CN117131372A

CN117131372A - Model training methods, devices, computer equipment, storage media and program products

Info

Publication number: CN117131372A
Application number: CN202310994346.7A
Authority: CN
Inventors: 党娜
Original assignee: Bank of China Ltd
Current assignee: Bank of China Ltd
Priority date: 2023-08-08
Filing date: 2023-08-08
Publication date: 2023-11-28

Abstract

The application relates to a model training method, a device, computer equipment, a storage medium and a program product, which relate to the technical field of artificial intelligence, wherein a preset interview attribute mapping table comprising attribute characteristics of staff and weights corresponding to the attribute characteristics is utilized to determine each interview scoring sample, then each interview scoring sample is utilized to train an initial interview result classification model to obtain an intermediate interview result classification model, finally, a target interview result classification model is determined according to the intermediate interview result classification model and historical interview results corresponding to the historical interview scores of each historical staff, and the staff can be evaluated by utilizing the target interview result classification model to obtain the evaluation result of the staff. By adopting the method, the accuracy of the evaluation result obtained by evaluating the job seeker can be improved.

Description

Model training method, apparatus, computer device, storage medium, and program product

Technical Field

The present application relates to the field of artificial intelligence technology, and in particular, to a model training method, apparatus, computer device, storage medium, and program product.

Background

The interview is usually used as a necessary link for screening the job seeker, and the bank interviewer can judge the work suitability of the job seeker by interviewing the education background, the mastered degree of professional knowledge, the work attitude and other aspects of the job seeker.

Typically, a bank interviewer evaluates a job seeker through personal experience according to information presented by the job seeker to obtain an evaluation result. However, the evaluation result obtained by evaluating the job seeker using personal experience is low in accuracy.

Disclosure of Invention

In view of the foregoing, it is desirable to provide a model training method, apparatus, computer device, storage medium, and program product that can utilize a trained target interview result classification model to evaluate an evaluation result obtained by a job seeker, thereby improving the accuracy of the obtained evaluation result.

In a first aspect, the present application provides a model training method. The method comprises the following steps:

determining each interview scoring sample by using a preset interview attribute mapping table; the interview attribute mapping table comprises attribute features of job seekers and weights corresponding to the attribute features;

training an initial interview result classification model by using each interview scoring sample to obtain an intermediate interview result classification model;

Determining a target interview result classification model according to the intermediate interview result classification model and the historical interview results corresponding to the historical interview scores of the historical staff seekers; the target interview result classification model is used for evaluating job seekers.

In one embodiment, the training the initial interview result classification model with each of the interview scoring samples to obtain an intermediate interview result classification model includes:

classifying each interview scoring sample according to the first initial center object and the second initial center object to determine a first initial region and a second initial region; the first initial center object and the second initial center object are any two samples in each interview scoring sample;

and training an initial interview result classification model according to the first initial region and the second initial region to obtain the intermediate interview result classification model.

In one embodiment, the classifying each of the interview scoring samples according to the first initial center object and the second initial center object to determine a first initial region and a second initial region includes:

determining a first distance from the first initial center object and a second distance from the second initial center object for a first other interview scoring sample; the first other interview scoring samples include samples of each of the interview scoring samples other than the first initial center object, the second initial center object;

And determining the first initial area and the second initial area according to the first distance and the second distance corresponding to the first other interview scoring samples.

In one embodiment, the determining the first initial region and the second initial region according to the first distance and the second distance corresponding to the first other interview scoring samples includes:

if the first distance corresponding to the first other interview scoring samples is smaller than the corresponding second distance, dividing the first other interview scoring samples into a first initial area;

and if the first distance corresponding to the first other interview scoring samples is not smaller than the corresponding second distance, dividing the first other interview scoring samples into a second initial area.

In one embodiment, the training an initial interview result classification model according to the first initial region and the second initial region to obtain the intermediate interview result classification model includes:

if the first initial area or the second initial area meets the center object updating condition, updating the first initial center object to obtain a first new center object, and updating the second initial center object to obtain a second new center object according to each interview scoring sample in the second initial area;

And returning to the step of determining a new first distance between the second other interview scoring sample and the first new center object and a new second distance between the second other interview scoring sample and the second new center object, and updating the first initial region and the second initial region according to the new first distance and the new second distance corresponding to the second other interview scoring sample to obtain a first region and a second region, so as to train the initial interview result classification model until the first region and the second region do not meet the center object updating condition, and obtaining the intermediate interview result classification model.

In one embodiment, the determining the target interview result classification model according to the intermediate interview result classification model and the historical interview results corresponding to the historical interview scores of the historical staff members includes:

inputting each historical interview score into the intermediate interview result classification model to obtain a prediction result corresponding to each historical interview score;

determining a first total number of historical interview scores for which the predicted outcome is the same as the historical interview outcome corresponding to the predicted outcome;

the target interview result classification model is determined based on the first total number and the second total number of historical interview scores.

In one embodiment, the determining the target interview result classification model based on the first total number and the second total number of historical interview scores includes:

if the first total number is greater than or equal to a preset threshold value, determining the intermediate interview result classification model as the target interview result classification model;

and if the first total number is smaller than the preset threshold value, adjusting weights corresponding to the attribute features in the interview attribute mapping table to obtain a new interview attribute mapping table, returning to execute the step of determining each new interview scoring sample by using the new interview attribute mapping table, obtaining a new first total number until the new first number is larger than or equal to the preset threshold value, and taking a middle interview result classification model corresponding to the new first number larger than or equal to the preset threshold value as the target interview result classification model.

In one embodiment, the determining the interview scoring sample using the preset interview attribute mapping table includes:

determining the product result of the score corresponding to the attribute feature and the weight corresponding to the attribute feature of each job seeker based on the preset interview attribute mapping table;

taking the summation result of the product result of the weights corresponding to the attribute features as the interview scoring sample.

In a second aspect, the application further provides a model training device. The device comprises:

the first determining module is used for determining each interview scoring sample by utilizing a preset interview attribute mapping table; the interview attribute mapping table comprises attribute features of job seekers and weights corresponding to the attribute features;

the second determining module is used for training an initial interview result classification model by utilizing each interview scoring sample to obtain an intermediate interview result classification model;

and the third determining module is used for determining a target interview result classification model according to the intermediate interview result classification model and the historical interview results corresponding to the historical interview scores of the historical staff seekers.

In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor implementing the steps of the above method when the processor executes the computer program.

In a fourth aspect, the present application also provides a computer-readable storage medium. The computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the steps of the above method.

In a fifth aspect, the present application also provides a computer program product. The computer program product comprising a computer program which, when executed by a processor, implements the steps of the above method.

According to the model training method, the device, the computer equipment, the storage medium and the program product, the interview attribute mapping table comprising the attribute characteristics of the job seeker and the weights corresponding to the attribute characteristics is utilized to determine each interview scoring sample, then the initial interview result classification model is trained by utilizing each interview scoring sample to obtain the intermediate interview result classification model, finally the target interview result classification model is determined according to the intermediate interview result classification model and the historical interview results corresponding to the historical interview scores of each historical job seeker, the job seeker can be evaluated by utilizing the target interview result classification model, and the evaluation result of the job seeker is obtained. In the traditional technology, the interviewer evaluates the job seeker through personal experience according to the information displayed by the job seeker, so that the accuracy of the evaluation result is low. According to the application, the interview scoring sample for training is obtained according to the attribute characteristics of the job seeker and the weights corresponding to the attribute characteristics, the interview scoring sample is utilized to obtain the intermediate interview result classification model, and the intermediate interview result classification model is further verified based on the historical interview result to obtain the target interview result classification model with high accuracy, so that the target interview result classification model is utilized to evaluate the job seeker, and the accuracy of the evaluation result obtained by evaluating the job seeker can be improved.

Drawings

FIG. 1 is a schematic flow chart of a model training method according to an embodiment of the present application;

FIG. 2 is a schematic flow chart of a method for determining a classification model of an intermediate interview result according to an embodiment of the present application;

FIG. 3 is a schematic flow chart of a method for determining a first initial region and a second initial region according to an embodiment of the present application;

FIG. 4 is a second flow chart of a method for determining a first initial region and a second initial region according to an embodiment of the present application;

FIG. 5 is a second flow chart of a method for determining a classification model of an intermediate interview result according to an embodiment of the present application;

FIG. 6 is a schematic flow chart of a method for determining a classification model of a target interview result according to an embodiment of the present application;

FIG. 7 is a second flowchart of a method for determining a classification model of a target interview result according to an embodiment of the present application;

FIG. 8 is a flowchart of a method for determining an interview score sample according to an embodiment of the present application;

FIG. 9 is a block diagram of a model training apparatus according to an embodiment of the present application;

fig. 10 is an internal structure diagram of a computer device according to an embodiment of the present application.

Detailed Description

The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.

Interviews are usually used as a necessary link for screening job seekers, and interviewees can judge the work suitability of job seekers by interviewing the conditions of multiple aspects such as education background, mastering degree of professional knowledge, working attitude and the like of the job seekers. In general, the interviewer evaluates the job seeker through personal experience according to information presented by the job seeker to obtain an evaluation result. However, the evaluation result obtained by evaluating the job seeker using personal experience is low in accuracy.

In order to solve the problems, a model training method, a device, computer equipment, a storage medium and a program product are provided, wherein the accuracy of an evaluation result obtained by evaluating a job seeker can be improved, and the evaluation result obtained by evaluating the job seeker is utilized by the obtained target interview result classification model. The model training method provided by the embodiment of the application can be applied to a server, and the server can be an independent server or a server cluster formed by a plurality of servers.

In one embodiment, fig. 1 is one of flow diagrams of a model training method according to an embodiment of the present application, and referring to fig. 1, a model training method is provided, which includes the following steps:

s101, determining each interview scoring sample by using a preset interview attribute mapping table; the interview attribute mapping table comprises attribute features of job seekers and weights corresponding to the attribute features.

The preset interview attribute mapping table may include a plurality of attribute features of the job seeker, and a weight corresponding to each attribute feature. For example, the attribute features of the job seeker may include a working year and a skill specialty, where the working year is less than 5 years with a corresponding weight of 0.2, the working year is 5-10 years with a corresponding weight of 0.3, the working year is greater than 10 years with a corresponding weight of 0.5, the skill specialty is proficient with a corresponding weight of 0.5, familiarity with a corresponding weight of 0.3, and knowledge of the corresponding weight is 0.2.

Specifically, the score corresponding to the attribute feature of the job applicant and the weight corresponding to the attribute feature can be multiplied to obtain the score of the job applicant in the attribute feature, and the score of the attribute feature is multiplied by a preset correction coefficient to obtain the score of the attribute feature after correction. And calculating the scores of the corrected attribute features of the job seeker by using the method, and adding the scores of the corrected attribute features to obtain an interview score sample corresponding to the job seeker.

Receiving the above example, presetting a correction coefficient to be 0.9, wherein the corresponding score of the working period is 10 points, the working period of the job seeker A is 6 years, the score of the job seeker A in the working period is 3 points, and the score of the job seeker A in the working period after correction is 2.7 points; the corresponding score of the skill specialty is 30 points, the skill specialty of the job applicant A is skilled, the score of the job applicant A in the skill specialty is 15 points, and the score of the job applicant A in the skill specialty after correction is 13.5 points. The job seeker a corresponds to a interview score sample of 16.2 points.

S102, training an initial interview result classification model by using each interview scoring sample to obtain an intermediate interview result classification model.

Specifically, the interview scoring samples may be classified for the first time according to a preset first initial center object and a preset second initial center object, and samples of the interview scoring samples except the first initial center object and the second initial center object may be classified into a first initial area centered on the first initial center object or a second initial area centered on the second initial center object. And determining a first new center object and a second new center object from the first initial area respectively, dividing the interview grading samples except the first new center object and the second new center object into the first area or the second area, repeatedly executing the method for determining the first new center object and the second new center object, training the initial interview result classification model until the first new center object in the first area or the second new center object in the second area is unchanged, and obtaining the intermediate interview result classification model.

S103, determining a target interview result classification model according to the intermediate interview result classification model and the historical interview results corresponding to the historical interview scores of the historical staff seekers; the target interview result classification model is used for evaluating job seekers.

The historical interview score is a known interview score, and the prediction result corresponding to the historical interview score comprises pass and fail.

In this embodiment, a plurality of historical interview scores may be input to the intermediate interview result classification model, prediction results corresponding to the plurality of historical interview scores may be obtained, a first total number of historical interview scores, in which the prediction results are the same as the historical interview results corresponding to the prediction results, may be determined, and if a percentage obtained by dividing the first total number by the second total number of the historical interview scores is greater than or equal to a preset percentage, the intermediate interview result classification model may be used as the target interview result classification model; and if the percentage obtained by dividing the first total number by the second total number of the historical interview scores is smaller than the preset percentage, adjusting the weight corresponding to the attribute features in the interview attribute mapping table to obtain a new interview attribute mapping table, determining each new interview score sample according to the new interview attribute mapping table, training an initial interview result classification model by using each new interview score sample, and obtaining a new intermediate interview result classification model until the percentage obtained by dividing the first total number by the second total number of the historical interview scores is larger than or equal to the preset percentage, and taking the new intermediate interview result classification model as a target interview result classification model.

In the model training method, the preset interview attribute mapping table comprising attribute features of the job seeker and weights corresponding to the attribute features is utilized to determine each interview scoring sample, then the initial interview result classification model is trained by utilizing each interview scoring sample to obtain the intermediate interview result classification model, finally, a target interview result classification model is determined according to the intermediate interview result classification model and the historical interview results corresponding to the historical interview scores of each historical job seeker, and the target interview result classification model can be utilized to evaluate the job seeker to obtain the evaluation result of the job seeker. In the traditional technology, the interviewer evaluates the job seeker through personal experience according to the information displayed by the job seeker, so that the accuracy of the evaluation result is low. According to the application, the interview scoring sample for training is obtained according to the attribute characteristics of the job seeker and the weights corresponding to the attribute characteristics, the interview scoring sample is utilized to obtain the intermediate interview result classification model, and the intermediate interview result classification model is further verified based on the historical interview result to obtain the target interview result classification model with high accuracy, so that the target interview result classification model is utilized to evaluate the job seeker, and the accuracy of the evaluation result obtained by evaluating the job seeker can be improved.

In one embodiment, fig. 2 is one of flow diagrams of a method for determining an intermediate interview result classification model according to the embodiment of the present application, referring to fig. 2, the embodiment relates to how to train an initial interview result classification model by using each interview scoring sample to obtain a possible implementation manner of the intermediate interview result classification model, and on the basis of the embodiment, S102 includes:

s201, classifying each interview scoring sample according to a first initial center object and a second initial center object to determine a first initial area and a second initial area; the first initial center object and the second initial center object are any two samples in each interview scoring sample.

Wherein the first initial center object and the second initial center object may be any two samples of the interview scoring samples. The first initial region may be a region centered on the first initial center object, and the second initial region may be a region centered on the second initial center object.

Specifically, a first other interview scoring sample except the first initial center object and the second initial center object may be determined according to centering on the first initial center object, a first distance from the first initial center object and a second distance corresponding to the second initial center object may be determined, and the first distance and the second distance may be multiplied by a preset distance correction coefficient, to obtain a corrected first distance and a corrected second distance. And dividing the first other interview scoring samples with the corrected first distance being smaller than the corresponding corrected second distance into a first initial region, and dividing the first other interview scoring samples with the corrected first distance being not smaller than the corresponding corrected second distance into a second initial region.

S202, training an initial interview result classification model according to the first initial region and the second initial region to obtain an intermediate interview result classification model.

Specifically, if the first initial area or the second initial area meets a central object updating condition, the central object updating condition may be that the first initial area and the second initial area do not reach the preset updating times, updating the first initial central object to obtain a first new central object, and updating the second initial central object to obtain a second new central object according to each interview scoring sample in the second initial area; and returning to the step of determining the new first distance between the second other interview scoring sample and the first new center object and the new second distance between the second other interview scoring sample and the second new center object, and updating the first initial area and the second initial area according to the new first distance and the new second distance corresponding to the second other interview scoring sample to obtain a first area and a second area so as to train the initial interview result classification model until the first area and the second area do not meet the center object updating condition, thereby obtaining the intermediate interview result classification model.

In this embodiment, the first initial area and the second initial area are determined by classifying each interview scoring sample according to the first initial center object and the second initial center object, and the initial interview result classification model is trained according to the first initial area and the second initial area to obtain the intermediate interview result classification model, so that the accuracy of the intermediate interview result classification model can be improved.

In one embodiment, fig. 3 is one of the flow charts of a first initial area and a second initial area determining method provided in the embodiment of the present application, referring to fig. 3, the embodiment refers to how to classify each interview scoring sample according to a first initial center object and a second initial center object to determine a possible implementation manner of the first initial area and the second initial area, where, based on the above embodiment, S201 includes:

s301, determining a first distance between a first other interview scoring sample and a first initial center object and a second distance between the first other interview scoring sample and a second initial center object; the first other interview scoring samples include samples of the interview scoring samples other than the first initial center object, the second initial center object.

Specifically, a first distance to the first initial center object and a second distance corresponding to the second initial center may be determined from a first other interview scoring sample centered on the first initial center object and the second initial center object.

For example, a first other interview scoring sample a is 3 at a first distance from a first initial central object and a second distance from a second initial central object is 5; the first other interview scoring sample B was a first distance of 6 from the first initial center object and a second distance of 3 from the second initial center object.

S302, determining a first initial area and a second initial area according to a first distance and a second distance corresponding to the first other interview scoring samples.

Specifically, the first distance and the second distance may be multiplied by a preset distance correction coefficient, respectively, to obtain a corrected first distance and a corrected second distance. And dividing the first other interview scoring samples with the corrected first distance being smaller than the corresponding corrected second distance into a first initial region, and dividing the first other interview scoring samples with the corrected first distance being not smaller than the corresponding corrected second distance into a second initial region.

For example, the preset correction coefficient is 0.9, the first distance between the first other interview scoring sample a and the first initial center object is 3, the corrected first distance is 2.7, the second distance between the first interview scoring sample a and the second initial center object is 5, and the corrected second distance is 4.5; the first distance between the first other interview scoring sample B and the first initial center object is 6, the corrected first distance is 5.4, the second distance between the first other interview scoring sample B and the second initial center object is 3, and the corrected second distance is 2.7. The first other interview scoring sample a may be partitioned into a first initial region and the first other interview scoring sample B partitioned into a second initial region.

In this embodiment, by determining the first distance between the first other interview scoring sample and the first initial center object and the second distance between the first other interview scoring sample and the second initial center object, and determining the first initial region and the second initial region according to the first distance and the second distance corresponding to the first other interview scoring sample, the first classification is equivalent to the first classification of the interview scoring sample, and the first other interview scoring sample in the first initial region and the first other interview scoring sample in the second initial region after the first classification can be determined, which is equivalent to the preliminary prediction of the interview scoring sample.

In an embodiment, fig. 4 is a second flowchart of a first initial area and second initial area determining method according to the embodiment of the present application, referring to fig. 4, the embodiment relates to how to determine a possible implementation manner of the first initial area and the second initial area according to a first distance and a second distance corresponding to a first other interview scoring sample, where, based on the embodiment, S302 includes:

s401, if the first distance corresponding to the first other interview scoring samples is smaller than the corresponding second distance, dividing the first other interview scoring samples into a first initial area.

In this embodiment, if the first distance corresponding to the first other interview scoring sample is less than the second distance corresponding to the first other interview scoring sample, such first other interview scoring sample may be partitioned into the first initial region.

For example, if the first other interview scoring sample a is a first distance of 5 from the first initial center object and the second distance of 8 from the second initial center object, then the first other interview scoring sample a may be partitioned into a first initial region.

And S402, if the first distance corresponding to the first other interview scoring samples is not smaller than the second distance corresponding to the first other interview scoring samples, dividing the first other interview scoring samples into a second initial area.

In this embodiment, if the first distance corresponding to the first other interview scoring sample is not less than the second distance corresponding to the first other interview scoring sample, such first other interview scoring sample may be partitioned into the second initial region.

For example, if the first other interview score sample B is a first distance of 6 from the first initial center object and the second distance of 1 from the second initial center object, then the first other interview score sample B may be partitioned into the second initial region.

In this embodiment, if the first distance corresponding to the first other interview scoring sample is smaller than the corresponding second distance, the first other interview scoring sample is divided into the first initial area, and if the first distance corresponding to the first other interview scoring sample is not smaller than the corresponding second distance, the first other interview scoring sample is divided into the second initial area, so that the first other interview scoring sample in the first initial area and the first other interview scoring sample in the second initial area after the first classification can be determined, and preliminary prediction is performed on the interview scoring sample.

In an embodiment, fig. 5 is a second flow chart of a method for determining an intermediate interview result classification model according to the embodiment of the present application, where the embodiment relates to how to train an initial interview result classification model according to a first initial area and a second initial area to obtain a possible implementation manner of the intermediate interview result classification model, and on the basis of the embodiment, S202 includes:

s501, if the first initial area or the second initial area meets the center object updating condition, updating the first initial center object to obtain a first new center object, and updating the second initial center object according to each interview scoring sample in the second initial area to obtain a second new center object.

Wherein the center object update condition may include that a position of the division of the interview score sample into the first initial region or the second initial region is not fixed, that the first center object or the second center object is still different from the previous time, and the like.

Specifically, if the first initial region or the second initial region satisfies the center object update condition, a median or a mean value of the first initial region may be calculated, the median or the mean value is taken as a first new center object, and a median or a mean value of the second initial region is calculated, and the median or the mean value is taken as a second new center object.

S502, returning to execute the step of determining the new first distance between the second other interview scoring samples and the first new center object and the new second distance between the second other interview scoring samples and the second new center object, and updating the first initial area and the second initial area according to the new first distance and the new second distance corresponding to the second other interview scoring samples to obtain a first area and a second area so as to train the initial interview result classification model until the first area and the second area do not meet the center object updating condition, thereby obtaining the intermediate interview result classification model.

Wherein the second other interview scoring samples include samples of the interview scoring samples other than the first new center object and the second new center object.

In the embodiment of the present application, the step of determining the new first distance between the second other interview scoring sample and the first new center object and the new second distance between the second other interview scoring sample and the second new center object may be performed in a returning manner, if the new first distance corresponding to the second other interview scoring sample is smaller than the corresponding new second distance, the second other interview scoring sample may be divided into the first area, and if the new first distance corresponding to the second other interview scoring sample is not smaller than the corresponding new second distance, the second other interview scoring sample may be divided into the second area until the first area and the second area do not satisfy the center object updating condition, thereby obtaining the intermediate interview result classification model.

In this embodiment, if the first initial area or the second initial area satisfies the center object updating condition, the first initial center object is updated to obtain a first new center object, and the second initial center object is updated to obtain a second new center object according to each interview score sample in the second initial area, the step of determining a new first distance between the second other interview score sample and the first new center object and a new second distance between the second other interview score sample and the second new center object is performed, and the first initial area and the second initial area are updated to obtain the first area and the second area according to the new first distance and the new second distance corresponding to the second other interview score sample, so as to train the initial interview result classification model until the first area and the second area do not satisfy the center object updating condition, so as to obtain an intermediate interview result classification model, and the intermediate interview result classification model with high accuracy can be obtained by continuously classifying the interview score samples, and further, the accuracy of the evaluation result obtained by evaluating staff can be improved.

In one embodiment, fig. 6 is one of the flow diagrams of a method for determining a target interview result classification model according to the embodiment of the present application, where the embodiment relates to how to determine a possible implementation manner of the target interview result classification model according to the intermediate interview result classification model and the historical interview results corresponding to the historical interview scores of the historical staff members, and on the basis of the above embodiment, the step S103 includes:

S601, inputting each historical interview score into an intermediate interview result classification model to obtain a prediction result corresponding to each historical interview score.

The historical interview score is a known interview score, and the prediction result corresponding to the historical interview score includes pass and fail, for example, the historical interview score of the historical job seeker B is 68 points, and the prediction result is pass.

In the embodiment of the application, one historical interview score is input into the intermediate interview result classification model, so that a predicted result corresponding to the historical interview score can be obtained, and a plurality of historical interview scores are respectively input into the intermediate interview result classification model, so that a predicted result corresponding to the historical interview scores can be obtained.

S602, determining a first total number of historical interview scores, wherein the predicted result is identical to the historical interview result corresponding to the predicted result.

The historical interview result is a known result, and the predicted result is a result obtained after the result is input into the intermediate interview result classification model.

Specifically, it may be first determined whether the predicted outcome corresponding to each of the historical interview scores is the same as the historical interview outcome corresponding to the predicted outcome, and then a first total number of historical interview scores for which the predicted outcome is the same as the historical interview outcome corresponding to the predicted outcome may be determined.

For example, there are 10 predictions in total, and the same historical interview score as the corresponding historical interview result of the predictions out of the 10 predictions is 8.

S603, determining a target interview result classification model according to the first total number and the second total number of the historical interview scores.

Specifically, a percentage obtained by dividing the first total number by the second total number of the historical interview scores may be determined, and if the percentage is greater than or equal to a preset percentage, the intermediate interview result classification model may be used as the target interview result classification model. And if the percentage is smaller than the preset percentage, adjusting the weight corresponding to the attribute characteristics in the interview attribute mapping table to obtain a new interview attribute mapping table, re-obtaining a new interview scoring sample, and determining a new intermediate interview result classification model until the percentage is larger than or equal to the preset percentage, and determining the new intermediate interview result classification model as a target interview result classification model.

For example, the preset percentage is 80%, the second total number of the historical interview scores is 10, the first total number of the historical interview scores, which are the same as the corresponding historical interview results of the predicted results, is 8, it may be determined that the percentage obtained by dividing the first total number by the second total number of the historical interview scores is equal to the preset percentage, and the interview result classification model may be used as the target interview result classification model.

In this embodiment, by inputting each historical interview score to the intermediate interview result classification model to obtain a prediction result corresponding to each historical interview score, determining a first total number of historical interview scores, where the prediction result is the same as the historical interview result corresponding to the prediction result, and determining the target interview result classification model according to the first total number and the second total number of the historical interview scores, the intermediate interview result classification model can be verified to obtain a more accurate target interview result classification model.

In an embodiment, fig. 7 is a second flowchart of a method for determining a target interview result classification model according to an embodiment of the present application, where the embodiment relates to a possible implementation manner of determining the target interview result classification model according to the first total number and the second total number of historical interview scores, and on the basis of the embodiment, S603 includes:

and S701, if the first total number is greater than or equal to a preset threshold value, determining the intermediate interview result classification model as a target interview result classification model.

In this embodiment, if the first total number of the historical interview scores, for which the prediction result is the same as the historical interview result corresponding to the prediction result, is greater than or equal to the preset threshold, the intermediate interview result classification model may be used as the target interview result classification model.

For example, if the preset threshold is 6 and the first total number of the historical interview scores, of which the predicted results are the same as the historical interview results corresponding to the predicted results, is 7, the intermediate interview result classification model may be used as the target interview result classification model.

S702, if the first total number is smaller than a preset threshold, adjusting weights corresponding to all attribute features in the interview attribute mapping table to obtain a new interview attribute mapping table, returning to execute the step of determining all new interview scoring samples by using the new interview attribute mapping table, obtaining a new first total number until the new first number is greater than or equal to the preset threshold, and taking an intermediate interview result classification model corresponding to the new first number greater than or equal to the preset threshold as a target interview result classification model.

In this embodiment, if the first total number of historical interview scores, where the predicted result is the same as the historical interview result corresponding to the predicted result, is smaller than a preset threshold, the weight corresponding to each attribute feature in the interview attribute mapping table may be adjusted to obtain a new interview attribute mapping table, and the step of determining each new interview score sample using the new interview attribute mapping table is performed back to obtain the new first total number, until the new first number is greater than or equal to the preset threshold, then the intermediate interview result classification model corresponding to the new first number greater than or equal to the preset threshold may be used as the target interview result classification model. The adjusting the weights corresponding to the attribute features in the interview attribute mapping table may be to multiply the weights corresponding to the attribute features with a preset weight correction coefficient to obtain new weights.

In this embodiment, if the first total number is greater than or equal to a preset threshold, determining that the intermediate interview result classification model is the target interview result classification model, if the first total number is less than the preset threshold, adjusting weights corresponding to all attribute features in the interview attribute mapping table to obtain a new interview attribute mapping table, and returning to execute the step of determining each new interview scoring sample by using the new interview attribute mapping table, and obtaining the new first total number until the new first number is greater than or equal to the preset threshold, and taking the intermediate interview result classification model corresponding to the new first number greater than or equal to the preset threshold as the target interview result classification model, so that a more accurate target interview result classification model can be obtained.

In one embodiment, fig. 8 is a flowchart of a method for determining an interview score sample according to the embodiment of the present application, where the embodiment relates to a possible implementation manner of determining the interview score sample by using a preset interview attribute mapping table, and on the basis of the foregoing embodiment, the step S101 includes:

s801, determining a product result of scores corresponding to attribute features and weights corresponding to the attribute features of all job seekers based on a preset interview attribute mapping table.

In the embodiment of the application, the preset interview attribute mapping table can comprise a plurality of attribute features of the job seeker and weights corresponding to each attribute feature, and the scores corresponding to the attribute features of each job seeker can be multiplied by the weights corresponding to the attribute features to obtain a product result.

For example, as shown in table 1, the preset interview attribute mapping table includes gender, home address, working year, overtime acceptance, skill expertise, character matching degree, and the like, and each attribute feature has a corresponding weight. The score corresponding to the gender is 10 points, and the gender of the job seeker A is male, and the score result of the score corresponding to the gender and the weight corresponding to the gender of the job seeker A is 7 points.

S802, taking the summation result of the product results of the weights corresponding to the attribute features as an interview scoring sample.

Specifically, the summation result of the product results of the weights corresponding to the attribute features of one job seeker can be used as an interview scoring sample, and the summation result of the product results of the weights corresponding to the attribute features of a plurality of job seekers can be determined according to the method, so that a plurality of interview scoring samples can be obtained.

With reference to table 2, the result of the product of the weights corresponding to the gender of the job seeker a is 7 points, the result of the product of the weights corresponding to the home address is 2 points, the result of the product of the weights corresponding to the working years is 35 points, the result of the product of the weights corresponding to the overtime acceptance is 10 points, the result of the product of the weights corresponding to the skill expertise is 20 points, the result of the product corresponding to the character matching degree is 10 points, and the interview scoring sample corresponding to the job seeker a is 84 points.

In this embodiment, the product result of the score corresponding to the attribute feature and the weight corresponding to the attribute feature of each job seeker is determined based on the preset interview attribute mapping table, and the summation result of the product result of the weights corresponding to the attribute features is used as the interview scoring sample, so that the interview scoring sample of the job seeker can be obtained in an omnibearing manner through the diversified attribute features, and further, the accuracy of evaluating the job seeker can be improved.

It should be understood that, although the steps in the flowcharts related to the above embodiments are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.

Based on the same inventive concept, the embodiment of the application also provides a model training device for realizing the model training method. The implementation of the solution provided by the device is similar to the implementation described in the above method, so the specific limitation in one or more embodiments of the model training device provided below may be referred to above for the limitation of the model training method, which is not repeated here.

In one embodiment, fig. 9 is a block diagram of a model training apparatus according to an embodiment of the present application, and referring to fig. 9, a model training apparatus 900 is provided, including: a first determination module 901, a second determination module 902, and a third determination module 903, wherein:

a first determining module 901, configured to determine each interview scoring sample by using a preset interview attribute mapping table; the interview attribute mapping table comprises attribute features of job seekers and weights corresponding to the attribute features.

And a second determining module 902, configured to train the initial interview result classification model by using each interview scoring sample, so as to obtain an intermediate interview result classification model.

A third determining module 903, configured to determine a target interview result classification model according to the intermediate interview result classification model and the historical interview results corresponding to the historical interview scores of each historical job seeker; the target interview result classification model is used for evaluating job seekers.

In one embodiment, the second determining module 902 includes:

the first determining submodule is used for classifying each interview scoring sample according to the first initial center object and the second initial center object so as to determine a first initial area and a second initial area; the first initial center object and the second initial center object are any two samples in each interview scoring sample.

And the training sub-module is used for training the initial interview result classification model according to the first initial region and the second initial region to obtain the intermediate interview result classification model.

In one embodiment, the first determination submodule includes:

a first determining unit for determining a first distance of the first other interview scoring sample from the first initial center object and a second distance from the second initial center object; the first other interview scoring samples include samples of the interview scoring samples other than the first initial center object, the second initial center object.

And the second determining unit is used for determining the first initial area and the second initial area according to the first distance and the second distance corresponding to the first other interview scoring samples.

In one embodiment, the first determining unit is specifically configured to divide the first other interview scoring samples into the first initial region if the first distance corresponding to the first other interview scoring samples is smaller than the second distance corresponding to the first other interview scoring samples; and if the first distance corresponding to the first other interview scoring samples is not smaller than the corresponding second distance, dividing the first other interview scoring samples into a second initial area.

In one embodiment, the training submodule includes:

and the updating unit is used for updating the first initial center object to obtain a first new center object if the first initial area or the second initial area meets the center object updating condition, and updating the second initial center object according to each interview grading sample in the second initial area to obtain a second new center object.

And the third determining unit is used for returning to execute the steps of determining the new first distance between the second other interview scoring sample and the first new center object and the new second distance between the second other interview scoring sample and the second new center object, updating the first initial area and the second initial area according to the new first distance and the new second distance corresponding to the second other interview scoring sample to obtain a first area and a second area, and training the initial interview result classification model until the first area and the second area do not meet the center object updating condition to obtain the intermediate interview result classification model.

In one embodiment, the third determination module 903 includes:

and the second determination submodule is used for inputting each historical interview score into the intermediate interview result classification model to obtain a prediction result corresponding to each historical interview score.

And a third determination sub-module for determining a first total number of historical interview scores for which the predicted outcome is the same as the historical interview outcome to which the predicted outcome corresponds.

And a fourth determination sub-module for determining a target interview result classification model based on the first total number and the second total number of historical interview scores.

In one embodiment, the fourth determining submodule is specifically configured to determine the intermediate interview result classification model as the target interview result classification model if the first total number is greater than or equal to a preset threshold; and if the first total number is smaller than the preset threshold value, adjusting weights corresponding to all attribute features in the interview attribute mapping table to obtain a new interview attribute mapping table, returning to execute the step of determining all new interview scoring samples by using the new interview attribute mapping table, obtaining a new first total number until the new first number is larger than or equal to the preset threshold value, and taking an intermediate interview result classification model corresponding to the new first number larger than or equal to the preset threshold value as a target interview result classification model.

In one embodiment, the first determining module 901 includes:

and a fifth determining sub-module, configured to determine a product result of the score corresponding to the attribute feature and the weight corresponding to the attribute feature of each job seeker based on a preset interview attribute mapping table.

And a sixth determining submodule, configured to take a summation result of product results of weights corresponding to the attribute features as an interview scoring sample.

The various modules in the model training apparatus described above may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 10. The computer device includes a processor, a memory, an Input/Output interface (I/O) and a communication interface. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is for storing data. The input/output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a model training method.

It will be appreciated by those skilled in the art that the structure shown in FIG. 10 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.

In one embodiment, a computer device is provided comprising a memory and a processor, the memory having stored therein a computer program, the processor when executing the computer program performing the steps of:

determining a target interview result classification model according to the intermediate interview result classification model and the historical interview results corresponding to the historical interview scores of the historical staff members; the target interview result classification model is used for evaluating job seekers.

In one embodiment, the processor when executing the computer program further performs the steps of:

and training an initial interview result classification model according to the first initial region and the second initial region to obtain an intermediate interview result classification model.

determining a first distance from the first initial center object and a second distance from the second initial center object for the first other interview scoring samples; the first other interview scoring samples include samples of each interview scoring sample other than the first initial central object, the second initial central object;

and determining a first initial area and a second initial area according to the first distance and the second distance corresponding to the first other interview scoring samples.

if the first initial area or the second initial area meets the center object updating condition, updating the first initial center object to obtain a first new center object, and updating the second initial center object according to each interview grading sample in the second initial area to obtain a second new center object;

and returning to the step of determining the new first distance between the second other interview scoring sample and the first new center object and the new second distance between the second other interview scoring sample and the second new center object, and updating the first initial area and the second initial area according to the new first distance and the new second distance corresponding to the second other interview scoring sample to obtain a first area and a second area so as to train the initial interview result classification model until the first area and the second area do not meet the center object updating condition, thereby obtaining the intermediate interview result classification model.

determining a first total number of historical interview scores for which the predicted result is the same as the historical interview result corresponding to the predicted result;

a target interview result classification model is determined based on the first total number and the second total number of historical interview scores.

if the first total number is greater than or equal to a preset threshold value, determining the intermediate interview result classification model as a target interview result classification model;

and if the first total number is smaller than the preset threshold value, adjusting weights corresponding to all attribute features in the interview attribute mapping table to obtain a new interview attribute mapping table, returning to execute the step of determining all new interview scoring samples by using the new interview attribute mapping table, obtaining a new first total number until the new first number is larger than or equal to the preset threshold value, and taking an intermediate interview result classification model corresponding to the new first number larger than or equal to the preset threshold value as a target interview result classification model.

Determining the product result of the score corresponding to the attribute feature and the weight corresponding to the attribute feature of each job seeker based on a preset interview attribute mapping table;

and taking the summation result of the product results of the weights corresponding to the attribute features as an interview scoring sample.

In one embodiment, a computer readable storage medium is provided having a computer program stored thereon, which when executed by a processor, performs the steps of:

In one embodiment, the computer program when executed by the processor further performs the steps of:

In one embodiment, a computer program product is provided comprising a computer program which, when executed by a processor, performs the steps of:

It should be noted that, the user information (including but not limited to user equipment information, user personal information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present application are information and data authorized by the user or sufficiently authorized by each party, and the collection, use and processing of the related data need to comply with the related laws and regulations and standards of the related country and region.

Those skilled in the art will appreciate that implementing all or part of the above-described methods in accordance with the embodiments may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the embodiments provided herein may include at least one of a relational database and a non-relational database. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processor referred to in the embodiments provided in the present application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, or the like, but is not limited thereto.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The foregoing examples illustrate only a few embodiments of the application, which are described in detail and are not to be construed as limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of the application should be assessed as that of the appended claims.

Claims

1. A method of model training, the method comprising:

2. The method of claim 1, wherein training an initial interview result classification model using each of the interview scoring samples to obtain an intermediate interview result classification model, comprising:

classifying each interview scoring sample according to a first initial center object and a second initial center object to determine a first initial region and a second initial region; the first initial center object and the second initial center object are any two samples in the interview scoring samples;

3. The method of claim 2, wherein classifying each of the interview scoring samples based on a first initial center object and a second initial center object to determine a first initial region and a second initial region comprises:

4. The method of claim 3, wherein the determining a first initial region and a second initial region from the first distance and the second distance corresponding to the first other interview scoring sample comprises:

5. The method of claim 4, wherein training an initial interview result classification model based on the first initial region and the second initial region to obtain the intermediate interview result classification model comprises:

and returning to the step of determining a new first distance between a second other interview scoring sample and the first new center object and a new second distance between the second other interview scoring sample and the second new center object, and updating the first initial region and the second initial region according to the new first distance and the new second distance corresponding to the second other interview scoring sample to obtain a first region and a second region so as to train the initial interview result classification model until the first region and the second region do not meet the center object updating condition, thereby obtaining the intermediate interview result classification model.

6. The method of claim 1, wherein the determining the target interview result classification model based on the intermediate interview result classification model and the historical interview results corresponding to the historical interview scores of each historical job applicant comprises:

determining the target interview result classification model based on the first total number and the second total number of historical interview scores.

7. The method of claim 6, wherein the determining the target interview result classification model based on the first total number and the second total number of historical interview scores comprises:

and if the first total number is smaller than the preset threshold value, adjusting weights corresponding to all the attribute features in the interview attribute mapping table to obtain a new interview attribute mapping table, returning to execute the step of determining each new interview scoring sample by using the new interview attribute mapping table, obtaining a new first total number until the new first number is larger than or equal to the preset threshold value, and taking an intermediate interview result classification model corresponding to the new first number larger than or equal to the preset threshold value as the target interview result classification model.

8. The method of claim 1, wherein determining the interview scoring samples using a preset interview attribute mapping table comprises:

determining a product result of a score corresponding to the attribute feature and a weight corresponding to the attribute feature of each job seeker based on a preset interview attribute mapping table;

and taking the summation result of the product result of the weights corresponding to the attribute features as the interview scoring sample.

9. A model training apparatus, the apparatus comprising:

the third determining module is used for determining a target interview result classification model according to the intermediate interview result classification model and the historical interview results corresponding to the historical interview scores of the historical staff seekers; the target interview result classification model is used for evaluating job seekers.

10. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 8 when the computer program is executed.

11. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 8.

12. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements the steps of the method of any one of claims 1 to 8.