based on dynamic weight collaborative filtering

Abstract

The invention discloses a job application and recruitment recommendation system and method based on dynamic weight collaborative filtering, belongs to the technical field of intelligent services, and aims to solve the technical problem of how to realize accurate pushing of information in an employment service scene. Comprising a job hunting user database used for storing job hunting information of each job hunting user and a job hunting person portrait constructed based on the job hunting information; the enterprise user database is used for storing post information of each enterprise user and an enterprise portrait constructed based on the post information; the government affair information pushing module is used for pushing related government affair information to the job hunting user and providing personalized planning consultation service; the position information pushing module is used for pushing a job-hunting and recruitment short message to the job-hunting user who performs unemployment registration based on the position information of the related enterprise user; the job hunting matching module retrieves matched enterprise users for the specific job hunting users; and the vacancy quick matching module is used for recommending the job application information of the related job application users to the enterprise users.

Description

Technical Field

[0001] The present invention relates to the field of intelligent service technology, and in particular to a job search and recruitment recommendation system and method based on dynamic weighted collaborative filtering. Background Art

[0002] In order to improve the public employment service system, promote the overall improvement of service capabilities, and achieve high-quality full employment, it is necessary to focus on building a "public employment service ecosystem" that covers the entire population, runs through the entire process, radiates the entire region, and is convenient and efficient, so as to promote the improvement of the quality and efficiency of public employment services in a wider range, more levels, and deeper degree, and promote high-quality economic and social development.

[0003] How to achieve accurate information push in employment service scenarios is a technical problem that needs to be solved. Summary of the Invention

[0004] The technical task of the present invention is to address the above shortcomings and provide a job search and recruitment recommendation system and method based on dynamic weighted collaborative filtering to solve the technical problem of how to achieve accurate push of information in employment service scenarios.

[0005] In a first aspect, the present invention provides a job search and recruitment recommendation system based on dynamic weighted collaborative filtering, comprising a government information push module, a job information push module, a job search matching module, a vacancy matching module, a service applet, a job search user database, and an enterprise user database; The service mini-program is designed for corporate users and job seekers, and is used to support job seekers in registering and logging in, posting job information, and performing job search operations. It is also used to provide job seekers with employment guidance, and to support corporate users in registering and logging in, posting job information, and performing recruitment operations. The job seeker database is used to store each job seeker's job search information and a job seeker profile constructed based on the job search information. The job seeker profile includes the job seeker's features and feature vectors in various dimensions. The job seeker's features are obtained by filtering the job search information. The enterprise user database is used to store the job information of each enterprise user and the enterprise profile built based on the job information. The enterprise profile includes the characteristics and feature vectors of the enterprise user in various dimensions. The characteristics of the enterprise user are obtained by filtering the job information. The government information push module is used to push relevant government information to job seekers based on their job search information, and to provide personalized planning consulting services based on their permissions and permission validity period. The job information push module is used to find job seekers who have registered for unemployment on the same day. It searches and matches based on corporate and job seeker profiles and dynamic weighted collaborative filtering. It then pushes job recruitment text messages to job seekers who have registered for unemployment based on the job information of related corporate users. The job search matching module is used to search and match based on corporate and job seeker profiles through dynamic weighted collaborative filtering, retrieve matching corporate users for specific job seekers, and send recommended information to specific job seekers based on the job information of related corporate users. Specific job seekers include job seekers whose job information has not been paid attention to by companies within a predetermined period of time and job seekers who have difficulty finding employment; The vacancy matching module is used to search and match based on corporate profiles and job seeker profiles through dynamic weighted collaborative filtering. It retrieves matching job seekers for corporate users who have not recruited the required personnel within the set deadline and recommends the job search information of relevant job seekers to corporate users.

[0006] Preferably, for job seekers who are applying for unemployment registration, the job information push module is configured to perform the following operations: Calculate the similarity between job search information and job position information based on dynamic weighted collaborative filtering, match relevant corporate users to job seekers based on the similarity, and send job search and recruitment information to job seekers; Determine the job seeker's registration status in the service mini-program. If the job seeker has not registered in the service mini-program, guide the job seeker to register and log in to the service mini-program and perform related operations. If the job seeker has registered in the service mini-program, match the job seeker with appropriate services based on the job search information posted by the job seeker in the service mini-program and the job search operations performed. Appropriate services include pushing employment guidance, government information, and recruitment information from related corporate users.

[0007] Preferably, the job matching module is configured to perform the following operations: Regularly scan the job-seeking user database, identify job-seeking users whose job information has not been followed by companies within a predetermined period of time and job-seeking users who have difficulty finding employment as specific job-seeking users, and set labels for these specific job-seeking users; For specific job seekers, search and matching are performed based on corporate profiles and job seeker profiles through dynamic weighted collaborative filtering to retrieve matching corporate users for the specific job seekers, and recommendation information is sent to the specific job seekers based on the job information of related corporate users.

[0008] As a preferred method, search and matching is performed based on the company profile and the job seeker profile through dynamic weighted collaborative filtering, including the following operations: Customize the weight of each job search information and position information; For enterprise users, set priority weights for each enterprise user based on the release time of job information; Vector similarity calculation is performed based on the feature vectors corresponding to the job application information and the feature vectors corresponding to the position information. When performing vector similarity calculation, a weighted sum is performed based on the weight of the job application information, the weight of the position information, and the priority weight of the enterprise user, and retrieval and matching is performed based on the final similarity calculation result.

[0009] In a second aspect, the present invention provides a job search and recruitment recommendation method based on dynamic weighted collaborative filtering, which implements intelligent job search and recruitment recommendations between enterprise users and job seekers based on a job search and recruitment recommendation system based on dynamic weighted collaborative filtering as described in any one of the first aspects, including the following steps: Build user profiles: Based on the job application information of job seekers, a job seeker profile is constructed. The job seeker profile includes the characteristics and feature vectors of the job seeker in various dimensions. Based on the job application information of enterprise users, a corporate profile is constructed. The corporate profile includes the characteristics and feature vectors of the enterprise users in various dimensions. The characteristics of job seekers are obtained by screening job application information, and the characteristics of enterprise users are obtained by screening job information. Government information push: Push relevant government information to job seekers based on their job search information, and provide personalized planning consulting services based on their permissions and permission validity periods; Job information push: Find job seekers who have registered for unemployment on the same day, perform search and matching based on company and job seeker profiles through dynamic weighted collaborative filtering, and push job recruitment SMS messages to job seekers who have registered for unemployment based on the job information of related company users; Job matching: Based on the company profile and job seeker profile, dynamic weighted collaborative filtering is used to search and match, retrieve matching company users for specific job seekers, and send recommended information to specific job seekers based on the job information of related company users. Specific job seekers include job seekers whose job information has not been paid attention to by companies within a predetermined period of time and job seekers who have difficulty finding employment; Vacancy Matching: Based on corporate and job seeker profiles, dynamic weighted collaborative filtering is used for search and matching. For corporate users who have not recruited the required personnel within the set period, matching job seekers are retrieved, and the job search information of relevant job seekers is recommended to corporate users.

[0010] As a preferred option, job seekers who are applying for unemployment registration should perform the following operations: Calculate the similarity between job search information and job position information based on dynamic weighted collaborative filtering, match relevant corporate users to job seekers based on the similarity, and send job search and recruitment information to job seekers; Determine the job seeker's registration status in the service mini-program. If the job seeker has not registered in the service mini-program, guide the job seeker to register and log in to the service mini-program and perform related operations. If the job seeker has registered in the service mini-program, match the job seeker with appropriate services based on the job search information posted by the job seeker in the service mini-program and the job search operations performed. Appropriate services include pushing employment guidance, government information, and recruitment information from related corporate users.

[0011] Preferably, job matching includes the following operations: Regularly scan the job-seeking user database, identify job-seeking users whose job information has not been followed by companies within a predetermined period of time and job-seeking users who have difficulty finding employment as specific job-seeking users, and set labels for these specific job-seeking users; For specific job seekers, search and matching are performed based on corporate profiles and job seeker profiles through dynamic weighted collaborative filtering to retrieve matching corporate users for the specific job seekers, and recommendation information is sent to the specific job seekers based on the job information of related corporate users.

[0012] As a preferred method, search and matching is performed based on the company profile and the job seeker profile through dynamic weighted collaborative filtering, including the following operations: Customize the weight of each job search information and position information; For enterprise users, set priority weights for each enterprise user based on the release time of job information; Vector similarity calculation is performed based on the feature vectors corresponding to the job application information and the feature vectors corresponding to the position information. When performing vector similarity calculation, a weighted sum is performed based on the weight of the job application information, the weight of the position information, and the priority weight of the enterprise user, and retrieval and matching is performed based on the final similarity calculation result.

[0013] The job search and recruitment recommendation system and method based on dynamic weighted collaborative filtering of the present invention has the following advantages: 1. Adopting an intelligent push algorithm with dynamic weighted collaborative filtering to simulate the manual policy push model, simplifying the workload of personnel and reducing the error rate of policy push; 2. It has greatly solved the problems of job seekers finding employment, obtaining employment consultation, and matching jobs. The employment matching is intelligent, electronic, and can be pushed quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0015] The present invention will be further described below with reference to the accompanying drawings.

[0016] Figure 1 This is a schematic diagram of the interface of a job search and recruitment recommendation system based on dynamic weighted collaborative filtering in Example 1. DETAILED DESCRIPTION

[0017] The present invention will be further described below with reference to the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it. However, the embodiments given are not intended to limit the present invention. Unless there is a conflict, the embodiments of the present invention and the technical features in the embodiments may be combined with each other.

[0018] The embodiments of the present invention provide a job search and recruitment recommendation system and method based on dynamic weighted collaborative filtering, which are used to solve the technical problem of how to achieve accurate push of information in employment service scenarios.

[0019] Embodiment 1: The present invention provides a job search and recruitment recommendation system based on dynamic weighted collaborative filtering, comprising a government information push module, a job information push module, a job search matching module, a vacancy matching module, a service applet, a job search user database, and an enterprise user database.

[0020] The service mini program is aimed at corporate users and job seekers. It is used to support job seekers in registering and logging in, posting job information, and performing job search operations. It is used to provide employment guidance to job seekers and to support corporate users in registering and logging in, posting job information, and performing recruitment operations.

[0021] The job seeker database is used to store the job search information of each job seeker and a job seeker portrait constructed based on the job search information. The job seeker portrait includes the job seeker's features and feature vectors in various dimensions, wherein the job seeker's features are obtained by screening the job search information.

[0022] The enterprise user database is used to store the job information of each enterprise user and the enterprise portrait built based on the job information. The enterprise portrait includes the characteristics and feature vectors of the enterprise user in various dimensions, among which the characteristics of the enterprise user are obtained by filtering the job information.

[0023] In this embodiment, the dimensions of the enterprise portrait include industry attributes, recruitment preferences, and historical behaviors. The features under industry attributes include manufacturing and IT services, and the encoding method is One-Hot encoding. The features under recruitment preferences include experience weight and educational requirements, and the encoding method is numerical standardization (0-1). The features under historical behaviors include average feedback time and interview conversion rate, and the encoding method is time series embedding.

[0024] The government information push module is used to push relevant government information to job seekers based on their job search information, and to provide personalized planning consulting services based on their permissions and permission time limit.

[0025] The government information push module in this embodiment intelligently pushes relevant policy information based on individual job search information, including employment policies, social security policies, and social security preferential policies (such as employment assistance policies for people aged 40-50, people with disabilities, and the unemployed). This helps job seekers fully understand the social security system and protect their rights and interests. At the same time, the system can provide personalized social security planning and consulting services based on individual social security payment status, achieving comprehensive coverage of social security services. The time-efficiency function formula is as follows: ; in, Indicates the number of days remaining in the validity period.

[0026] The policy vectorization algorithm is as follows: class PolicyEncoder:-_init__(self):def self.keywords ={"4858 personnel":8,"disabled people":1,"unemployment insurance":2} # Policy label dictionary def encode(self, policy_text): #Use BERT to extract semantic features vec_semantic = bert_model.encode(policy_text) # Structured feature encoding vec_structured = [1 if kw in policy_text else for kw inself.keywords] return np.concatenate([vec_semantic,vec_structured]). The job information push module is used to find job seekers who have registered for unemployment on the same day. It performs search and matching based on corporate portraits and job seeker portraits through dynamic weighted collaborative filtering, and pushes job recruitment text messages to job seekers who have registered for unemployment based on the job information of relevant corporate users.

[0027] For job seekers who are applying for unemployment registration, the job information push module of this embodiment is used to perform the following operations: (1) Calculate the similarity between job search information and job position information based on dynamic weighted collaborative filtering, match relevant enterprise users to job seekers based on the similarity, and send job search and recruitment information to job seekers; (2) Determine the registration status of the job seeker in the service mini-program. If the job seeker has not registered in the service mini-program, guide the job seeker to register and log in to the service mini-program and perform relevant operations. If the job seeker has registered in the service mini-program, match the job seeker with appropriate services based on the job search information posted by the job seeker in the service mini-program and the job search operations performed. Appropriate services include pushing employment guidance, government information, and recruitment information of relevant corporate users.

[0028] In this embodiment, the module regularly finds out the people who have registered for unemployment on the same day from the backflow data, and sends them job search and recruitment text messages (sends 2 to 3 intelligently recommended company information), and automatically determines whether this person has registered for the service applet by comparing personal information and registration status. If not, the relevant link address will be sent to guide them to seek employment, recruitment and training through the applet. If already registered, the user's usage behavior and published information will be intelligently analyzed, and accurate push will be made according to actual needs. For example, if the user has registered for training needs, training plans that match their training needs will be pushed first. In the applet, information such as intelligently matched companies and related training will be pushed, as well as unemployment insurance policies, unemployment benefit issuance policies, etc., reminders to handle social security policies during unemployment, and links to relevant addresses for social security applications; push personal training needs, push training plans to be carried out, etc. Accurate service push such as Figure 1 shown.

[0029] The job search matching module is used to search and match based on corporate portraits and job seeker portraits through dynamic weighted collaborative filtering, retrieve matching corporate users for specific job seekers, and send recommendation information to specific job seekers based on the job information of relevant corporate users. Specific job seekers include job seekers whose job information has not been paid attention to by companies within a predetermined period of time and job seekers who have difficulty finding employment.

[0030] As a specific implementation of the job matching module, this module is used to perform the following operations: (1) Regularly scan the job-seeking user database, identify job-seeking users whose job search information has not been followed by companies within a predetermined period of time and job-seeking users who have difficulty finding employment as specific job-seeking users, and set labels for specific job-seeking users; (2) For specific job seekers, search and match are performed based on corporate profiles and job seeker profiles through dynamic weighted collaborative filtering to retrieve matching corporate users for the specific job seekers, and recommend information is sent to the specific job seekers based on the job information of the relevant corporate users.

[0031] This module targets job search information and information on people with employment difficulties that have not received attention from companies for a long time. Through automatic system monitoring and intelligent matching, as well as manual operations by back-end managers, it recommends these job search information to suitable companies. Based on the company profile, the enterprise urgency algorithm calculation model is designed. The calculation method is as follows: def calculate_urgency(job_post): # Time urgency time_factor =1 / (1 + math.exp(-8.3*(job_post['days_open']- 7))) # Job Priority priority_map = {'key positions': 1.2, 'ordinary positions': 1.0, 'reserve positions': 8.8} return job_post['vacancy_count']* time_factor * priority_map[job_post['type']]. Finally, the job seekers’ skills and experience are matched with the company’s recruitment needs, and the matching results are pushed to the relevant companies. Back-end managers can manually intervene and make adjustments as needed.

[0032] At the same time, the module will display the matching and recommendation results to job seekers and notify them of relevant recommendation information to enhance users' sense of the platform's services and personalized care.

[0033] The vacancy matching module is used to search and match based on corporate profiles and job seeker profiles through dynamic weighted collaborative filtering. It retrieves matching job seekers for corporate users who have not recruited the required personnel within the set deadline and recommends the job search information of relevant job seekers to corporate users.

[0034] In this embodiment, search and matching based on enterprise profiles and job seeker profiles through dynamic weighted collaborative filtering includes the following operations: (1) Customize the weight of each job application and position information; (2) For enterprise users, set priority weights for each enterprise user based on the release time of job information; (3) Calculate vector similarity based on the feature vectors corresponding to the job application information and the feature vectors corresponding to the job position information. When calculating vector similarity, perform a weighted sum based on the weight of the job application information, the weight of the job position information, and the priority weight of the enterprise user, and perform retrieval and matching based on the final similarity calculation result.

[0035] This module automatically monitors job postings by companies that have not filled their vacancies within the set time limit and intelligently matches and recommends these jobs.

[0036] This module searches the job applicant database for matching candidates based on job requirements and recruitment conditions, identifies qualified candidates, and recommends these candidates to relevant companies. First, company positions may have attributes such as industry, skill requirements, work experience, salary range, and geographic location. These features need to be encoded into vectors to calculate similarity. This is shown below: job_vector = one hot_encode(industry) #Industry category tfidf(skill requirements) # Skill requirements normalize(salary_range) # Salary range geo hash(location. # Geographic location encoding Urgency(days_posted #Urgency indicator.

[0037] At the same time, the characteristics of job seekers also need similar processing, such as skills, work experience, educational background, expected salary, etc. As shown below: candidate_vector=[ skill_graph_embedding(skills) # Skill graph embedding work_exp_score(years), # Experience quantification education_level_mapping(degree), #education level mapping expected salary_normalized #Expected salary.

[0038] Next, the time factor is a key point. The longer a job has been posted, the more urgent it may be, and therefore, it should be given a higher weight when matching. This may require designing a time decay function, such that over time, the weight of the job's urgency gradually increases until it reaches a certain threshold.

[0039] Next, consider the dynamic adjustment strategy. Dynamic weighted collaborative filtering requires integrating multiple factors, such as job requirement match, candidate skill match, and time urgency. A comprehensive similarity formula might be designed to combine these factors with varying weights. For example, job skill match could account for 60%, time urgency 30%, and other factors 10%. Table 1 shows the dynamic adjustment strategy.

[0040] Table 1. Dynamic adjustment strategy

[0041] At the same time, to ensure the accuracy of recommendations, backend administrators can also conduct manual review and adjustments. Once the system completes matching and recommendation, the relevant positions and job seeker information will be displayed on their respective interfaces, allowing both parties to better understand each other's compatibility and increase the success rate of recruitment and job search.

[0042] The system of this embodiment is applicable to the entire process of employment services, including services for the unemployed, assistance to disadvantaged groups, and precise matching of government and enterprises.

[0043] Example 2: The present invention provides a job search and recruitment recommendation method based on dynamic weighted collaborative filtering, which implements intelligent job search and recruitment recommendations between corporate users and job seekers based on the system disclosed in Example 1, including five steps: building user portraits, pushing government information, pushing job information, job matching, and quick matching of vacancies.

[0044] Step S100 constructs a user portrait: constructs a job seeker portrait based on the job-seeking information of the job-seeking user, the job seeker portrait includes the characteristics and feature vectors of the job seeker in various dimensions, and constructs a corporate portrait based on the job information of the corporate user, the corporate portrait includes the characteristics and feature vectors of the corporate user in various dimensions, wherein the characteristics of the job seeker are obtained by screening the job-seeking information, and the characteristics of the corporate user are obtained by screening the job information.

[0045] In this embodiment, the dimensions of the enterprise portrait include industry attributes, recruitment preferences, and historical behaviors. The features under industry attributes include manufacturing and IT services, and the encoding method is One-Hot encoding. The features under recruitment preferences include experience weight and educational requirements, and the encoding method is numerical standardization (0-1). The features under historical behaviors include average feedback time and interview conversion rate, and the encoding method is time series embedding.

[0046] Step S200: Pushing government information: Pushing relevant government information to job seekers based on their job-seeking information, and providing personalized planning consulting services based on their permissions and permission time limit.

[0047] In this embodiment, when pushing government information, relevant policy information is intelligently pushed based on personal job search information, including employment policies, social security policies, and social security preferential policies (such as employment assistance policies for people aged 40-50, people with disabilities, and the unemployed), helping job seekers fully understand the social security system and protect their own rights and interests. At the same time, the system can provide personalized social security planning and consulting services based on the individual's social security payment status, achieving comprehensive coverage of social security services. The time-effectiveness function formula is as follows: ; in, Indicates the number of days remaining in the validity period.

[0048] The policy vectorization algorithm is as follows: class PolicyEncoder:-_init__(self):def self.keywords ={"4858 personnel":8,"disabled people":1,"unemployment insurance":2} # Policy label dictionary def encode(self, policy_text): #Use BERT to extract semantic features vec_semantic = bert_model.encode(policy_text) # Structured feature encoding vec_structured = [1 if kw in policy_text else for kw inself.keywords] return np.concatenate([vec_semantic,vec_structured]). Step S300: Job information push: Search for job seekers who have registered for unemployment on the same day, perform search and matching based on corporate profiles and job seeker profiles through dynamic weighted collaborative filtering, and push job recruitment text messages to job seekers who have registered for unemployment based on the job information of related corporate users.

[0049] For job seekers who are applying for unemployment registration, the job information push module of this embodiment is used to perform the following operations: (1) Calculate the similarity between job search information and job position information based on dynamic weighted collaborative filtering, match relevant enterprise users to job seekers based on the similarity, and send job search and recruitment information to job seekers; (2) Determine the registration status of the job seeker in the service mini-program. If the job seeker has not registered in the service mini-program, guide the job seeker to register and log in to the service mini-program and perform relevant operations. If the job seeker has registered in the service mini-program, match the job seeker with appropriate services based on the job search information posted by the job seeker in the service mini-program and the job search operations performed. Appropriate services include pushing employment guidance, government information, and recruitment information of relevant corporate users.

[0050] In this embodiment, this step regularly identifies individuals who have registered for unemployment that day from the backflow data and sends them job search and recruitment text messages (including 2 to 3 intelligently recommended company information). By comparing personal information and registration status, it automatically determines whether the individual has registered for the service mini-program. If not, a link is sent to guide them through the mini-program for job search, recruitment, and training. If they are registered, the system intelligently analyzes their usage behavior and posted information, and delivers targeted information based on their actual needs. For example, if the user has registered for training, training programs that match their needs will be prioritized. The mini-program also pushes information about intelligently matched companies and related training programs, as well as policies related to unemployment insurance and unemployment benefits. It reminds individuals to apply for social security policies during unemployment, with links to relevant social security application addresses. It also pushes information about individual training needs and upcoming training plans.

[0051] Step S400: Job matching: Based on the corporate portrait and job seeker portrait, search and match through dynamic weight collaborative filtering, retrieve matching corporate users for specific job seekers, and send recommendation information to specific job seekers based on the job information of related corporate users. Specific job seekers include job seekers whose job information has not been paid attention to by the company within a predetermined period of time and job seekers who have difficulty finding employment.

[0052] As a specific implementation of job matching, this step includes the following operations: (1) Regularly scan the job-seeking user database, identify job-seeking users whose job search information has not been followed by companies within a predetermined period of time and job-seeking users who have difficulty finding employment as specific job-seeking users, and set labels for specific job-seeking users; (2) For specific job seekers, search and match are performed based on corporate profiles and job seeker profiles through dynamic weighted collaborative filtering to retrieve matching corporate users for the specific job seekers, and recommend information is sent to the specific job seekers based on the job information of the relevant corporate users.

[0053] This step targets job search information and information on people with employment difficulties that have not received attention from companies for a long time. Through automatic monitoring and intelligent matching by the system, as well as manual operations by back-end managers, these job search information are recommended to suitable companies. Based on the company profile, the enterprise urgency algorithm calculation model is designed, and the calculation method is as follows: def calculate_urgency(job_post): # Time urgency time_factor =1 / (1 + math.exp(-8.3*(job_post['days_open']- 7))) # Job Priority priority_map = {'key positions': 1.2, 'ordinary positions': 1.0, 'reserve positions': 8.8} return job_post['vacancy_count']* time_factor * priority_map[job_post['type']]. Finally, the job seekers’ skills and experience are matched with the company’s recruitment needs, and the matching results are pushed to the relevant companies. Back-end managers can manually intervene and make adjustments as needed.

[0054] At the same time, this step will display the matching and recommendation results to job seekers and notify them of relevant recommendation information to enhance users' sense of the platform's service presence and personalized care.

[0055] Step S500: Vacancy matching: Based on the corporate profile and job seeker profile, dynamic weight collaborative filtering is used to search and match, retrieve matching job seekers for corporate users who have not recruited the required personnel within the set period, and recommend the job search information of relevant job seekers to corporate users.

[0056] In this embodiment, search and matching based on enterprise profiles and job seeker profiles through dynamic weighted collaborative filtering includes the following operations: (1) Customize the weight of each job application and position information; (2) For enterprise users, set priority weights for each enterprise user based on the release time of job information; (3) Calculate vector similarity based on the feature vectors corresponding to the job application information and the feature vectors corresponding to the job position information. When calculating vector similarity, perform a weighted sum based on the weight of the job application information, the weight of the job position information, and the priority weight of the enterprise user, and perform retrieval and matching based on the final similarity calculation result.

[0057] This step targets job postings by companies that have not filled their vacancies within the set time limit. This function will automatically monitor and intelligently match and recommend these jobs.

[0058] This step searches through the candidate database based on job requirements and recruitment criteria, identifying qualified candidates and recommending their information to relevant companies. First, company positions may have attributes such as industry, skill requirements, work experience, salary range, and geographic location. These features need to be encoded into vectors to calculate similarity. This is shown below: job_vector = one hot_encode(industry) #Industry category tfidf(skill requirements) # Skill requirements normalize(salary_range) # Salary range geo hash(location. # Geographic location encoding Urgency(days_posted #Urgency indicator.

[0059] At the same time, the characteristics of job seekers also need similar processing, such as skills, work experience, educational background, expected salary, etc. As shown below: candidate_vector=[ skill_graph_embedding(skills) # Skill graph embedding work_exp_score(years), # Experience quantification education_level_mapping(degree), #education level mapping expected salary_normalized #Expected salary.

[0060] Next, the time factor is a key point. The longer a job has been posted, the more urgent it may be, and therefore, it should be given a higher weight when matching. This may require designing a time decay function, such that over time, the weight of the job's urgency gradually increases until it reaches a certain threshold.

[0061] Next, we need to dynamically adjust the strategy. Dynamically weighted collaborative filtering requires integrating multiple factors, such as job requirement match, candidate skill match, and time urgency. This may require designing a comprehensive similarity formula that combines these factors with varying weights. For example, job skill match could weigh 60%, time urgency 30%, and other factors 10%.

[0062] At the same time, to ensure the accuracy of recommendations, backend administrators can also conduct manual review and adjustments. Once the system completes matching and recommendation, the relevant positions and job seeker information will be displayed on their respective interfaces, allowing both parties to better understand each other's compatibility and increase the success rate of recruitment and job search.

[0063] The method of this embodiment can execute the system disclosed in Example 1 to realize intelligent push notification of job search and recruitment.

[0064] The above is a detailed introduction to the job search and recruitment recommendation system and method based on dynamic weighted collaborative filtering provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for general technical personnel in this field, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A job recruitment recommendation system based on dynamic weighted collaborative filtering, characterized in that: It includes government information push module, job information push module, job search matching module, vacancy matching module, service applet, job search user database and enterprise user database; The service mini-program is designed for corporate users and job seekers, and is used to support job seekers in registering and logging in, posting job information, and performing job search operations. It is also used to provide job seekers with employment guidance, and to support corporate users in registering and logging in, posting job information, and performing recruitment operations. The job seeker database is used to store each job seeker's job search information and a job seeker profile constructed based on the job search information. The job seeker profile includes the job seeker's features and feature vectors in various dimensions. The job seeker's features are obtained by filtering the job search information. The enterprise user database is used to store the job information of each enterprise user and the enterprise profile built based on the job information. The enterprise profile includes the characteristics and feature vectors of the enterprise user in various dimensions. The characteristics of the enterprise user are obtained by filtering the job information. The government information push module is used to push relevant government information to job seekers based on their job search information, and to provide personalized planning consulting services based on their permissions and permission validity period. The job information push module is used to find job seekers who have registered for unemployment on the same day. It searches and matches based on corporate and job seeker profiles and dynamic weighted collaborative filtering. It then pushes job recruitment text messages to job seekers who have registered for unemployment based on the job information of related corporate users. The job search matching module is used to search and match based on corporate and job seeker profiles through dynamic weighted collaborative filtering, retrieve matching corporate users for specific job seekers, and send recommended information to specific job seekers based on the job information of related corporate users. Specific job seekers include job seekers whose job information has not been paid attention to by companies within a predetermined period of time and job seekers who have difficulty finding employment; The vacancy matching module is used to search and match based on corporate profiles and job seeker profiles through dynamic weighted collaborative filtering. It retrieves matching job seekers for corporate users who have not recruited the required personnel within the set deadline and recommends the job search information of relevant job seekers to corporate users.

2. The job search and recruitment recommendation system based on dynamic weighted collaborative filtering according to claim 1 is characterized in that: For job seekers who have registered for unemployment, the job information push module is used to perform the following operations: Calculate the similarity between job search information and job position information based on dynamic weighted collaborative filtering, match relevant corporate users to job seekers based on the similarity, and send job search and recruitment information to job seekers; Determine the job seeker's registration status in the service mini-program. If the job seeker has not registered in the service mini-program, guide the job seeker to register and log in to the service mini-program and perform related operations. If the job seeker has registered in the service mini-program, match the job seeker with appropriate services based on the job search information posted by the job seeker in the service mini-program and the job search operations performed. Appropriate services include pushing employment guidance, government information, and recruitment information from related corporate users.

3. The job search and recruitment recommendation system based on dynamic weighted collaborative filtering according to claim 1 is characterized in that: The job matching module is used to perform the following operations: Regularly scan the job-seeking user database, identify job-seeking users whose job information has not been followed by companies within a predetermined period of time and job-seeking users who have difficulty finding employment as specific job-seeking users, and set labels for these specific job-seeking users; For specific job seekers, search and matching are performed based on corporate profiles and job seeker profiles through dynamic weighted collaborative filtering to retrieve matching corporate users for the specific job seekers, and recommendation information is sent to the specific job seekers based on the job information of related corporate users.

4. The job search and recruitment recommendation system based on dynamic weighted collaborative filtering according to any one of claims 1 to 3, characterized in that: Search and matching is performed based on company and candidate profiles through dynamic weighted collaborative filtering, including the following operations: Customize the weight of each job search information and position information; For enterprise users, set priority weights for each enterprise user based on the release time of job information; Vector similarity calculation is performed based on the feature vectors corresponding to the job application information and the feature vectors corresponding to the position information. When performing vector similarity calculation, a weighted sum is performed based on the weight of the job application information, the weight of the position information, and the priority weight of the enterprise user, and retrieval and matching is performed based on the final similarity calculation result.

5. A job search and recruitment recommendation method based on dynamic weighted collaborative filtering, characterized in that: A job search and recruitment recommendation system based on dynamic weighted collaborative filtering according to any one of claims 1 to 4 is used to implement intelligent job search and recruitment recommendations between enterprise users and job seekers, including the following steps: Build user profiles: Based on the job application information of job seekers, a job seeker profile is constructed. The job seeker profile includes the characteristics and feature vectors of the job seeker in various dimensions. Based on the job application information of enterprise users, a corporate profile is constructed. The corporate profile includes the characteristics and feature vectors of the enterprise users in various dimensions. The characteristics of job seekers are obtained by screening job application information, and the characteristics of enterprise users are obtained by screening job information. Government information push: Push relevant government information to job seekers based on their job search information, and provide personalized planning consulting services based on their permissions and permission validity periods; Job information push: Find job seekers who have registered for unemployment on the same day, perform search and matching based on company and job seeker profiles through dynamic weighted collaborative filtering, and push job recruitment SMS messages to job seekers who have registered for unemployment based on the job information of related company users; Job matching: Based on the company profile and job seeker profile, dynamic weighted collaborative filtering is used to search and match, retrieve matching company users for specific job seekers, and send recommended information to specific job seekers based on the job information of related company users. Specific job seekers include job seekers whose job information has not been paid attention to by companies within a predetermined period of time and job seekers who have difficulty finding employment; Vacancy Matching: Based on corporate and job seeker profiles, dynamic weighted collaborative filtering is used for search and matching. For corporate users who have not recruited the required personnel within the set period, matching job seekers are retrieved, and the job search information of relevant job seekers is recommended to corporate users.

6. The job search and recruitment recommendation method based on dynamic weighted collaborative filtering according to claim 5, characterized in that: For job seekers who have registered for unemployment, perform the following operations: Calculate the similarity between job search information and job position information based on dynamic weighted collaborative filtering, match relevant corporate users to job seekers based on the similarity, and send job search and recruitment information to job seekers; Determine the job seeker's registration status in the service mini-program. If the job seeker has not registered in the service mini-program, guide the job seeker to register and log in to the service mini-program and perform related operations. If the job seeker has registered in the service mini-program, match the job seeker with appropriate services based on the job search information posted by the job seeker in the service mini-program and the job search operations performed. Appropriate services include pushing employment guidance, government information, and recruitment information from related corporate users.

7. The job search and recruitment recommendation method based on dynamic weighted collaborative filtering according to claim 5, characterized in that: Job matching includes the following operations: Regularly scan the job-seeking user database, identify job-seeking users whose job information has not been followed by companies within a predetermined period of time and job-seeking users who have difficulty finding employment as specific job-seeking users, and set labels for these specific job-seeking users; For specific job seekers, search and matching are performed based on corporate profiles and job seeker profiles through dynamic weighted collaborative filtering to retrieve matching corporate users for the specific job seekers, and recommendation information is sent to the specific job seekers based on the job information of related corporate users.

8. The job search and recruitment recommendation method based on dynamic weighted collaborative filtering according to claim 5, characterized in that: Search and matching is performed based on company and candidate profiles through dynamic weighted collaborative filtering, including the following operations: Customize the weight of each job search information and position information; For enterprise users, set priority weights for each enterprise user based on the release time of job information; Vector similarity calculation is performed based on the feature vectors corresponding to the job application information and the feature vectors corresponding to the position information. When performing vector similarity calculation, a weighted sum is performed based on the weight of the job application information, the weight of the position information, and the priority weight of the enterprise user, and retrieval and matching is performed based on the final similarity calculation result.

Images