Intelligent position recommendation method and system based on artificial intelligence

Abstract

The invention relates to the technical field of intelligent position information recommendation, in particular to an intelligent position recommendation method based on artificial intelligence, which comprises the following steps: collecting basic information and behavior data of a user, and constructing a user portrait; acquiring position information, extracting key information to establish a position information base, and introducing an expert system to supplement the position information; understanding implicit significance of position description through a natural language processing method and a semantic analysis method; performing feature extraction on the user portrait and the position information, and constructing a feature data set; training a recommendation model; user feedback is actively inquired through an active learning algorithm; and generating a position recommendation list for the user through the trained recommendation model. The method optimizes description of position information, is more convenient for users to understand, and solves the problem that accurate user portraits cannot be constructed due to lack of data for users and new positions in specific fields through an active learning algorithm and introduction of an expert system.

Description

Technical Field

[0001] The present invention relates to the technical field of job information intelligent recommendation, and in particular to an artificial intelligence-based job information intelligent recommendation method and system. Background Art

[0002] With the development of big data and artificial intelligence technology, applying big data and artificial intelligence to the matching of positions and job seekers can improve the accuracy of job recommendations, enhance the application experience and success rate.

[0003] For example, the patent with announcement number CN109241446B discloses a method and system for job recommendation, including the following steps S1: obtaining user access data and generating a data matrix; step S2: executing a supervised learning algorithm on the data matrix; step S3: executing an unsupervised learning algorithm on the data matrix and generating prediction result data; step S4: generating job pre-recommendation data based on the prediction result data and character portrait data; according to the device information, access content, operation behavior, etc. of the visitor, through the machine learning method of job pre-recommendation, the system automatically learns, completes user analysis, and accurately recommends job information; allows users to discover more relevant and interesting job information, and improves the user's viewing content. It combines character portraits and machine learning algorithms.

[0004] Combining character portraits with machine learning algorithms is a common method used by existing intelligent job recommendation systems. This method is highly dependent on data and requires accurate data to build character portraits. However, for users in some specific fields or some emerging professions, the lack of data will affect the construction of character portraits and the construction and training of recommendation models. This results in the recommended positions not meeting user requirements, which in turn affects the recruitment process. Summary of the invention

[0005] The present invention aims to provide a technical solution to solve the problems in the background technology.

[0006] The technical solution of the present invention is: an artificial intelligence-based job intelligent recommendation method, the method comprising the following steps:

[0007] Collect users’ basic information and behavior data to build user portraits;

[0008] Obtaining job information, extracting key information to establish a job information database, and introducing an expert system to supplement the job information, wherein the key information includes job title and job description;

[0009] Use natural language processing and semantic analysis methods to understand the implicit meaning of job descriptions, identify inaccurate or ambiguous descriptions, and generate feedback information;

[0010] Extract features from user portraits and job information to build a feature data set;

[0011] Train the recommendation model to learn the matching relationship between users and positions;

[0012] Actively ask for user feedback through active learning algorithms;

[0013] Generate a job recommendation list for users through the trained recommendation model;

[0014] Evaluate the recommendation effect through evaluation indicators;

[0015] The evaluation index includes one or more of click-through rate and user satisfaction, the basic information of the user includes educational background, work experience, and skill certificates, the behavior data of the user includes browsing history and job application history of the user, and the recommendation model is a collaborative filtering model;

[0016] The collecting of basic information and behavior data of users is as follows: collecting basic information and behavior data of users through registration information, user behavior logs, and questionnaire surveys.

[0017] Preferably, the step of obtaining job information and extracting key information to establish a job information database comprises the following steps:

[0018] Use web crawlers to crawl job information;

[0019] Extracting key information from the job information by using a text processing algorithm, wherein the key information includes one or more of the company name, the work location, and the required skills;

[0020] Supplement and revise job information with the knowledge of industry experts to improve its accuracy and completeness.

[0021] Preferably, the introduction of an expert system to supplement the job information comprises the following steps:

[0022] Build an expert database to include industry experts’ contact information, professional fields, and skills;

[0023] Use keyword extraction algorithms to extract job keywords from job description information and / or feedback information, match them with expert professional field information, push feedback information to matched experts, and receive feedback information from experts to supplement or modify job information;

[0024] The matching with the expert professional field information is achieved by calculating the similarity between the keywords and the professional field information.

[0025] Preferably, the step of extracting features from user portraits and job information to construct a feature data set includes the following steps:

[0026] The working years, skill proficiency, relevant skill level and job interest level data are extracted from the user's basic information and behavior data, and after normalization, a user information feature vector is formed: i=1、2、Λ、n;X 1 Indicates working years, X 2 Indicates skill proficiency, X 3 Indicates the relevant skill level, X 4 Indicates the level of interest in the position;

[0027] Constructing a user feature vector set from multiple user information feature vectors Where n represents the number of users, n≥3;

[0028] Extract working years, working location information, and skill requirement information from the job information, and normalize them to form a job information feature vector j=1、2、Λ、m;Y 1 Indicates years of working experience, Y 2 Indicates work location information, Y 3 Indicates skill requirement information;

[0029] Multiple job information feature vectors constitute a job information feature set:

[0030] Where m represents the number of positions, m≥3;

[0031] The feature data set includes a user feature vector set and a position information feature set.

[0032] Preferably, the training of the recommendation model so that the recommendation model learns the matching relationship between users and positions includes the following steps:

[0033] A plurality of historical user data are obtained from a database, wherein the historical user data includes a historical user feature vector and a historical job recommendation list corresponding to the historical user feature vector, the historical data are divided into a training set and a test set, and a recommendation model is trained using a collaborative filtering algorithm, specifically:

[0034] Select two historical user feature vectors from the training set, one as the feature vector of the user to be recommended, and the other as the feature vector of the target user;

[0035] The cosine similarity algorithm is used to calculate the similarity between the feature vector of the user to be recommended and the feature vector of the target user. If the similarity value is greater than 96%, it is considered that the user corresponding to the feature vector of the target user has the same job interest as the user to be recommended.

[0036] The target user's job recommendation list is used as a list of recommended jobs to be determined and recommended to the user to be recommended;

[0037] The list of recommended positions to be determined is compared with the list of recommended positions corresponding to the feature vector of the user to be recommended, and the duplication degree is obtained through the duplication checking algorithm. If the duplication degree is greater than 90%, the list to be recommended is considered valid.

[0038] Preferably, the method of understanding the implicit meaning of the job description, identifying inaccurate or ambiguous descriptions, and generating feedback information by natural language processing and semantic analysis methods includes the following steps:

[0039] Preprocess the job description text: remove irrelevant characters and stop words; break the text into words or phrases; identify the word's part of speech; unify different word forms;

[0040] Perform feature extraction: identify job skills and job requirements in job descriptions;

[0041] Perform topic modeling: Use the LDA algorithm to identify topics in the text;

[0042] Conduct semantic analysis: analyze the dependency relationship between words in a sentence; extract organizations, names, and location entities; and determine the sentiment of the text;

[0043] Understanding the hidden meaning:

[0044] Use contextual information to understand the specific meaning of vocabulary;

[0045] Identify requirements or characteristics that are not directly expressed;

[0046] Identify ambiguous words in descriptions;

[0047] Check for inconsistencies in descriptions, such as conflicting job requirements;

[0048] Feedback information is generated, the feedback information including a list of questions and position description text.

[0049] Preferably, evaluating the recommendation effect by using evaluation indicators comprises the following steps:

[0050] Use website analysis tools to record users' click behavior on positions in the recommended job list, specifically to calculate the click-through rate (CTR);

[0051] The calculated click-through rate is compared with the preset target value C. If |CTR-C|≤W, the position recommended by the recommendation model is considered reasonable. Otherwise, the click-through rate is stored and a prompt message indicating that the click-through rate is unqualified is generated, where W is the preset click threshold.

[0052] Preferably, actively asking for user feedback through an active learning algorithm comprises the following steps:

[0053] Obtain the positions with unsatisfactory click-through rates in the job list, and feed back the job descriptions of these positions to users to request feedback;

[0054] Obtain user feedback information, identify professional field information of the position in the user feedback information, send the user feedback information to a matching expert, and obtain feedback information from the expert;

[0055] Replace or supplement the original job description based on user feedback and expert feedback, and update the job information in the database.

[0056] The present invention also provides an artificial intelligence-based intelligent job recommendation system, comprising a processor, a memory, a communication module and a display terminal, and the system executes the artificial intelligence-based intelligent job recommendation method.

[0057] The present invention also provides a computer-readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the artificial intelligence-based intelligent job recommendation method.

[0058] Beneficial effects of the present invention:

[0059] 1. The present invention extracts features from user portraits and job information and constructs a feature data set. By constructing a feature data set containing user portraits and job information, it can be used for subsequent tasks such as user portrait analysis and job matching, helping us to better understand and meet user needs.

[0060] 2. The present invention introduces an expert system to supplement job information. By constructing an expert database and using a keyword extraction algorithm to match the expert's professional field with job description information, expert knowledge can be effectively introduced to supplement and correct job information. This method not only improves the quality of job information, but also enhances the professional level and user trust of the job recommendation system. Through the participation of experts, the problem of some new positions lacking detailed data information and being unable to build accurate user portraits can be solved. At the same time, it can also ensure the consistency of job information with industry standards and actual needs, thereby improving the success rate of recruitment.

[0061] 3. In the present invention, if the click-through rate does not meet the conditions, the click-through rate data is stored, and a prompt message indicating that the click-through rate is unqualified is generated, and the user feedback is asked through an active learning algorithm; the user's feedback information is sent to matching experts, who should be selected based on their professional field information; the expert's feedback information is received, which may include corrections, supplements or other suggestions for the job description; the original job description is replaced or supplemented based on the feedback information from the user and the expert, so that the job description is more professional and accurate so that the applicant can better understand it; by updating the job information in the database, more accurate data is provided for feature extraction, the effect of recommendation model training is improved, and the accuracy of job recommendations is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a flow chart of an artificial intelligence-based intelligent job recommendation method of the present invention. DETAILED DESCRIPTION

[0063] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious variations. The basic principles of the present invention defined in the following description can be applied to other embodiments, variations, improvements, equivalents, and other technical solutions that do not deviate from the spirit and scope of the present invention.

[0064] It is to be understood that the term "one" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element may be one, while in another embodiment, the number of the element may be multiple, and the term "one" should not be understood as a limitation on the quantity.

[0065] refer to Figure 1 The present invention provides a technical solution: an artificial intelligence-based job intelligent recommendation method, the method comprising the following steps:

[0066] Collect users’ basic information and behavior data to build user portraits;

[0067] Obtaining job information, extracting key information to establish a job information database, and introducing an expert system to supplement the job information; the key information includes job title and job description;

[0068] Use natural language processing and semantic analysis methods to understand the implicit meaning of job descriptions, identify inaccurate or ambiguous descriptions, and generate feedback information;

[0069] Extract features from user portraits and job information to build a feature data set;

[0070] Train the recommendation model to learn the matching relationship between users and positions;

[0071] Actively ask for user feedback through active learning algorithms;

[0072] Generate a job recommendation list for users through the trained recommendation model;

[0073] Evaluate the recommendation effect through evaluation indicators; the evaluation indicators include one or more of click-through rate and user satisfaction;

[0074] The collection of basic information and behavior data of users includes: collecting basic information and behavior data of users through registration information, user behavior logs, and questionnaires;

[0075] The basic information of the user includes educational background, work experience, and skill certificates; the behavior data of the user includes browsing history and job application history of the user; and the recommendation model is a collaborative filtering model.

[0076] The collecting of basic information and behavior data of users is as follows: collecting basic information and behavior data of users through registration information, user behavior logs, and questionnaire surveys.

[0077] First of all, we can obtain some basic information about the user through registration information. This information, such as name, age, gender, and contact information, can help us understand the user's basic characteristics and background.

[0078] Secondly, we can collect user behavior data through user behavior logs. When users use our platform or application, they will generate a series of operation records, such as browsing pages, clicking buttons, submitting forms, etc. By analyzing these behavior logs, we can understand users' usage habits, preferences and needs, so as to provide them with better personalized services.

[0079] In addition, we can also collect basic information and behavior data of users through questionnaires. Questionnaires are a method of asking questions directly to users. A series of questions can be designed to understand the user's educational background, work experience, skill certificates, etc. At the same time, we can also ask users about their browsing history, application history, etc. to further understand their needs and preferences.

[0080] We can collect basic information and behavior data of users through registration information, user behavior logs and questionnaires. These data will provide us with valuable insights, help us better understand users, and provide them with more accurate services and recommendations.

[0081] The process of obtaining job information, extracting key information and establishing a job information database is as follows:

[0082] Crawl job information: Determine a list of recruitment websites to crawl, and develop or use ready-made web crawler tools to crawl job information, including the time of job posting, title, job details page URL, etc.

[0083] Extract key information and build a job information database:

[0084] Apply natural language processing techniques, such as regular expressions, word segmentation, named entity recognition, etc., to extract information from job detail pages;

[0085] Text processing algorithms can be used to extract key information from the crawled page content, such as job title, company name, job location, job description, required skills, etc.

[0086] Expert database construction: Collect the contact information of industry experts, including their professional fields and skills. Present the extracted job information to experts in related fields and ask them to provide feedback. Based on the feedback from experts, supplement and revise the job information to ensure the accuracy and completeness of the information.

[0087] Database design: Design a database structure suitable for storing job information to ensure data consistency and queryability. Import the extracted and corrected data into the database to establish a job information database. Set up a regular crawling and updating process to ensure that the data in the job information database is up to date.

[0088] By using web crawlers to crawl job information from websites and combining text processing algorithms to extract key information, a job information database can be effectively established. At the same time, the accuracy and completeness of job information can be improved through the knowledge and feedback of industry experts. Such a job information database not only provides job seekers with a rich selection of positions, but also provides effective talent matching services for recruiting companies.

[0089] Introduce an expert system to supplement job information, specifically:

[0090] 1. Build an expert database: collect the contact information, professional fields and skills of industry experts to ensure the accuracy and update of information.

[0091] 2. Keyword extraction and matching: Select keyword extraction algorithms, such as TF-IDF, LDA topic model, etc., to extract keywords from job descriptions. Apply these algorithms to job description information and user feedback information to extract representative key fields. Find the most matching expert by calculating the similarity between the extracted keywords and the expert's professional field information.

[0092] 3. Expert feedback loop: Push user feedback information or job descriptions to matching experts to request their professional opinions and feedback. Based on the expert’s feedback, supplement or modify the job information to improve the accuracy and completeness of the information.

[0093] By building an expert database and using a keyword extraction algorithm to match the expert's professional field with the job description information, expert knowledge can be effectively introduced to supplement and correct job information. This method not only improves the quality of job information, but also enhances the professional level and user trust of the job recommendation system. At the same time, through the participation of experts, the consistency of job information with industry standards and actual needs can be ensured, thereby improving the success rate of recruitment.

[0094] Extract features from user portraits and job information and build a feature dataset, including the following steps:

[0095] Data on years of work experience, skill proficiency, relevant skill level, and job interest level are extracted from the user's basic information and behavior data. These data are important components of user portraits and can reflect the user's professional background and ability level.

[0096] After normalization, the extracted data is used to form a user information feature vector i=1, 2, Λ, n; among them, X 1 Indicates working years, X 2 Indicates skill proficiency, X 3 Indicates the relevant skill level, X 4 Indicates the level of interest in the position.

[0097] Multiple user information feature vectors constitute a user feature vector set Where n represents the number of users, n≥3; this set contains the feature vectors of all users, which can be used for subsequent data analysis and model training. The years of work experience, work location information, and skill requirements are extracted from the job information. These data are important components of job profiles and can reflect the needs and requirements of the job.

[0098] After normalization, the extracted data is used to form a position information feature vector j=1, 2, Λ, m. Among them, Y 1 Indicates years of working experience, Y 2 Indicates work location information, Y 3 Represents skill requirement information. Multiple job information feature vectors constitute a job information feature set Where m represents the number of positions, m ≥ 3. This set contains the feature vectors of all positions and can be used for subsequent data analysis and recommendation model training.

[0099] Through the above steps, we have successfully built a feature dataset containing user portraits and job information. This dataset can be used for subsequent tasks such as user portrait analysis and job matching, helping us better understand and meet user needs.

[0100] The process of training the recommendation model:

[0101] 1. Data preparation: Get historical user data from the database, including user feature vectors (such as skills, experience, educational background, etc.) and the corresponding historical job recommendation list. Divide the historical data into a training set and a test set, usually with a ratio of 80% training set and 20% test set.

[0102] 2. Training model: Select two historical user feature vectors from the training set, one as the feature vector of the user to be recommended, and the other as the feature vector of the target user. After normalization, use the cosine similarity algorithm to calculate the similarity between the two users:

[0103] If the calculated similarity value is greater than 96%, it is considered that the two users have a high degree of similarity in job interests; otherwise, the two users are considered to be dissimilar.

[0104] For example, the user information of user 1 (user to be recommended) is: working experience is 3 years, skill proficiency level is level 4, related skill level is level 2, and position interest level is level 1; the constructed feature vector is:

[0105]

[0106] The information of user 2 (target user) is: working experience is 2 years, skill proficiency level is level 2, relevant skill level is level 3, and job interest level is level 2; the constructed feature vector is:

[0107] The maximum value in the two vectors is 4 and the minimum value is 1, and they are normalized using the min-max normalization formula:

[0108] The normalized feature vector of user 1 is:

[0109] The normalized feature vector of user 2 is:

[0110] The dot product of the two eigenvectors is: 0.67×0.33+0.33+0.67×0.33≈0.7722;

[0111] Calculate cosine similarity: This means that User 2 is not similar to User 1, and User 2's recommended job list will not be pushed to User 1.

[0112] 3. Generate recommended job list:

[0113] The target user's job recommendation list is used as a list of recommended jobs to be determined and recommended to the user to be recommended. The list of recommended jobs to be determined is compared with the historical job recommendation list corresponding to the feature vector of the user to be recommended, and the degree of repetition is calculated:

[0114]

[0115] If the repetition rate is greater than 90%, the list to be recommended is considered valid, otherwise a new target user is matched.

[0116] The principle of using the similarity between positions to predict positions that the user may be interested in is similar to this, and will not be described in detail in this embodiment.

[0117] According to the results of the recommendation effect evaluation, adjust the parameters of similarity calculation or recommendation algorithm to optimize the performance of the recommendation model. Use the cross-validation method to evaluate the stability and generalization ability of the recommendation model. Finally, use the test set to evaluate the accuracy, recall and other performance indicators of the recommendation model.

[0118] By using collaborative filtering algorithms and cosine similarity calculations, the recommendation model can be effectively trained to learn the matching relationship between users and positions. This method can help the system identify user groups with similar job interests and recommend suitable positions for them. Through continuous model optimization and testing, the accuracy of the recommendation system and user satisfaction can be further improved.

[0119] Steps to understand the implicit meaning in job descriptions and generate feedback information through natural language processing and semantic analysis methods:

[0120] Understand the implicit meaning in the job description: Remove irrelevant characters and stop words from the text, which usually do not carry useful information and affect the analysis results. Break the text into words or phrases, which is the basis of most natural language processing tasks. Identify the part of speech of words (such as noun, verb, adjective, etc.), which helps to understand sentence structure and semantics. Unify different word forms and reduce the diversity of vocabulary, which helps clustering and topic modeling.

[0121] Feature extraction: Identify the skills and requirements required for the position from the job description, which may include specific technical abilities, work experience, educational background, etc.

[0122] The Latent Dirichlet Allocation (LDA) algorithm is used to identify topics in the text, which helps in understanding the core content and context of the job description.

[0123] Analyze dependencies between words in a sentence, which helps understand the structure and semantics of the sentence. Extract entities such as organizations, names, and locations, which are important for understanding the specific context of the position. Determine the sentiment of the text, such as positive or negative, which can provide insights into job satisfaction or expectations. Use contextual information to understand the specific meaning of words, especially in different industry contexts.

[0124] Identify requirements or characteristics that are not directly expressed in descriptions, such as "quick learning ability" may mean "inexperience is also acceptable." Identify vague terms in descriptions, such as "good communication skills" may need further clarification to refer to oral or written communication.

[0125] Check for contradictory information in the description, such as requiring both "entry-level experience" and "five years of management experience." Based on the above analysis, generate a list of questions to point out inaccuracies, ambiguities, or contradictions in the job description. Job Description Text: Provide the original job description text for reference along with the list of questions.

[0126] By combining natural language processing and semantic analysis methods, we can gain a deep understanding of the implicit meaning in job descriptions and identify inaccurate or ambiguous descriptions.

[0127] Use website analysis tools to record users’ click behaviors on recommended positions and use them to evaluate the rationality of the recommendation model. The following are the detailed steps and explanations:

[0128] Use website analytics tools to record user click behavior: Use website analytics tools (such as Google Analytics) to track user click behavior on each job in the recommended job list.

[0129] The click-through rate (CTR) is calculated based on the collected data: (total number of job clicks / total number of job impressions) × 100%.

[0130] The calculated click-through rate is compared with the preset target value C or industry standard. If |CTR-C|≤W, the recommended position is considered reasonable. If the click-through rate does not meet the preset conditions, that is, |CTR-C|>W, the click-through rate value is stored and a prompt message indicating that the click-through rate is unqualified is generated.

[0131] The application of the active learning algorithm in this embodiment includes the following steps:

[0132] Identify the prompt information of unqualified click-through rate and take the corresponding job recommendation list as the optimization object;

[0133] Build a feedback interface to provide users with the job descriptions corresponding to these positions and request users to provide feedback;

[0134] Collect user feedback information through online questionnaires, feedback forms or direct user communication, analyze user feedback information through NLP technology, identify the professional field information of the position and the inaccurate or ambiguous parts in the job description, and send this part of information to matching experts;

[0135] Receive feedback from experts, which may include corrections, additions or other suggestions to the job posting;

[0136] Based on the feedback from users and experts, the original job description is replaced or supplemented, thereby updating the job information in the database to ensure the accuracy and timeliness of the job information.

[0137] By using website analysis tools to record and analyze user click behaviors, the performance of the recommendation model can be effectively evaluated, and whether further user and expert feedback is needed can be determined based on whether the click rate meets the standard. This not only improves the quality of job information, but also enhances the interaction between users and the system, thereby improving the overall user experience and satisfaction.

[0138] The present invention also provides an artificial intelligence-based intelligent job recommendation system, comprising a processor, a memory, a communication module and a display terminal, and the system executes the artificial intelligence-based intelligent job recommendation method.

[0139] The present invention also provides a computer-readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the artificial intelligence-based intelligent job recommendation method.

[0140] The embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. The embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part, and / or installed from a removable medium. When the computer program is executed by the central processing unit (CPU), the above-mentioned functions defined in the method of the present application are executed. It should be noted that the above-mentioned computer-readable medium of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, a system, device or device of an electrical, magnetic, optical, electromagnetic, infrared segment, or semiconductor, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more wire segments, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, electrical wire, optical cable, RF, etc., or any suitable combination of the foregoing.

[0141] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0142] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments. Without departing from the principles, the implementation methods of the present invention may be deformed or modified in any way.

Claims

1. An artificial intelligence-based job intelligent recommendation method, characterized in that: The method comprises the following steps: Collect users’ basic information and behavior data to build user portraits; Obtaining job information, extracting key information to establish a job information database, and introducing an expert system to supplement the job information, wherein the key information includes job title and job description; Use natural language processing and semantic analysis methods to understand the implicit meaning of job descriptions, identify inaccurate or ambiguous descriptions, and generate feedback information; Extract features from user portraits and job information to build a feature data set; Train the recommendation model to learn the matching relationship between users and positions; Actively ask for user feedback through active learning algorithms; Generate a job recommendation list for users through the trained recommendation model; Evaluate the recommendation effect through evaluation indicators; The evaluation index includes one or more of click-through rate and user satisfaction, the basic information of the user includes educational background, work experience, and skill certificates, the behavior data of the user includes browsing history and job application history of the user, and the recommendation model is a collaborative filtering model; The collecting of basic information and behavior data of users is as follows: collecting basic information and behavior data of users through registration information, user behavior logs, and questionnaire surveys.

2. The method for intelligent job recommendation based on artificial intelligence according to claim 1, characterized in that: The step of obtaining job information and extracting key information to establish a job information database includes the following steps: Use web crawlers to crawl job information; Extracting key information from the job information by using a text processing algorithm, wherein the key information includes one or more of the company name, the work location, and the required skills; Supplement and revise job information with the knowledge of industry experts to improve its accuracy and completeness.

3. The method for intelligent job recommendation based on artificial intelligence according to claim 1, characterized in that: The introduction of the expert system to supplement the job information includes the following steps: Build an expert database to include industry experts’ contact information, professional fields, and skills; Use keyword extraction algorithms to extract job keywords from job description information and / or feedback information, match them with expert professional field information, push feedback information to matched experts, and receive feedback information from experts to supplement or modify job information; The matching with the expert professional field information is achieved by calculating the similarity between the keywords and the professional field information.

4. The method for intelligent job recommendation based on artificial intelligence according to claim 1, characterized in that: The feature extraction of user portraits and position information and the construction of feature data sets include the following steps: The working years, skill proficiency, relevant skill level and job interest level data are extracted from the user's basic information and behavior data, and after normalization, a user information feature vector is formed: Vρ i ′=(X1, X2, X3, X4); i=1, 2, Λ, n; X1 represents years of work experience, X2 represents skill proficiency, X3 represents relevant skill level, and X4 represents job interest level; A user feature vector set A = {Vρ1′, Vρ2′, ..., Vρ n ′}; where n represents the number of users, n≥3; Extract working years, working location information, and skill requirement information from the job information, and after normalization, form a job information feature vector Uρ j ′=(Y1, Y2, Y3), j=1, 2, Λ, m; Y1 represents the years of work experience, Y2 represents the work location information, and Y3 represents the skill requirements information; Multiple job information feature vectors constitute a job information feature set: B={Uρ1′, Uρ2′,..., Uρ′ m }, where m represents the number of positions, m ≥ 3; The feature data set includes a user feature vector set and a position information feature set.

5. The method for intelligent job recommendation based on artificial intelligence according to claim 1, characterized in that: The training of the recommendation model so that the recommendation model learns the matching relationship between users and positions includes the following steps: A plurality of historical user data are obtained from a database, wherein the historical user data includes a historical user feature vector and a historical job recommendation list corresponding to the historical user feature vector, the historical data are divided into a training set and a test set, and a recommendation model is trained using a collaborative filtering algorithm, specifically: Select two historical user feature vectors from the training set, one as the feature vector of the user to be recommended, and the other as the feature vector of the target user; The cosine similarity algorithm is used to calculate the similarity between the feature vector of the user to be recommended and the feature vector of the target user. If the similarity value is greater than 96%, it is considered that the user corresponding to the feature vector of the target user has the same job interest as the user to be recommended. The target user's job recommendation list is used as a list of recommended jobs to be determined and recommended to the user to be recommended; The list of recommended positions to be determined is compared with the list of recommended positions corresponding to the feature vector of the user to be recommended, and the duplication degree is obtained through the duplication checking algorithm. If the duplication degree is greater than 90%, the list to be recommended is considered valid.

6. The method for intelligent job recommendation based on artificial intelligence according to claim 1, characterized in that: The method of using natural language processing and semantic analysis to understand the implicit meaning of the job description, identify inaccurate or ambiguous descriptions, and generate feedback information includes the following steps: Preprocess the job description text: remove irrelevant characters and stop words; break the text into words or phrases; identify the word's part of speech; unify different word forms; Perform feature extraction: identify job skills and job requirements in job descriptions; Perform topic modeling: Use the LDA algorithm to identify topics in the text; Conduct semantic analysis: analyze the dependency relationship between words in a sentence; extract organizations, names, and location entities; and determine the sentiment of the text; Understand implicit meaning: Use contextual information to understand the specific meaning of words; Identify requirements or characteristics that are not directly expressed; Identify ambiguous words in descriptions; Check for inconsistencies in descriptions, such as conflicting job requirements; Feedback information is generated, the feedback information including a list of questions and position description text.

7. The method for intelligent job recommendation based on artificial intelligence according to claim 1, characterized in that: The evaluation of the recommendation effect by using the evaluation index comprises the following steps: Use website analysis tools to record users' click behavior on positions in the recommended job list, specifically to calculate the click-through rate (CTR); The calculated click-through rate is compared with the preset target value C. If |CTR-C|≤W, the position recommended by the recommendation model is considered reasonable. Otherwise, the click-through rate is stored and a prompt message indicating that the click-through rate is unqualified is generated, where W is the preset click threshold.

8. The method for intelligent job recommendation based on artificial intelligence according to claim 1, characterized in that: The method of actively asking for user feedback through an active learning algorithm includes the following steps: Obtain the positions with unsatisfactory click-through rates in the job list, and feed back the job descriptions of these positions to users to request feedback; Obtain user feedback information, identify professional field information of the position in the user feedback information, send the user feedback information to a matching expert, and obtain feedback information from the expert; Replace or supplement the original job description based on user feedback and expert feedback, and update the job information in the database.

9. An artificial intelligence-based job recommendation system, comprising a processor, a memory, a communication module and a display terminal, characterized in that: The system executes an artificial intelligence-based intelligent job recommendation method as described in any one of claims 1-8 above.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement an artificial intelligence-based intelligent job recommendation method as described in any one of claims 1 to 8.

Images