Human and post matching system based on artificial intelligence

Abstract

The invention relates to the technical field of data processing, in particular to an artificial intelligence-based person and post matching system, which comprises an acquisition module, a judgment module, an evaluation module, a determination module, an adjustment module, a correction module and a matching module. The method comprises the following steps: acquiring various key data, judging concerned posts in combination with browsing frequency and browsing speed according to keywords of resumes and recruitment, evaluating recruitment indexes and resume indexes by utilizing a model, and determining a candidate post list in combination with historical training frequency and historical promotion speed; the resume index or the similarity threshold value is dynamically adjusted according to the delivery and interview success rate of the candidate post list, the candidate post list is corrected according to the number of times of adjustment, and the recommended post is matched, so that the recommendation result better fits the target and the post, the matching accuracy and effectiveness are improved, and the user experience is improved. The problems of inaccurate matching and slow response speed caused by incomplete data acquisition and excessive dependence on a static model are effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a human-post matching system based on artificial intelligence. Background Art

[0002] In today's rapidly developing labor market, enterprises are facing increasingly fierce competition and need to find suitable talents efficiently and accurately to maintain their competitive advantages. At the same time, job seekers also expect to quickly find positions that match their abilities and interests. Traditional human-post matching methods often rely on manual resume screening and interviews, which are not only time-consuming and laborious, but also easily affected by subjective factors, resulting in low matching efficiency and brain drain. Therefore, a human-post matching system based on artificial intelligence has emerged, aiming to improve the recruitment efficiency and quality through advanced technical means and achieve precise matching between enterprises and job seekers.

[0003] The patent document with the publication number CN116578933A discloses a talent resume matching method and system based on artificial intelligence big data. The method includes: S100, obtaining resume information and recruitment requirement information, and respectively collecting a first data set in the resume information and a second data set in the recruitment requirement information book through a preset intelligent model; S200, respectively performing big data collection according to the first data set and the second data set to obtain a first associated data set and a second associated data set; S300, respectively inputting the first associated data set and the second associated data set into a training model to obtain a first label tree and a second label tree; and evaluating the matching degree between the resume information and the recruitment requirement information according to the comparison between the first label tree and the second label tree.

[0004] It can be seen that the talent resume matching method based on artificial intelligence big data has the following problems: relying only on static resume information and recruitment requirement information, it is impossible to comprehensively evaluate the suitability of job seekers, resulting in inaccurate matching results; the static model causes the matching results to be unable to reflect the changes in the market and the behavior of job seekers in a timely manner, reducing the response speed and matching accuracy of the system; the matching results rely too much on static information and cannot comprehensively evaluate the comprehensive abilities and potentials of job seekers, thus affecting the accuracy and effectiveness of the matching. Summary of the Invention

[0005] Therefore, the present invention provides a human-post matching system based on artificial intelligence to overcome the problems of inaccurate matching and slow response speed caused by incomplete data collection and excessive reliance on static models in the prior art through multi-dimensional real-time data and a dynamic adjustment mechanism.

[0006] To achieve the above object, the present invention provides a human-post matching system based on artificial intelligence, including:

[0007] A collection module is used to collect, in real time after a target uploads a resume containing resume keywords, the recruitment keywords, historical training frequencies, historical promotion speeds, the target's browsing frequencies, browsing speeds, application success rates, and interview success rates of each job to be matched in the job database.

[0008] A determination module, which is connected to the collection module, is used to determine a number of concerned jobs according to the resume keywords, the recruitment keywords, a preset similarity threshold, the browsing frequency, and the browsing speed.

[0009] An evaluation module, which is respectively connected to the collection module and the determination module, is used to evaluate a resume index based on a preset artificial intelligence model according to the resume keywords, and to evaluate a recruitment index according to the recruitment keywords of each concerned job.

[0010] A determination module, which is respectively connected to the collection module and the evaluation module, is used to determine a list of candidate jobs according to the recruitment index, the resume index, the historical training frequency, and the historical promotion speed of each concerned job.

[0011] An adjustment module, which is respectively connected to the collection module and the determination module, is used to adjust the resume index or adjust the preset similarity threshold according to the list of candidate jobs, the application success rate, and the interview success rate.

[0012] A correction module, which is connected to the adjustment module and the determination module, is used to correct the list of candidate jobs according to the number of times of adjusting the resume index and the number of times of adjusting the preset similarity threshold within a preset correction duration to form a corrected job list.

[0013] A matching module, which is connected to the correction module, is used to match a number of recommended jobs according to the corrected job list.

[0014] Further, the determination module includes:

[0015] A vector conversion unit is used to convert the recruitment keywords into vector form to obtain a number of recruitment key vectors, and to convert the resume keywords into vector form to obtain a resume key vector.

[0016] A temporary determination unit, which is connected to the vector conversion unit, is used to determine a number of temporary jobs according to the resume key vector and each recruitment key vector.

[0017] A browsing frequency determination unit, which is connected to the temporary determination unit, is used to determine a number of marked jobs according to the browsing frequency of each temporary job.

[0018] A browsing speed determination unit, which is connected to the browsing frequency determination unit, is used to determine a marked position as a concerned position when the browsing speed of each of the marked positions is less than a preset browsing speed threshold, so as to determine a number of concerned positions.

[0019] Further, the temporary determination unit includes:

[0020] A similarity calculation sub-unit, which is used to calculate the cosine similarity between the resume key vector and each of the recruitment key vectors to obtain a number of similarities;

[0021] A temporary determination sub-unit, which is connected to the similarity calculation sub-unit, is used to determine the to-be-matched position as a temporary position when the similarity is greater than a preset similarity threshold, so as to determine a number of temporary positions.

[0022] Further, the browsing frequency determination unit includes:

[0023] A browsing frequency fluctuation calculation sub-unit, which is used to calculate the standard deviation of the browsing frequency within a preset determination duration to obtain a browsing frequency fluctuation value;

[0024] A browsing frequency average calculation sub-unit, which is used to calculate the average value of the browsing frequency within the preset determination duration to obtain a browsing frequency average;

[0025] A browsing frequency determination sub-unit, which is respectively connected to the browsing frequency fluctuation calculation sub-unit and the browsing frequency average calculation sub-unit, is used to determine the temporary position as a marked position when the browsing frequency fluctuation value is less than a preset browsing frequency fluctuation threshold and the browsing frequency average is greater than a preset browsing frequency average threshold, so as to determine a number of marked positions.

[0026] Further, the determination module includes:

[0027] An index comparison unit, which is used to compare the recruitment index and the resume index to obtain an index comparison result;

[0028] A candidate determination unit, which is connected to the index comparison unit, is used to determine the concerned position as a candidate position when the index comparison result is that the resume index is greater than the recruitment index, so as to determine a number of candidate positions.

[0029] A sorting unit, which is connected to the candidate determination unit, is used to determine a candidate position list according to the historical training frequency and the historical promotion speed of each of the candidate positions.

[0030] Further, the sorting unit includes:

[0031] A training curve drawing sub-unit, which is used to draw a change curve of the historical training frequency within a preset sorting duration to obtain a number of training curves;

[0032] A promotion curve drawing subunit, configured to draw a change curve of the historical promotion speed within the preset sorting duration, and obtain a plurality of promotion curves;

[0033] A correlation calculation subunit, which is respectively connected to the training curve drawing subunit and the promotion curve drawing subunit, and is configured to calculate the correlation coefficient between the training curve and the promotion curve, and obtain a plurality of correlation coefficient values;

[0034] A sorting subunit, which is connected to the correlation calculation subunit, and is configured to sort all the correlation coefficient values of the candidate positions from high to low, and obtain a candidate position list.

[0035] Further, the adjustment module includes:

[0036] A delivery deviation calculation unit, configured to calculate the relative deviation between the delivery success rate and the preset expected delivery success rate, and obtain a delivery deviation;

[0037] An interview deviation calculation unit, configured to calculate the relative deviation between the interview success rate and the preset expected interview success rate, and obtain an interview deviation;

[0038] A deviation index calculation unit, which is respectively connected to the delivery deviation calculation unit and the interview deviation calculation unit, and is configured to perform weighted summation on the delivery deviation, the preset delivery weight, the interview deviation, and the preset interview weight, and obtain a deviation index;

[0039] A deviation comparison unit, which is connected to the deviation index calculation unit, and is configured to compare the deviation index with the preset deviation index range, and obtain a deviation comparison result;

[0040] An adjustment unit, which is connected to the deviation comparison unit, and is configured to, when the deviation comparison result is that the deviation index is greater than the maximum value of the preset deviation index range, adjust the resume index according to the deviation index, or, when the deviation comparison result is that the deviation index is less than the minimum value of the preset deviation index range, adjust the preset similarity threshold according to the delivery deviation.

[0041] Further, the adjustment unit includes:

[0042] A first adjustment subunit, configured to, when the deviation index is greater than the maximum value of the preset deviation index range, adjust the resume index according to the relative deviation between the deviation index and the maximum value of the preset deviation index range and a preset first adjustment coefficient, and obtain an adjusted resume index, and the relative deviation and the adjusted resume index are negatively correlated.

[0043] A second adjustment subunit, configured to, when the deviation index is less than or equal to the minimum value of the preset deviation index range, adjust the preset similarity threshold according to a preset delivery deviation threshold, a relative deviation of the delivery deviation, and a preset second adjustment coefficient, to obtain an adjusted similarity threshold, where the relative deviation is positively correlated with the adjusted resume index.

[0044] Further, the correction module includes:

[0045] A frequency acquisition unit, configured to acquire the number of times the adjusted resume index is obtained within the preset correction duration to form an index adjustment frequency, and acquire the number of times the adjusted similarity threshold is obtained within the preset correction duration to form a threshold adjustment frequency;

[0046] A frequency difference calculation unit, connected to the frequency acquisition unit, configured to calculate the absolute value of the difference between the index adjustment frequency and the threshold adjustment frequency to obtain a frequency difference value;

[0047] A correction unit, connected to the frequency difference calculation, configured to correct the candidate position list according to the frequency difference value, the index adjustment frequency, and the threshold adjustment frequency.

[0048] Further, the correction unit includes:

[0049] A first correction subunit, configured to, when the frequency difference value is greater than a preset frequency difference threshold and the index adjustment frequency is greater than the threshold adjustment frequency, correct the candidate position list based on the adjusted resume index to form a corrected position list, and, when the frequency difference value is greater than the preset frequency difference threshold and the index adjustment frequency is less than the threshold adjustment frequency, correct the candidate position list based on the adjusted similarity threshold to form a corrected position list;

[0050] An adjustment fluctuation calculation subunit, configured to, when the frequency difference value is less than or equal to the preset frequency difference threshold, calculate the standard deviation of the index adjustment frequency within the preset correction duration to obtain an index adjustment fluctuation value, and calculate the standard deviation of the threshold adjustment frequency within the preset correction duration to obtain a threshold adjustment fluctuation value;

[0051] A fluctuation deviation calculation subunit, connected to the adjustment fluctuation calculation subunit, configured to calculate the relative deviation between the index adjustment fluctuation value and a preset index adjustment threshold to obtain an index adjustment deviation, and calculate the relative deviation between the threshold adjustment fluctuation value and a preset threshold adjustment threshold to obtain a threshold adjustment deviation;

[0052] A second correction subunit, connected to the fluctuation deviation calculation subunit, is configured to correct the candidate position list based on the adjusted resume index to form a corrected position list when the index adjustment deviation is greater than the threshold adjustment deviation, and to correct the candidate position list based on the adjusted similarity threshold to form a corrected position list when the index adjustment deviation is less than or equal to the threshold adjustment deviation.

[0053] Compared with the prior art, the beneficial effects of the present invention are as follows. By collecting the keywords of the target uploaded resume, the relevant data of each position in the position database, and the behavior data of the target, and based on the keywords of the resume and recruitment, combined with the browsing frequency and browsing speed, the concerned positions are determined, and the positions that not only meet the position requirements but also reflect the true interests of the candidates are screened out. Then, an artificial intelligence model is used to evaluate the resume and the concerned positions to obtain a recruitment index and a resume index, providing a quantitative basis for position matching. Furthermore, the candidate position list is determined according to the recruitment index, the resume index, the historical training frequency, and the historical promotion speed, comprehensively considering the attractiveness and development prospects of the positions. The resume index or the similarity threshold is dynamically adjusted according to the delivery and interview success rates of the candidate position list, and the candidate position list is re-determined according to the number of adjustments and the recommended positions are matched, making the recommended results more in line with the target's job hunting situation and the actual matching degree of the positions, improving the accuracy and effectiveness of the matching, helping the target find a more suitable position, and effectively solving the problems of inaccurate matching and slow response speed in the prior art due to incomplete data collection and over-reliance on static models.

[0054] Furthermore, through vector transformation, the originally abstract text information can be quantified into specific numerical vectors, thus providing an operable data basis for subsequent matching calculations. The vector form can more intuitively reflect the similarity and correlation between keywords, facilitating precise matching using mathematical methods. Then, based on the relationship between the recruitment key vector and the resume key vector, the temporary positions are determined, which can more objectively and accurately screen out the positions with a higher matching degree with the resume. The concerned positions that the job seekers are really interested in are screened out through the browsing frequency and browsing speed. The browsing frequency reflects the degree of attention of the job seekers to the positions. The browsing speed further reflects the staying time of the job seekers on the position information. When the browsing speed is less than the preset browsing speed threshold, the positions that the job seekers are interested in after browsing can be screened out, indicating that the position information is more in line with the expectations of the job seekers. By first screening out the marked positions according to the browsing frequency and then further screening out the concerned positions in combination with the browsing speed, the positions that the job seekers are really interested in can be more accurately located, avoiding misjudgment that may be caused by relying only on single-dimensional data.

[0055] Furthermore, by calculating the cosine similarity between the key vectors of the resume and each recruitment key vector, a quantifiable basis is provided for job matching. Cosine similarity can effectively measure the directional correlation between two vectors and is not affected by the vector length, thus more accurately reflecting the semantic association between keywords. Then, when the similarity is greater than the threshold, the job to be matched is determined as a temporary job, which can effectively screen out the jobs with a higher matching degree with the resume, avoid the interference of low-correlation jobs, and improve the efficiency and accuracy of job screening.

[0056] Furthermore, by considering both the fluctuation value and the mean value simultaneously, the true interest of the job seeker in the job can be evaluated more comprehensively. The fluctuation value of the browsing frequency reflects the stability of the job seeker's browsing behavior for the job. A smaller fluctuation value means that the job seeker's attention to the job is relatively stable within the preset duration, rather than occasional or random browsing. The mean value of the browsing frequency, on the other hand, reflects the overall attention of the job seeker to the job. A higher mean value indicates that the job seeker has a higher interest in the job. Only when the fluctuation value is less than the preset threshold and the mean value is greater than the preset threshold, the job is determined as a marked job, which can effectively screen out the jobs that the job seeker is truly stable and highly concerned about, avoiding misjudgment caused by accidental factors.

[0057] Furthermore, by determining the candidate jobs through the recruitment index and the resume index, the matching degree between the job seeker and the job can be intuitively measured. The recruitment index reflects the requirements and expectations of the job for the job seeker, while the resume index reflects the abilities and qualifications of the job seeker. When the resume index is greater than the recruitment index, it indicates that the abilities and qualifications of the job seeker exceed the requirements of the job to a certain extent. Then, the historical training frequency and the historical promotion speed are used to sort the candidate jobs, which reflects the training and development opportunities provided by the enterprise where the job is located. A higher training frequency and promotion speed mean that the enterprise pays more attention to the growth and development of employees, and the job seeker may have greater development space and better career prospects in such a job. The finally determined list of candidate jobs not only meets the abilities and needs of the job seeker, but also provides better career development opportunities for the job seeker.

[0058] Furthermore, by plotting the change curves of the training frequency and the promotion speed and calculating their correlation coefficient, the relationship between the training and development opportunities of the job can be quantitatively evaluated. A higher correlation coefficient indicates a strong positive correlation between the training opportunities and the promotion speed of the job, that is, the more training provided by the enterprise, the faster the promotion speed of the employees, which reflects that the enterprise where the job is located provides better training and development opportunities for employees. Sorting the candidate jobs based on the correlation coefficient value can give priority to recommending those jobs that pay more attention to the growth and development of employees to the job seeker. It not only considers the current matching degree of the job, but also takes into account the career development needs of the job seeker, making the recommended jobs more attractive and practical.

[0059] Furthermore, by calculating the relative deviations of the application success rate and the interview success rate from their expected values, the gap between the actual performance and the expected goal of job seekers during the job hunting process can be quantitatively evaluated. Then, the application deviation and the interview deviation are weighted and summed after being multiplied by preset weights respectively to obtain a deviation index, fully considering the different importance of the application and interview links in the job hunting process, and balancing their impacts through weight allocation. Finally, according to the comparison result between the deviation index and the preset deviation index range, the resume index or the similarity threshold is selectively adjusted, making the entire matching process more scientific, flexible and adaptable, and effectively improving the efficiency and quality of the person-job matching.

[0060] Furthermore, when the deviation index exceeds the maximum value of the preset range, it indicates that the current matching effect is not good and the resume index overestimates the suitability of the candidate. By calculating the relative deviation of the deviation index from the maximum value and combining it with the preset first adjustment coefficient to reduce the resume index, the over-optimistic estimation in the matching process can be effectively reduced, making the matching result closer to the actual requirements. On the contrary, when the deviation index is lower than the minimum value of the preset range, it indicates that the matching result is too loose, possibly due to the similarity threshold being set too low. At this time, according to the relative deviation of the application deviation from the preset application deviation threshold and the preset second adjustment coefficient to increase the similarity threshold, the matching standard can be tightened to prevent too many unqualified candidates from entering the subsequent process. It can not only flexibly adjust the matching strategy according to the actual matching effect, but also precisely control the adjustment amplitude in a quantitative manner to ensure the stability and adaptability of the system.

[0061] Furthermore, by counting the number of times of adjusting the resume index and the similarity threshold within the preset correction duration and calculating the absolute value of the difference between the two numbers, a number difference value is obtained, which reflects the balance state between the resume index adjustment and the similarity threshold adjustment of the system. According to the number difference value and the specific number of adjustments, the candidate job list is corrected, and the matching result can be optimized dynamically according to the frequency difference of the adjustment behavior.

[0062] Furthermore, by comparing the number difference value with the preset number difference threshold, when the number difference value is greater than the preset threshold, it indicates that there is a significant imbalance in the adjustment behavior between the resume index and the similarity threshold. At this time, selecting the dimension with more adjustment times for correction can directly optimize the main problems of the system. When the number difference value is less than or equal to the preset threshold, it indicates that the adjustment behavior is relatively balanced in the two dimensions. At this time, the stability of the adjustment needs to be further considered. By calculating the standard deviation of the number of adjustments to obtain a fluctuation value, and further calculating the relative deviation of the fluctuation value from the preset threshold, the stability of the adjustment behavior can be quantified. Selecting the dimension with a smaller fluctuation deviation for correction can avoid the matching deviation caused by the instability of the adjustment behavior. Description of the Drawings

[0063] Figure 1 This is a schematic diagram of the person-job matching system based on artificial intelligence in this embodiment;

[0064] Figure 2 This is a decision logic diagram for the browsing speed determination unit in this embodiment to determine the concerned positions;

[0065] Figure 3 This is a decision logic diagram for the temporary determination subunit in this embodiment to determine the temporary positions;

[0066] Figure 4 This is a decision logic diagram for the browsing frequency determination subunit in this embodiment to determine the marked positions. Detailed implementation manners

[0067] In order to make the objectives and advantages of the present invention clearer, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0068] The preferred implementation manners of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these implementation manners are only used to explain the technical principles of the present invention and do not limit the protection scope of the present invention.

[0069] Please refer to Figure 1 as shown, which is a schematic diagram of the person-job matching system based on artificial intelligence in this embodiment;

[0070] This embodiment provides a person-job matching system based on artificial intelligence, including:

[0071] An acquisition module, configured to collect in real time the recruitment keywords, historical training frequency, historical promotion speed, browsing frequency, browsing speed, application success rate, and interview success rate of each to-be-matched position in the position database after a target uploads a resume containing resume keywords.

[0072] A determination module, connected to the acquisition module, configured to determine a number of concerned positions according to the resume keywords, the recruitment keywords, a preset similarity threshold, the browsing frequency, and the browsing speed;

[0073] An evaluation module, respectively connected to the acquisition module and the determination module, configured to evaluate a resume index based on a preset artificial intelligence model according to the resume keywords, and evaluate a recruitment index according to the recruitment keywords of each of the concerned positions;

[0074] A determination module, respectively connected to the acquisition module and the evaluation module, configured to determine a candidate position list according to the recruitment index, the resume index, the historical training frequency, and the historical promotion speed of each of the concerned positions;

[0075] An adjustment module, which is respectively connected to the acquisition module and the determination module, is used to adjust the resume index or the preset similarity threshold according to the candidate position list, the delivery success rate, and the interview success rate;

[0076] A correction module, which is connected to the adjustment module and the determination module, is used to correct the candidate position list according to the number of times of adjusting the resume index and the number of times of adjusting the preset similarity threshold within a preset correction duration, so as to form a corrected position list;

[0077] A matching module, which is connected to the correction module, is used to match a number of recommended positions according to the corrected position list.

[0078] In this embodiment, the target refers to job seekers who use recruitment software to apply for jobs, the position database refers to the set of all data stored in the recruitment software system, including but not limited to position information, enterprise information, and historical data, and each position to be matched refers to all positions in the position database that are currently open and need to recruit suitable candidates.

[0079] After the target uploads a resume containing resume keywords, the acquisition module collects multi-dimensional data in the position database in real time. Resume keywords refer to key information such as skills, experience, and educational background related to the position mentioned by the job seeker in the resume; recruitment keywords refer to key information such as skills, experience, and educational background related to the position requirements mentioned by the recruiter in the position description; both can be extracted using natural language processing technology. The historical training frequency refers to the number of training sessions provided by the enterprise to employees in the past certain period, which reflects the degree of attention of the enterprise to employee development and the investment in training resources; the historical promotion speed refers to the average time interval for employees to be promoted within the enterprise, which reflects the career development opportunities within the enterprise and the efficiency of the promotion mechanism, and both can be calculated by analyzing the human resource management system of the recruiter using SQL queries or data analysis tools. The browsing frequency of the target refers to the frequency at which the job seeker browses different positions on the recruitment software, and the browsing speed refers to the average time for the job seeker to browse each position page on the recruitment software, and both can be collected through the user behavior logs of the recruitment software. The delivery success rate refers to the proportion of job seekers who obtain interview opportunities after submitting their resumes, and the interview success rate refers to the proportion of job seekers who obtain employment after participating in interviews, and both can be calculated through the position database using SQL queries or data warehouse tools.

[0080] The resume index refers to a quantitative indicator obtained by comprehensively evaluating various information (such as skills, experience, educational background, etc.) in a job seeker's resume through a pre-set artificial intelligence model, which reflects the overall ability and background of the job seeker. The recruitment index refers to a quantitative indicator obtained by comprehensively evaluating various requirements (such as skills, experience, educational qualifications, etc.) in a job description through a pre-set artificial intelligence model, which reflects the requirements for abilities of the position.

[0081] The pre-set artificial intelligence model is a model constructed based on machine learning-based classification or regression, and is used to evaluate the resume index and recruitment index according to the input feature data (resume keywords, recruitment keywords).

[0082] 1. Model Overview

[0083] The pre-set artificial intelligence model is based on the BERT architecture, capable of processing text data and extracting semantic information, thus achieving precise matching. The model combines the deep learning ability of BERT and multi-task learning, and processes the text features of resumes and job descriptions respectively, and can effectively capture the semantic and context information in the text.

[0084] 2. Initial Model Architecture

[0085] 2.1 Resume Feature Extraction Module

[0086] The resume feature extraction module is based on the BERT architecture and is used to process the semantic features of resume texts.

[0087] Input layer: Receives the pre-processed resume text, and its dimension is (\text{batch_size}, \text{sequence_length}).

[0088] BERT layer:

[0089] Multi-head self-attention mechanism: Used to capture long-range dependencies in the text.

[0090] Feed-forward neural network: Used to further process the features.

[0091] Residual connection and layer normalization: Used to stabilize the training process.

[0092] Fully connected layer: Maps the output of the BERT layer to a low-dimensional resume feature vector.

[0093] Output layer: Outputs the resume feature vector for subsequent analysis.

[0094] 2.2 Job Feature Extraction Module

[0095] The job feature extraction module is also based on the BERT architecture and is used to process the semantic features of job description texts.

[0096] Input layer: Receives the preprocessed job description text, with dimensions (\text{batch_size}, \text{sequence_length}).

[0097] BERT layer:

[0098] Multi-head self-attention mechanism: Used to capture long-range dependencies in the text.

[0099] Feed-forward neural network: Used to further process features.

[0100] Residual connection and layer normalization: Used to stabilize the training process.

[0101] Fully connected layer: Maps the output of the BERT layer to a low-dimensional job feature vector.

[0102] Output layer: Outputs the job feature vector for subsequent analysis.

[0103] 2.3 Matching evaluation module

[0104] The matching evaluation module is used to evaluate the matching degree between the resume feature vector and the job feature vector, and generate a resume index and a recruitment index.

[0105] Input layer: Receives the resume feature vector and the job feature vector.

[0106] Similarity calculation layer: Calculates the similarity between the resume feature vector and the job feature vector, usually using cosine similarity or dot product.

[0107] Fully connected layer: Maps the similarity to a resume index and a recruitment index.

[0108] Output layer: Outputs the resume index and the recruitment index.

[0109] 3. Core parameters

[0110] 3.1 BERT parameters

[0111] Input dimension:

[0112] batch_size: The number of data samples in each training batch.

[0113] sequence_length: The length of the text sequence.

[0114] BERT layer parameters:

[0115] Multi-head self-attention mechanism:

[0116] num_heads: The number of attention heads, used to process features in different subspaces in parallel.

[0117] hidden_dim: The hidden dimension of each attention head.

[0118] Feed-forward neural network:

[0119] feed_forward_dim: The intermediate dimension of the feed-forward network.

[0120] Activation function: Usually ReLU or GELU is used.

[0121] Residual connection and layer normalization: Used to stabilize the training process.

[0122] Output dimension: The dimension of the low-dimensional feature vector (such as 128 or 256).

[0123] 3.2 Matching evaluation parameters

[0124] Similarity calculation: Cosine similarity or dot product.

[0125] Fully connected layer:

[0126] Output dimension: The dimension of the resume index and the recruitment index (such as 128 or 256).

[0127] 4. Training process

[0128] 4.1 Input data preparation

[0129] Resume text: The resume text is preprocessed, including word segmentation, encoding, etc., to form an input dataset with the dimension (\text{batch_size}, \text{sequence_length}).

[0130] Job description text: The job description text is preprocessed, including word segmentation, encoding, etc., to form an input dataset with the dimension (\text{batch_size}, \text{sequence_length}).

[0131] 4.2 Model training

[0132] Loss function:

[0133] Resume feature extraction: Use the mean squared error (MSE) loss.

[0134] Job feature extraction: Use the mean squared error (MSE) loss.

[0135] Matching evaluation: Use the mean squared error (MSE) loss or cross-entropy loss.

[0136] Optimizer: Use the Adam optimizer with the learning rate set to 2×10-5.

[0137] Training process: Update the model parameters through backpropagation. The training objective is to enable the model to accurately extract the features of resumes and job descriptions and evaluate their matching degree.

[0138] 4.3 Preset training duration

[0139] Data scale: During training, the model needs to learn a sufficient amount of historical data to identify patterns, and the scale of the dataset directly affects the training duration.

[0140] Convergence: As the loss function decreases, the model gradually converges, and the training duration is also affected by the convergence speed of the model.

[0141] 5. Trained model

[0142] Model parameters:

[0143] Weights and biases of the BERT layer.

[0144] Weights and biases of the fully connected layer.

[0145] Output features:

[0146] Resume feature vector.

[0147] Job feature vector.

[0148] Resume index and recruitment index.

[0149] 6. Using the trained model

[0150] Input new data: Input the new resume text and job description text into the trained model.

[0151] Model inference: The model extracts features and outputs the resume index and recruitment index for subsequent matching analysis.

[0152] The preset correction duration is a time window for statistical and analytical adjustment behaviors, which depends on the system's operating frequency, data update speed, and the dynamics of business requirements. It is usually set between 1 week and 1 month. In this embodiment, it is set to 2 weeks, which can accumulate sufficient adjustment data to provide a reliable statistical basis for the correction module and can also adapt to the common periodic changes in the recruitment market.

[0153] By collecting various key data in real time, several concerned positions are screened out according to the resume keywords, recruitment keywords, preset similarity threshold, browsing frequency, and browsing speed collected. Then, based on a preset artificial intelligence model, the resume index is evaluated using the resume keywords, and the recruitment index is evaluated according to the recruitment keywords of the concerned positions. Combining the recruitment index, resume index, historical training frequency, and historical promotion speed of the concerned positions, a list of candidate positions is determined. According to the list of candidate positions, delivery success rate, and interview success rate, the resume index or preset similarity threshold is adjusted. According to the number of adjustments within a preset correction duration, the list of candidate positions is corrected. According to the corrected list of candidate positions, several recommended positions are matched.

[0154] By collecting the keywords of the target's uploaded resume, the relevant data of each position in the position database, and the target's behavior data, and based on the resume and recruitment keywords, combined with the browsing frequency and browsing speed, the concerned positions are determined, and the positions that not only meet the position requirements but also reflect the candidate's true interests are screened out. Then, an artificial intelligence model is used to evaluate the resume and the concerned positions to obtain the recruitment index and the resume index, providing a quantitative basis for position matching. Then, according to the recruitment index, resume index, historical training frequency, and historical promotion speed, a list of candidate positions is determined, comprehensively considering the attractiveness and development prospects of the positions. According to the delivery and interview success rates of the list of candidate positions, the resume index or similarity threshold is dynamically adjusted, and the list of candidate positions is re-determined according to the number of adjustments and recommended positions are matched, making the recommendation results more in line with the target's job search situation and the actual matching degree of the positions, improving the accuracy and effectiveness of the matching, helping the target find a more suitable position, and effectively solving the problems of inaccurate matching and slow response speed in the prior art due to incomplete data collection and over-reliance on static models.

[0155] Please continue to refer to Figure 2 as shown, which is the decision logic diagram for the browsing speed determination unit of this embodiment to determine the concerned positions;

[0156] The determination module includes:

[0157] A vector conversion unit for converting the recruitment keywords into vector form to obtain several recruitment key vectors, and converting the resume keywords into vector form to obtain resume key vectors;

[0158] A temporary determination unit connected to the vector conversion unit for determining several temporary positions according to the resume key vectors and each of the recruitment key vectors;

[0159] A browsing frequency determination unit connected to the temporary determination unit for determining several marked positions according to the browsing frequency of each of the temporary positions;

[0160] A browsing speed determination unit, which is connected to the browsing frequency determination unit, is used to determine a marked position as a concerned position when the browsing speed of each of the marked positions is less than a preset browsing speed threshold, so as to determine a number of concerned positions.

[0161] The preset browsing speed threshold is a standard value for measuring browsing speed, which depends on the user behavior data of the recruitment software, the complexity of the position, the industry characteristics, and the average browsing habits of job seekers. It is usually set between 100 words per second and 300 words per second. In this embodiment, it is set to 200 words per second, which can effectively screen out the positions that job seekers are really interested in, exclude those positions that are only browsed quickly but not read in depth, so as to improve the quality and accuracy of the concerned positions.

[0162] By converting the recruitment keywords and resume keywords into vector forms respectively, a recruitment key vector and a resume key vector are obtained. Then, a number of temporary positions are determined according to these vectors. After that, according to the browsing frequencies of these temporary positions, a number of marked positions are further screened out. Finally, for these marked positions, if their browsing speed is less than the preset browsing speed threshold, they are determined as concerned positions, so as to determine a number of concerned positions.

[0163] Through vector transformation, the originally abstract text information can be quantified into specific numerical vectors, providing an operable data basis for subsequent matching calculations. The vector form can more intuitively reflect the similarity and correlation between keywords, facilitating precise matching using mathematical methods. Then, determining temporary positions based on the relationship between the recruitment key vector and the resume key vector can more objectively and accurately screen out positions with a higher matching degree with the resume. Screening out the concerned positions that job seekers are really interested in through browsing frequency and browsing speed, the browsing frequency reflects the degree of attention of job seekers to the position. And the browsing speed further reflects the staying time of job seekers on the position information. A browsing speed less than the preset browsing speed threshold can screen out the positions that job seekers are interested in after browsing, indicating that the position information is more in line with the expectations of job seekers. By first screening out marked positions according to browsing frequency and then further screening out concerned positions in combination with browsing speed, the positions that job seekers are really interested in can be more accurately located, avoiding misjudgments that may be caused by relying on single-dimensional data alone.

[0164] Please continue to refer to Figure 3 as shown, which is the determination logic diagram of the temporary determination subunit in this embodiment for determining temporary positions;

[0165] The temporary determination unit includes:

[0166] A similarity calculation subunit, which is used to calculate the cosine similarity between the resume key vector and each of the recruitment key vectors to obtain a number of similarities;

[0167] A temporary determination subunit, which is connected to the similarity calculation subunit, is used to determine that the to-be-matched position is a temporary position when the similarity is greater than a preset similarity threshold, so as to determine a number of temporary positions.

[0168] The preset similarity threshold is a benchmark value for judging the similarity between the resume key vector and the recruitment key vector, which depends on the precise matching requirements of the recruitment position, the differences in resume quality, and the diversity of recruitment positions. It is usually set between 0.5 and 0.8. In this embodiment, it is set to 0.7, which can effectively screen out positions highly relevant to the resume, reduce misjudgment, and improve the quality and accuracy of temporary positions.

[0169] By calculating the cosine similarity between the resume key vector and each recruitment key vector, a number of similarity values are obtained. Then, these similarity values are compared with the preset similarity threshold. When the similarity of a to-be-matched position is greater than the preset similarity threshold, it is determined that the position is a temporary position, thereby determining a number of temporary positions.

[0170] By calculating the cosine similarity between the resume key vector and each recruitment key vector, it provides a quantifiable basis for position matching. The cosine similarity can effectively measure the directional correlation between two vectors and is not affected by the vector length, thus more accurately reflecting the semantic association between keywords. Then, when the similarity is greater than the threshold, it is determined that the to-be-matched position is a temporary position, which can effectively screen out positions with a higher matching degree to the resume, avoid the interference of low-correlation positions, and improve the efficiency and accuracy of position screening.

[0171] Please continue to refer to Figure 4 as shown, which is the determination logic diagram of the marked positions determined by the browsing frequency determination subunit in this embodiment;

[0172] The browsing frequency determination unit includes:

[0173] A browsing frequency fluctuation calculation subunit, which is used to calculate the standard deviation of the browsing frequency within a preset determination duration to obtain a browsing frequency fluctuation value;

[0174] A browsing frequency mean calculation subunit, which is used to calculate the average value of the browsing frequency within the preset determination duration to obtain a browsing frequency mean;

[0175] A browsing frequency determination subunit, which is respectively connected to the browsing frequency fluctuation calculation subunit and the browsing frequency mean calculation subunit, and is used to determine that the temporary position is a marked position when the browsing frequency fluctuation value is less than a preset browsing frequency fluctuation threshold and the browsing frequency mean is greater than a preset browsing frequency mean threshold, so as to determine a number of marked positions.

[0176] The preset determination duration refers to the time interval used to calculate the browsing frequency fluctuation value and the average browsing frequency, which depends on the periodicity of the user behavior data on the recruitment platform and the change rule of the job popularity. It is usually set between 12 hours and 2 days. In this embodiment, it is set to 1 day to better balance the timeliness and stability of the data, which can not only reflect the attention trend of job seekers to the job within a relatively short time, but also avoid affecting the judgment accuracy due to excessive data fluctuation caused by too short time.

[0177] The preset browsing frequency fluctuation threshold is the standard deviation threshold used to measure the stability of the browsing frequency, which depends on the popularity of the job and the stability of the job seekers' attention. It is usually set between 0.1 and 0.3. In this embodiment, it is set to 0.2, which can effectively screen out the jobs with relatively stable browsing frequencies, avoid misjudgment caused by occasional high-frequency or low-frequency browsing of job seekers, and thus improve the reliability of the marked jobs.

[0178] The preset average browsing frequency threshold is the average value threshold used to measure the overall attention of the job, which depends on the attractiveness of the job and the market demand. It is usually set between 5 times / day and 10 times / day. In this embodiment, it is set to 8 times / day to ensure that the screened jobs have a high degree of attention, and at the same time avoid missing some potential jobs due to too high a threshold, and improve the quality and quantity of the marked jobs.

[0179] By calculating the standard deviation of the browsing frequency, the browsing frequency fluctuation value is obtained; at the same time, the average value of the browsing frequency is calculated to obtain the average browsing frequency; when the browsing frequency fluctuation value is less than the preset fluctuation threshold and the average browsing frequency is greater than the preset average threshold, it is determined that the temporary job is a marked job.

[0180] By considering both the fluctuation value and the average value, the true interest of job seekers in the job can be evaluated more comprehensively. The browsing frequency fluctuation value reflects the stability of the job seekers' browsing behavior for the job. A smaller fluctuation value means that the job seekers' attention to the job is relatively stable within the preset duration, rather than occasional or random browsing. The average browsing frequency, on the other hand, reflects the overall attention of job seekers to the job. A higher average value indicates that job seekers have a higher interest in the job. Only when the fluctuation value is less than the preset threshold and the average value is greater than the preset threshold, is the job determined to be a marked job, which can effectively screen out the jobs that job seekers are truly stable and highly concerned about, and avoid misjudgment caused by accidental factors.

[0181] Specifically, the determination module includes:

[0182] An index comparison unit for comparing the recruitment index and the resume index to obtain an index comparison result;

[0183] A candidate determination unit, which is connected to the index comparison unit, is used to determine that the concerned position is a candidate position when the index comparison result is that the resume index is greater than the recruitment index, so as to determine a number of candidate positions.

[0184] A sorting unit, which is connected to the candidate determination unit, is used to determine a candidate position list according to the historical training frequency and the historical promotion speed of each candidate position.

[0185] By comparing the recruitment index and the resume index, when the resume index is greater than the recruitment index, a number of candidate positions are determined. Based on the historical training frequency and the historical promotion speed of these candidate positions, a candidate position list is further determined.

[0186] Determining candidate positions through the recruitment index and the resume index can intuitively measure the matching degree between job seekers and positions. The recruitment index reflects the requirements and expectations of the position for job seekers, while the resume index reflects the abilities and qualifications of job seekers. When the resume index is greater than the recruitment index, it means that the abilities and qualifications of job seekers exceed the requirements of the position to a certain extent. Then, the historical training frequency and the historical promotion speed are used to sort the candidate positions, which reflects the cultivation and development opportunities provided by the enterprise where the position is located. A higher training frequency and promotion speed mean that the enterprise pays more attention to the growth and development of employees, and job seekers may have greater development space and better career prospects in such positions. The finally determined candidate position list not only meets the abilities and needs of job seekers, but also provides better career development opportunities for job seekers.

[0187] Specifically, the sorting unit includes:

[0188] A training curve drawing subunit, which is used to draw the change curve of the historical training frequency within a preset sorting time period to obtain a number of training curves;

[0189] A promotion curve drawing subunit, which is used to draw the change curve of the historical promotion speed within the preset sorting time period to obtain a number of promotion curves;

[0190] A correlation calculation subunit, which is respectively connected to the training curve drawing subunit and the promotion curve drawing subunit, and is used to calculate the correlation coefficient between the training curve and the promotion curve to obtain a number of correlation coefficient values;

[0191] A sorting subunit, which is connected to the correlation calculation subunit, is used to sort the correlation coefficient values of all the candidate positions from high to low to obtain a candidate position list.

[0192] The preset sorting duration is the time range used to statistically analyze the historical training frequency and the change in the historical promotion speed. It depends on factors such as the career development stage of job seekers, the industry characteristics of the positions, and the talent cultivation cycle of the enterprise, and is usually set between 1 year and 5 years. In this embodiment, it is set to 3 years, which conforms to the cycle law of employee cultivation and promotion in most enterprises, making the sorting result more valuable for reference.

[0193] By plotting the change curves of the historical training frequency and the historical promotion speed of each position within the preset sorting duration, several training curves and promotion curves are obtained; then, the correlation coefficients between the training curves and the promotion curves are calculated to obtain several correlation coefficient values; all candidate positions are sorted according to the correlation coefficient values from high to low to obtain a list of candidate positions.

[0194] By plotting the change curves of the training frequency and the promotion speed and calculating their correlation coefficients, the relationship between the training and development opportunities of the positions can be quantitatively evaluated. A higher correlation coefficient indicates a strong positive correlation between the training opportunities of the position and the promotion speed, that is, the more training provided by the enterprise, the faster the promotion speed of employees, which reflects that the enterprise where the position is located has better cultivation and development opportunities for employees. Sorting the candidate positions based on the correlation coefficient values can preferentially recommend those positions that pay more attention to the growth and development of employees to job seekers. It not only considers the current matching degree of the positions but also takes into account the career development needs of job seekers, making the recommended positions more attractive and practical.

[0195] Specifically, the adjustment module includes:

[0196] A delivery deviation calculation unit for calculating the relative deviation between the delivery success rate and the preset expected delivery success rate to obtain a delivery deviation;

[0197] An interview deviation calculation unit for calculating the relative deviation between the interview success rate and the preset expected interview success rate to obtain an interview deviation;

[0198] A deviation index calculation unit, which is respectively connected to the delivery deviation calculation unit and the interview deviation calculation unit, for performing a weighted sum of the delivery deviation, the preset delivery weight, the interview deviation, and the preset interview weight to obtain a deviation index;

[0199] A deviation comparison unit, which is connected to the deviation index calculation unit, for comparing the deviation index with the preset deviation index range to obtain a deviation comparison result;

[0200] An adjustment unit, connected to the deviation comparison unit, is configured to adjust the resume index according to the deviation index when the deviation comparison result indicates that the deviation index is greater than the maximum value of the preset deviation index range, or to adjust the preset similarity threshold according to the delivery deviation when the deviation comparison result indicates that the deviation index is less than the minimum value of the preset deviation index range.

[0201] The preset expected delivery success rate refers to the ideal proportion of obtaining interview opportunities after submitting a resume, set according to industry average levels or historical data, and depends on industry competition, the popularity of job requirements, and the overall quality of job seekers. It is usually set between 30% and 70%. In this embodiment, it is set to 40%, which can better reflect the average situation of most positions, neither too high nor too low, and helps to reasonably evaluate the delivery effect of job seekers.

[0202] The preset expected interview success rate refers to the passing rate of interviews set according to industry average levels or historical data, and depends on the intensity of job competition, corporate recruitment standards, and the quality of job seekers. It is usually set between 20% and 40%. In this embodiment, it is set to 30%, which can better balance the expectations of job seekers and enterprises, neither too lenient nor too strict, and helps to reasonably evaluate the interview performance of job seekers.

[0203] The preset delivery weight refers to the weight ratio of the delivery deviation when calculating the deviation index, and depends on the importance of the delivery success rate to the overall job search effect of job seekers. It is usually set between 0.3 and 0.7. In this embodiment, it is set to 0.4, which can reasonably reflect the importance of the delivery success rate in the job search process, neither too high nor too low.

[0204] The preset interview weight refers to the weight ratio of the interview deviation when calculating the deviation index, and depends on the importance of the interview success rate to the final job search result of job seekers. It is usually set between 0.3 and 0.7. In this embodiment, it is set to 0.6, which can reasonably reflect the key role of the interview success rate in the job search process, higher than the delivery weight, reflecting the decisive role of the interview in the job search process, and helps to more accurately adjust the resume index and improve the matching degree of job seekers.

[0205] The preset deviation index range is a numerical range used to evaluate whether the deviation index is within a reasonable interval, depending on the system's business objectives, statistical analysis of historical data, and expectations for the matching effect. It is usually set between 0 and 1. In this embodiment, it is set between 0.3 and 0.7, which can effectively identify situations with large deviations, while avoiding frequent misjudgments caused by overly strict thresholds, thus balancing the stability and flexibility of the system, ensuring timely adjustment when the deviation is large, and maintaining the stability of the system when the deviation is small.

[0206] By calculating the relative deviation between the delivery success rate and the preset expected delivery success rate, the delivery deviation is obtained. At the same time, the relative deviation between the interview success rate and the preset expected interview success rate is calculated to obtain the interview deviation. Then, the delivery deviation and the interview deviation are respectively multiplied by the preset delivery weight and interview weight and then weighted and summed to obtain the deviation index. Next, the deviation index is compared with the preset deviation index range. If the deviation index is greater than the maximum value of the preset deviation index range, it indicates that the delivery and interview effects are significantly lower than expected. At this time, the resume index is adjusted according to the deviation index. If the deviation index is less than the minimum value of the preset deviation index range, it indicates that the delivery and interview effects are significantly higher than expected. At this time, the preset similarity threshold is adjusted according to the delivery deviation.

[0207] When the deviation index is higher than the maximum value of the preset range, it indicates that the current matching effect has a large gap with the expectation, which may be caused by inaccurate resume evaluation. By adjusting the resume index, the matching degree of the candidate can be re-evaluated, thereby improving the accuracy of the matching. When the deviation index is lower than the minimum value of the preset range, it indicates that the current matching effect far exceeds the expectation, which may be caused by too low similarity threshold. By adjusting the similarity threshold, the strictness of the matching can be improved to avoid over-matching.

[0208] By calculating the relative deviations of the delivery success rate and the interview success rate from the expected values, the gap between the actual performance of job seekers in the job hunting process and the expected goals can be quantitatively evaluated. Then, the delivery deviation and the interview deviation are respectively multiplied by the preset weights and then weighted and summed to obtain the deviation index, fully considering the different importance of the delivery and interview links in the job hunting process and balancing their impacts through weight allocation. Finally, according to the comparison result of the deviation index with the preset deviation index range, the resume index or the similarity threshold is selectively adjusted, making the entire matching process more scientific, flexible and adaptable, and effectively improving the efficiency and quality of the person-job matching.

[0209] Specifically, the adjustment unit includes:

[0210] A first adjustment subunit, configured to, when the deviation index is greater than the maximum value of the preset deviation index range, adjust the resume index according to the relative deviation between the deviation index and the maximum value of the preset deviation index range and a preset first adjustment coefficient to obtain an adjusted resume index, and the relative deviation and the adjusted resume index are negatively correlated.

[0211] A second adjustment subunit, configured to, when the deviation index is less than or equal to the minimum value of the preset deviation index range, adjust the preset similarity threshold according to the relative deviation between the preset delivery deviation threshold and the delivery deviation and a preset second adjustment coefficient to obtain an adjusted similarity threshold, and the relative deviation and the adjusted resume index are positively correlated.

[0212] The preset first adjustment coefficient is used to adjust the resume index. It depends on the applicant's performance stability during the application and interview process, as well as the company's expectations of the applicant's abilities. It is usually set between 0.5 and 0.9. In this embodiment, it is set to 0.7, which makes the resume index more sensitive to changes in the applicant's performance during the application and interview process, thereby more accurately matching them to positions.

[0213] The preset delivery deviation threshold measures the maximum acceptable deviation between the delivery success rate and the expected delivery success rate. It is determined by the analysis of historical data, the nature of the recruitment position, and the stringency of the recruitment goals, and is typically set between 5% and 15%. In this embodiment, it is set to 10%, which effectively balances the strictness and flexibility of the matching process, avoiding inaccurate matching results due to excessive deviation, while also avoiding excessively strict matching results due to too small a deviation.

[0214] The preset second adjustment coefficient is used to adjust the preset similarity threshold. It depends on the system's sensitivity to delivery bias and the similarity threshold adjustment strategy, and is typically set between 0.1 and 0.5. In this embodiment, it is set to 0.3 to ensure that the adjustment of the similarity threshold to delivery bias is appropriate, effectively reflecting the impact of delivery bias without over-adjusting the similarity threshold, thereby maintaining system stability and accuracy.

[0215] When the deviation index exceeds the maximum value of the preset deviation index range, the resume index is adjusted according to the relative deviation between the deviation index and the maximum value, and the preset first adjustment coefficient; when the deviation index is less than or equal to the minimum value of the preset deviation index range, the preset similarity threshold is adjusted according to the relative deviation between the preset delivery deviation threshold and the actual delivery deviation, and the preset second adjustment coefficient.

[0216] When the deviation index exceeds the maximum value of the preset range, it means that the current matching effect is poor and the resume index overestimates the suitability of the candidate. By calculating the relative deviation of the deviation index from the maximum value and combining it with the preset first adjustment coefficient to reduce the resume index, the over-optimistic estimation in the matching process is effectively reduced, making the matching results closer to actual needs. On the contrary, when the deviation index is lower than the minimum value of the preset range, it indicates that the matching result is too loose, and the similarity threshold may be set too low. At this time, increasing the similarity threshold based on the relative deviation between the delivery deviation and the preset delivery deviation threshold and the preset second adjustment coefficient can tighten the matching criteria and avoid too many unsuitable candidates from entering the subsequent process. It can not only flexibly adjust the matching strategy according to the actual matching effect, but also accurately control the adjustment range in a quantitative manner to ensure the stability and adaptability of the system.

[0217] Specifically, the correction module includes:

[0218] A frequency acquisition unit is configured to acquire the number of times the adjusted resume index is obtained within the preset correction duration to form an index adjustment frequency, and acquire the number of times the adjusted similarity threshold is obtained within the preset correction duration to form a threshold adjustment frequency;

[0219] A frequency difference calculation unit, which is connected to the frequency acquisition unit, is configured to calculate the absolute value of the difference between the index adjustment frequency and the threshold adjustment frequency to obtain a frequency difference value;

[0220] A correction unit, which is connected to the frequency difference calculation, is configured to correct the candidate job list according to the frequency difference value, the index adjustment frequency, and the threshold adjustment frequency.

[0221] By acquiring the number of times of the adjusted resume index to form an index adjustment frequency, and at the same time acquiring the number of times of the adjusted similarity threshold to form a threshold adjustment frequency. Subsequently, calculate the absolute value of the difference between the index adjustment frequency and the threshold adjustment frequency to obtain a frequency difference value. Finally, correct the candidate job list according to the frequency difference value and the specific situations of the index adjustment frequency and the threshold adjustment frequency.

[0222] By counting the number of times of the adjusted resume index and the adjusted similarity threshold within the preset correction duration, calculating the absolute value of the difference between the two frequencies to obtain a frequency difference value, which reflects the balance state between the resume index adjustment and the similarity threshold adjustment of the system. Correcting the candidate job list according to the frequency difference value and the specific adjustment frequencies can dynamically optimize the matching result according to the frequency difference of the adjustment behaviors.

[0223] Specifically, the correction unit includes:

[0224] A first correction subunit is configured to, when the frequency difference value is greater than a preset frequency difference threshold and the index adjustment frequency is greater than the threshold adjustment frequency, correct the candidate job list based on the adjusted resume index to form a corrected job list, and, when the frequency difference value is greater than the preset frequency difference threshold and the index adjustment frequency is less than the threshold adjustment frequency, correct the candidate job list based on the adjusted similarity threshold to form a corrected job list;

[0225] An adjustment fluctuation calculation subunit is configured to, when the frequency difference value is less than or equal to the preset frequency difference threshold, calculate the standard deviation of the index adjustment frequency within the preset correction duration to obtain an index adjustment fluctuation value, and calculate the standard deviation of the threshold adjustment frequency within the preset correction duration to obtain a threshold adjustment fluctuation value;

[0226] A fluctuation deviation calculation sub-unit, which is connected to the adjustment fluctuation calculation sub-unit, is used to calculate the relative deviation between the index adjustment fluctuation value and the preset index adjustment threshold to obtain the index adjustment deviation, and calculate the relative deviation between the threshold adjustment fluctuation value and the preset threshold adjustment threshold to obtain the threshold adjustment deviation;

[0227] A second correction sub-unit, which is connected to the fluctuation deviation calculation sub-unit, is used to, when the index adjustment deviation is greater than the threshold adjustment deviation, correct the candidate position list based on the adjusted resume index to form a corrected position list, and, when the index adjustment deviation is less than or equal to the threshold adjustment deviation, correct the candidate position list based on the adjusted similarity threshold to form a corrected position list.

[0228] The preset number difference threshold is a reference value for judging whether the difference between the index adjustment times and the threshold adjustment times is significant, depending on the adjustment frequency of the system, business requirements, and the fluctuation of historical data, and is usually set between 0 and 10. In this embodiment, it is set to 5, which can avoid unnecessary corrections caused by minor differences while ensuring the adjustment flexibility of the system, and improve the stability and adaptability of the system.

[0229] The preset index adjustment threshold is a reference value for judging whether the index adjustment fluctuation value is within a reasonable range, depending on the historical fluctuation of the resume index adjustment and the business requirements for matching accuracy, and is usually set between 0.1 and 0.5. In this embodiment, it is set to 0.3, which can effectively filter out misjudgments caused by accidental fluctuations, ensure the stability of index adjustment, and take into account the matching accuracy at the same time.

[0230] The preset threshold adjustment threshold is a reference value for judging whether the similarity threshold adjustment fluctuation value is within a reasonable range, depending on the historical fluctuation of the similarity threshold adjustment and the business requirements for matching strictness, and is usually set between 0.1 and 0.5. In this embodiment, it is set to 0.3, which can effectively avoid the instability of matching results caused by frequent adjustment of the similarity threshold, and ensure the rationality of the matching standard at the same time.

[0231] When the number difference value is greater than the preset number difference threshold, if the index adjustment times are greater than the threshold adjustment times, the candidate position list will be corrected based on the adjusted resume index; otherwise, the candidate position list will be corrected based on the adjusted similarity threshold. If the number difference value is less than or equal to the preset number difference threshold, the standard deviations of the index adjustment times and the threshold adjustment times within the preset correction duration are further calculated to obtain the index adjustment fluctuation value and the threshold adjustment fluctuation value. Then, the relative deviation of the index adjustment fluctuation value from the preset index adjustment threshold is calculated to obtain the index adjustment deviation, and the relative deviation of the threshold adjustment fluctuation value from the preset threshold adjustment threshold is calculated to obtain the threshold adjustment deviation. Finally, when the index adjustment deviation is greater than the threshold adjustment deviation, the candidate position list is corrected based on the adjusted resume index to form a corrected position list; otherwise, the candidate position list is corrected based on the adjusted similarity threshold to form a corrected position list. After selecting the adjustment object, the adjustment object is adjusted according to the adjustment method of the adjustment module, and the candidate position list is regenerated to obtain the corrected position list.

[0232] By comparing the number difference value and the preset number difference threshold, when the number difference value is greater than the preset threshold, it indicates that there is a significant imbalance between the adjustment behaviors in the resume index and the similarity threshold. At this time, selecting the dimension with more adjustment times for correction can directly optimize the main problems of the system. When the number difference value is less than or equal to the preset threshold, it indicates that the adjustment behaviors in the two dimensions are relatively balanced. At this time, it is necessary to further consider the stability of the adjustment. By calculating the standard deviation of the adjustment times to obtain the fluctuation value, and further calculating the relative deviation of the fluctuation value from the preset threshold, the stability of the adjustment behavior can be quantified. Selecting the dimension with a smaller fluctuation deviation for correction can avoid the matching deviation caused by the instability of the adjustment behavior.

[0233] So far, the technical solution of the present invention has been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. An artificial intelligence-based job-person matching system, characterized in that, Including: A collection module for collecting, in real time after a target uploads a resume containing resume keywords, the recruitment keywords, historical training frequencies, historical promotion speeds, the target's browsing frequencies, browsing speeds, application success rates, and interview success rates of each job to be matched in the job database. A determination module connected to the collection module for determining a number of concerned jobs according to the resume keywords, the recruitment keywords, a preset similarity threshold, the browsing frequency, and the browsing speed. An evaluation module connected to the collection module and the determination module respectively for evaluating a resume index based on a preset artificial intelligence model according to the resume keywords, and evaluating a recruitment index according to the recruitment keywords of each of the concerned jobs. A determination module connected to the collection module and the evaluation module respectively for determining a list of candidate jobs according to the recruitment indices, the resume index, the historical training frequency, and the historical promotion speed of each of the concerned jobs. An adjustment module connected to the collection module and the determination module respectively for adjusting the resume index or adjusting the preset similarity threshold according to the list of candidate jobs, the application success rate, and the interview success rate. A correction module connected to the adjustment module and the determination module for correcting the list of candidate jobs according to the number of times of adjusting the resume index and the number of times of adjusting the preset similarity threshold within a preset correction duration to form a corrected job list. A matching module connected to the correction module for matching a number of recommended jobs according to the corrected job list.

2. The artificial intelligence-based job-person matching system according to claim 1, wherein The determination module includes: A vector conversion unit for converting the recruitment keywords into a vector form to obtain a number of recruitment key vectors, and converting the resume keywords into a vector form to obtain a resume key vector. A temporary determination unit connected to the vector conversion unit for determining a number of temporary jobs according to the resume key vector and each of the recruitment key vectors. A browsing frequency determination unit connected to the temporary determination unit for determining a number of marked jobs according to the browsing frequency of each of the temporary jobs. A browsing speed determination unit connected to the browsing frequency determination unit for determining the marked jobs as concerned jobs when the browsing speed of each of the marked jobs is less than a preset browsing speed threshold to determine a number of concerned jobs.

3. The artificial intelligence-based job-person matching system according to claim 2, wherein The temporary determination unit includes: A similarity calculation sub-unit for calculating the cosine similarity between the resume key vector and each of the recruitment key vectors to obtain a number of similarities. A temporary determination sub-unit connected to the similarity calculation sub-unit for determining the job to be matched as a temporary job when the similarity is greater than a preset similarity threshold to determine a number of temporary jobs.

4. The artificial intelligence-based job-person matching system according to claim 3, wherein The browsing frequency determination unit includes: A browsing frequency fluctuation calculation sub-unit for calculating the standard deviation of the browsing frequency within a preset determination duration to obtain a browsing frequency fluctuation value. A browsing frequency average calculation sub-unit for calculating the average value of the browsing frequency within the preset determination duration to obtain a browsing frequency average. A browsing frequency determination sub-unit, which is respectively connected to the browsing frequency fluctuation calculation sub-unit and the browsing frequency average value calculation sub-unit, and is used to determine the temporary position as a marked position when the browsing frequency fluctuation value is less than a preset browsing frequency fluctuation threshold and the browsing frequency average value is greater than a preset browsing frequency average value threshold, so as to determine a number of marked positions.

5. The artificial intelligence-based job-person matching system according to claim 4, wherein The determination module includes: An index comparison unit, which is used to compare the recruitment index and the resume index to obtain an index comparison result; A candidate determination unit, which is connected to the index comparison unit, and is used to determine the concerned position as a candidate position when the index comparison result is that the resume index is greater than the recruitment index, so as to determine a number of candidate positions. A sorting unit, which is connected to the candidate determination unit, and is used to determine a candidate position list according to the historical training frequency and the historical promotion speed of each candidate position.

6. The artificial intelligence-based job-person matching system according to claim 5, wherein The sorting unit includes: A training curve drawing sub-unit, which is used to draw a change curve of the historical training frequency within a preset sorting duration to obtain a number of training curves; A promotion curve drawing sub-unit, which is used to draw a change curve of the historical promotion speed within the preset sorting duration to obtain a number of promotion curves; A correlation calculation sub-unit, which is respectively connected to the training curve drawing sub-unit and the promotion curve drawing sub-unit, and is used to calculate the correlation coefficient between the training curve and the promotion curve to obtain a number of correlation coefficient values; A sorting sub-unit, which is connected to the correlation calculation sub-unit, and is used to sort the correlation coefficient values of all candidate positions from high to low to obtain a candidate position list.

7. The artificial intelligence-based job-person matching system according to claim 6, wherein, The adjustment module includes: A delivery deviation calculation unit, which is used to calculate the relative deviation between the delivery success rate and a preset expected delivery success rate to obtain a delivery deviation; An interview deviation calculation unit, which is used to calculate the relative deviation between the interview success rate and a preset expected interview success rate to obtain an interview deviation; A deviation index calculation unit, which is respectively connected to the delivery deviation calculation unit and the interview deviation calculation unit, and is used to perform weighted summation on the delivery deviation, a preset delivery weight, the interview deviation, and a preset interview weight to obtain a deviation index; A deviation comparison unit, which is connected to the deviation index calculation unit, and is used to compare the deviation index with a preset deviation index range to obtain a deviation comparison result; An adjustment unit, which is connected to the deviation comparison unit, and is used to adjust the resume index according to the deviation index when the deviation comparison result is that the deviation index is greater than the maximum value of the preset deviation index range, or to adjust the preset similarity threshold according to the delivery deviation when the deviation comparison result is that the deviation index is less than the minimum value of the preset deviation index range.

8. The artificial intelligence-based job-person matching system according to claim 7, wherein The adjustment unit includes: A first adjustment sub-unit, which is used to adjust the resume index according to the relative deviation between the deviation index and the maximum value of the preset deviation index range and a preset first adjustment coefficient when the deviation index is greater than the maximum value of the preset deviation index range to obtain an adjusted resume index, and the relative deviation and the adjusted resume index are negatively correlated. A second adjustment subunit, configured to, when the deviation index is less than or equal to the minimum value of the preset deviation index range, adjust the preset similarity threshold according to a preset delivery deviation threshold, the relative deviation of the delivery deviation, and a preset second adjustment coefficient, to obtain an adjusted similarity threshold, where the relative deviation is positively correlated with the adjusted resume index.

9. The artificial intelligence-based job-person matching system according to claim 8, wherein The correction module includes: A count acquisition unit, configured to acquire the number of times the adjusted resume index is obtained within the preset correction duration to form an index adjustment count, and acquire the number of times the adjusted similarity threshold is obtained within the preset correction duration to form a threshold adjustment count; A count difference calculation unit, connected to the count acquisition unit, configured to calculate the absolute value of the difference between the index adjustment count and the threshold adjustment count to obtain a count difference value; A correction unit, connected to the count difference calculation unit, configured to correct the candidate job list according to the count difference value, the index adjustment count, and the threshold adjustment count.

10. The artificial intelligence-based job-person matching system according to claim 9, wherein The correction unit includes: A first correction subunit, configured to, when the count difference value is greater than a preset count difference threshold and the index adjustment count is greater than the threshold adjustment count, correct the candidate job list based on the adjusted resume index to form a corrected job list, and, when the count difference value is greater than the preset count difference threshold and the index adjustment count is less than the threshold adjustment count, correct the candidate job list based on the adjusted similarity threshold to form a corrected job list; An adjustment fluctuation calculation subunit, configured to, when the count difference value is less than or equal to the preset count difference threshold, calculate the standard deviation of the index adjustment count within the preset correction duration to obtain an index adjustment fluctuation value, and calculate the standard deviation of the threshold adjustment count within the preset correction duration to obtain a threshold adjustment fluctuation value; A fluctuation deviation calculation subunit, connected to the adjustment fluctuation calculation subunit, configured to calculate the relative deviation between the index adjustment fluctuation value and a preset index adjustment threshold to obtain an index adjustment deviation, and calculate the relative deviation between the threshold adjustment fluctuation value and a preset threshold adjustment threshold to obtain a threshold adjustment deviation; A second correction subunit, connected to the fluctuation deviation calculation subunit, configured to, when the index adjustment deviation is greater than the threshold adjustment deviation, correct the candidate job list based on the adjusted resume index to form a corrected job list, and, when the index adjustment deviation is less than or equal to the threshold adjustment deviation, correct the candidate job list based on the adjusted similarity threshold to form a corrected job list.

Images