CN118195561A - Human resource management method and system with intelligent accounting salary - Google Patents
Human resource management method and system with intelligent accounting salary Download PDFInfo
- Publication number
- CN118195561A CN118195561A CN202410335821.4A CN202410335821A CN118195561A CN 118195561 A CN118195561 A CN 118195561A CN 202410335821 A CN202410335821 A CN 202410335821A CN 118195561 A CN118195561 A CN 118195561A
- Authority
- CN
- China
- Prior art keywords
- employee
- abnormal
- staff
- working
- salary
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000007726 management method Methods 0.000 title claims abstract description 29
- 230000002159 abnormal effect Effects 0.000 claims abstract description 190
- 238000001514 detection method Methods 0.000 claims abstract description 134
- 206010000117 Abnormal behaviour Diseases 0.000 claims abstract description 92
- 230000005856 abnormality Effects 0.000 claims abstract description 50
- 238000000034 method Methods 0.000 claims abstract description 49
- 238000004364 calculation method Methods 0.000 claims abstract description 48
- 238000010586 diagram Methods 0.000 claims description 48
- 230000008569 process Effects 0.000 claims description 30
- 238000004458 analytical method Methods 0.000 claims description 25
- 238000000605 extraction Methods 0.000 claims description 23
- 238000011176 pooling Methods 0.000 claims description 21
- 239000013598 vector Substances 0.000 claims description 14
- 238000004080 punching Methods 0.000 claims description 12
- 238000013523 data management Methods 0.000 claims description 7
- 230000004927 fusion Effects 0.000 claims description 7
- 238000005096 rolling process Methods 0.000 claims description 7
- 238000007689 inspection Methods 0.000 claims description 3
- 230000006399 behavior Effects 0.000 abstract description 34
- 230000000694 effects Effects 0.000 abstract description 9
- 239000002699 waste material Substances 0.000 abstract description 2
- 230000008901 benefit Effects 0.000 description 5
- RWSOTUBLDIXVET-UHFFFAOYSA-N Dihydrogen sulfide Chemical compound S RWSOTUBLDIXVET-UHFFFAOYSA-N 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000011835 investigation Methods 0.000 description 2
- 229940101532 meted Drugs 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000007115 recruitment Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of salary calculation, in particular to a human resource management method and system with intelligent accounting salary. According to the invention, firstly, an abnormal behavior detection model is established, and the model is used for detecting the abnormality of staff at the working position and at the non-working position, so that the accuracy of behavior judgment is improved. The abnormal behavior detection model improves the effect of detecting the abnormal behaviors of staff, so that the accuracy of salary accounting is improved through sufficient abnormal data; secondly, the invention provides an abnormal content detection model for detecting the abnormality of the working equipment of staff, and accurately judging from the working content angle, thereby improving the accuracy of salary accounting; finally, the invention provides a staff salary checking method by combining the output results of the abnormal behavior detection model and the abnormal content detection model, and staff salary can be checked more accurately by the method, so that the waste of company financial resources is avoided.
Description
Technical Field
The invention relates to the technical field of salary calculation, in particular to a human resource management method and system with intelligent accounting salary.
Background
Human resource management is an important component of enterprise management, and can scientifically and effectively manage company staff. The accounting salary is a particularly important ring in the human resource management method, and the accurate accounting salary method not only can ensure the income of staff, but also can avoid unnecessary financial loss of companies.
The traditional accounting salary method is mainly calculated according to basic salary, benefits, rewards, performance and other aspects of staff; however, by means of the accounting salary method, only rough calculation is performed, and accurate statistics cannot be achieved. The staff is lazy and small in working time period, so that the company can also incorporate invalid working time into salary accounting, which is unfavorable for accurate accounting of staff effective working time and corresponding effective time, and leads to loss of enterprise economic resources. In order to solve the problem, the invention provides a human resource management method and a human resource management system with intelligent accounting salary.
Disclosure of Invention
A human resource management method and system with intelligent accounting salary are used for solving the technical problems. Firstly, judging whether an employee works seriously in a working time period by establishing an abnormal behavior detection model, and evaluating an abnormal grade to judge the behavior of the employee; secondly, in order to further judge the working state of the staff, the invention provides a working content detection model for judging whether the staff is carrying out the operation of related work or not; finally, the invention re-calculates the salary of the staff by combining the abnormal results of the staff working state and the working content.
The human resource management method with intelligent accounting salary comprises the following steps:
acquiring a first sign-in request and a first position of an employee at a first moment; the first position comprises position information in a preset area range; recording the first moment as a first sign-in time;
Acquiring a second sign-in request and a second position of the employee at a second moment; wherein the second position comprises position information in a preset area range; recording the second moment as a second check-in time;
Further, working out the working time of the staff according to the first check-in time and the second check-in time;
Further, collecting work information of the staff in the work time by using internal equipment of the company; wherein, the work information includes: working state information and working content information; the status information includes: the working state information and the non-working state information; the work content information includes: related work operation information and non-related work operation information.
Further, the working state information is input into an abnormal behavior detection model, and an abnormal working behavior detection list of the staff is output;
The abnormal behavior detection model is used for detecting whether the limb state of the staff is in a working state or not; the implementation process of the abnormal behavior detection model comprises the following steps:
Acquiring an office position image of the staff;
Judging whether the employee is in a working area at the current moment according to the image;
Recording as off-location time when the employee is not in the work area; acquiring the time of returning the staff to the working area according to the image, recording the time as the in-place time, and calculating the out-of-place time according to the out-of-place time and the in-place time;
Judging whether the dislocation duration is greater than a preset first duration threshold value or not; when the dislocation time length is larger than the first time length threshold value, marking abnormal states, and evaluating abnormal grades according to errors of the dislocation time length and the first time length threshold value; when the dislocation duration is equal to or less than the first time duration threshold, the marking state is normal;
Inputting an employee image into an abnormal behavior analysis network when the employee is in the working area; the abnormal behavior analysis network performs behavior analysis by inputting the employee image, and the specific steps include:
Step one, inputting the employee image into a first convolution layer to obtain a first characteristic diagram of the employee;
Step two, inputting the employee first feature map into a second convolution layer to obtain an employee second feature map;
step three, inputting the second characteristic diagram of the staff into a third rolling layer and a multi-scale information extraction module to obtain a third characteristic diagram of the staff and a first multi-scale characteristic diagram of the staff;
The implementation process of the multi-scale information extraction module comprises the following steps:
inputting the employee second feature map to a 1×1 convolution kernel to obtain an employee first convolution feature map; inputting the employee convolution feature map to a convolution kernel of 3×3 to obtain a second employee convolution feature map; the staff second convolution feature images are subjected to average pooling to obtain staff feature vectors;
inputting the employee second convolution feature map and the employee feature vector into a feature fusion layer, and outputting the employee first multi-scale feature map;
inputting the employee third feature map to a fourth convolution layer and the multi-scale information extraction module to obtain an employee fourth feature map and an employee second multi-scale feature map;
step five, inputting the fourth characteristic diagram of the staff into a fifth convolution layer to obtain a fifth characteristic diagram of the staff;
Step six, inputting the employee fifth feature map to a pool layer to obtain an employee pooling feature map;
And step seven, splicing the employee first multi-scale feature map, the employee second multi-scale feature map and the employee pooling feature map according to channel dimensions, and outputting the abnormal behavior type by a full-connection layer.
Further, the abnormal working behavior detection list is obtained according to the abnormal behavior type and the abnormal behavior duration.
Further, inputting the work content information into an abnormal content detection model; the abnormal content detection model is used for detecting whether the current working equipment of the staff is used for a relevant working state or not; the implementation process of the abnormal content detection model comprises the following steps:
Acquiring current work content information of the staff;
judging whether the input device of the current working device works or not;
When the input equipment is in an unoperated state, recording unoperated time; whether the non-working time length is greater than a preset second time length threshold value; marking abnormal states when the non-working time length is larger than the second time length threshold value, and evaluating abnormal grades according to errors of the non-working time length threshold value; when the non-working time length is smaller than or equal to the second time length threshold value, the marking state is normal;
when the input equipment is in a working state;
performing relevance matching on the current working content information and the operation of the input equipment; when the correlation matching degree is higher than a preset correlation matching degree threshold value, the marking state is normal; when the correlation matching degree is lower than the correlation matching degree threshold value, comparing the operation duration with a third duration threshold value; when the operation time length is greater than the third time length threshold value, marking the abnormal state, and evaluating the abnormal grade according to the error between the operation time length and the third time length threshold value; when the operation duration is less than or equal to the third duration threshold, the marking state is normal;
And obtaining the abnormal working content detection list according to the content abnormality and the abnormality duration of the working equipment.
Further, calculating overall abnormal values in the working time according to the abnormal working behavior detection list and the abnormal working content detection list, and judging overall abnormal grades; the calculation formula of the total abnormal value is as follows:
；
Wherein, Representing the total abnormal value of the staff in the working time period; /(I)A weight value expressed as an nth anomaly in the abnormal working behavior detection list; /(I)A value represented as an nth anomaly in the abnormal work behavior detection list; /(I)A weight value expressed as an mth abnormality in the abnormal work content detection list; and the value of the mth abnormality in the abnormal work content detection list is represented.
Further, calculating the amount of the fine on the same day in the working time according to the total abnormal value;
The calculation formula of the fine amount on the same day is as follows:
；
Wherein, Representing the amount of the fine on the current day; /(I)The abnormal value corresponding to the overall abnormal level k is expressed, and k is a natural number greater than or equal to 0; /(I)Penalty coefficients expressed as overall anomaly level k; /(I)Represented as penalty amount corresponding to the anomaly level k.
Further, calculating the accounting salary of the staff according to the current punishment amount; the calculation formula of the accounting salary is as follows:
；
wherein ESlaray represents the payroll amount for the employee; BS represents the basic salary of the employee; The amount of fines expressed as t days; PB represents the project prize for the employee.
A human resource management system with intelligent accounting salary, the system comprising: the system comprises an employee data management unit, an attendance unit, a working state examination unit, a salary statistics unit, a salary display unit and an abnormal feedback unit, wherein the employee data management unit is used for storing employee data in a company;
The attendance checking unit is used for recording the daily attendance and attendance card punching conditions of staff;
The working state examination unit is used for examining whether the staff works seriously in the working time;
The operation state examination unit includes: an abnormal behavior detection module and an abnormal work content detection module; the abnormal behavior detection module is used for detecting whether the limb state of the staff is in a working state or not; the implementation process of the abnormal behavior detection model comprises the following steps:
Acquiring an office position image of the staff;
Judging whether the employee is in a working area at the current moment according to the image;
Recording as off-location time when the employee is not in the work area; acquiring the time for the staff to return to the working area according to the image and recording the working time;
Calculating the dislocation duration according to the non-working time and the working time;
Judging whether the dislocation duration is greater than a preset first duration threshold value or not; when the dislocation time length is larger than the first time length threshold value, marking abnormal states, and evaluating abnormal grades according to errors of the dislocation time length and the first time length threshold value; when the dislocation duration is equal to or less than the first time duration threshold, the marking state is normal;
inputting an employee image into an abnormal behavior analysis network when the employee is in the working area; outputting the abnormal behavior type and the abnormal behavior duration of the employee;
The implementation specific steps of the abnormal behavior analysis network of the personnel comprise:
Step one, inputting the employee image into a first convolution layer to obtain a first characteristic diagram of the employee;
Step two, inputting the employee first feature map into a second convolution layer to obtain an employee second feature map;
step three, inputting the second characteristic diagram of the staff into a third rolling layer and a multi-scale information extraction module to obtain a third characteristic diagram of the staff and a first multi-scale characteristic diagram of the staff;
The implementation process of the multi-scale information extraction module comprises the following steps:
inputting the employee second feature map to a 1×1 convolution kernel to obtain an employee first convolution feature map; inputting the employee convolution feature map to a convolution kernel of 3×3 to obtain a second employee convolution feature map; the staff second convolution feature images are subjected to average pooling to obtain staff feature vectors;
inputting the employee second convolution feature map and the employee feature vector into a feature fusion layer, and outputting the employee first multi-scale feature map;
inputting the employee third feature map to a fourth convolution layer and the multi-scale information extraction module to obtain an employee fourth feature map and an employee second multi-scale feature map;
step five, inputting the fourth characteristic diagram of the staff into a fifth convolution layer to obtain a fifth characteristic diagram of the staff;
Step six, inputting the employee fifth feature map to a pool layer to obtain an employee pooling feature map;
And step seven, splicing the employee first multi-scale feature map, the employee second multi-scale feature map and the employee pooling feature map according to channel dimensions, and outputting the abnormal behavior type by a full-connection layer.
Assessing the abnormality grade of the employee according to the abnormality behavior type and the abnormality behavior duration;
The abnormal work content detection module is used for detecting whether the current work equipment of the staff is used for the relevant work state; the implementation process of the abnormal content detection model comprises the following steps:
Acquiring current work content information of the staff;
judging whether the input device of the current working device works or not;
when the input equipment is in an inactive state;
recording the non-working time; whether the non-working time length is greater than a preset second time length threshold value; marking abnormal states when the non-working time length is larger than the second time length threshold value, and evaluating abnormal grades according to errors of the non-working time length threshold value; when the non-working time length is smaller than or equal to the second time length threshold value, the marking state is normal;
when the input equipment is in a working state;
Performing relevance matching on the current working content information and the operation of the input equipment; when the correlation matching degree is higher than a preset correlation matching degree threshold value, the marking state is normal; when the correlation matching degree is lower than the correlation matching degree threshold value, comparing the duration of the operation with a third duration threshold value; when the operation time length is longer than the third threshold time length value, marking the abnormal state, and evaluating the abnormal grade according to the error between the operation time length and the third threshold time length value; and marking that the state is normal when the operation duration is less than or equal to the third threshold duration value.
The salary statistics unit includes: a salary calculation module and the salary accounting module;
The salary calculation module is used for calculating salary of the staff based on the salary influence data corresponding to the staff according to the salary calculation rules corresponding to the staff, and the salary accounting unit is used for accounting the salary condition of the staff;
The salary accounting module is used for recalculating the salary calculating module according to the information of the working state examining unit and calculating the amount of the fine on the same day in the working time according to the total abnormal grade;
Calculating the total abnormal value in the working time according to the abnormal working behavior detection and the abnormal working content detection, and judging the total abnormal grade; the calculation formula of the total abnormal value is as follows:
；
Wherein, Representing the total abnormal value of the staff in the working time period; /(I)A weight value expressed as an nth anomaly in the abnormal working behavior detection list; /(I)A value represented as an nth anomaly in the abnormal work behavior detection list; /(I)A weight value expressed as an mth abnormality in the abnormal work content detection list; and the value of the mth abnormality in the abnormal work content detection list is represented.
The salary accounting unit calculates the amount of the fine on the same day in the working time according to the total abnormal grade;
The calculation formula of the fine amount on the same day is as follows:
；
Wherein, Representing the amount of the fine on the current day; /(I)The abnormal value corresponding to the overall abnormal level k is expressed, and k is a natural number greater than or equal to 0; /(I)Penalty coefficients expressed as overall anomaly level k; /(I)Represented as penalty amount corresponding to the anomaly level k.
Calculating the accounting salary of the staff according to the punishment amount of the current day; the calculation formula of the accounting salary is as follows:
；
wherein ESlaray represents the payroll amount for the employee; BS represents the basic salary of the employee; The amount of fines expressed as t days; PB represents the project prize for the employee.
The salary display unit is used for displaying personal information of staff and salary;
The abnormal feedback unit is used for feeding back error information of the inspection unit by staff.
Compared with the prior art, the invention has the beneficial effects that:
1. The invention provides an abnormal behavior detection model which is used for detecting the abnormality of staff at a working position and at a non-working position; when the employee is not in the working position, comparing the dislocation duration with a set threshold value through calculation, judging the behavior state of the employee, and evaluating the abnormal grade; when an employee is in a working position, the invention adopts the abnormal behavior analysis network to analyze the behaviors of the employee, and the multiscale information extraction module is introduced into the network to more accurately acquire the behavior characteristics of the employee, thereby improving the accuracy of behavior judgment. The abnormal behavior detection model improves the effect of detecting the abnormal behaviors of staff, so that the accuracy of salary accounting is improved through accurate abnormal data.
2. In order to further judge the working state of staff, the invention provides an abnormal content detection model. The abnormal content detection model is used for detecting the abnormality of the working equipment of the staff, and the model is used for detecting the working state and the non-working state of the equipment respectively, so that the staff is prevented from operating from the related working content. The model is proposed to effectively avoid the situations of lazy work, small opening and the like of staff in the working time period; and setting a threshold value in the detection process of the abnormal content detection model to judge, and providing a certain error space for staff. The abnormal content detection model accurately judges from the angle of the working content, and improves the accuracy of salary accounting.
3. The invention provides a staff anomaly detection accounting salary method by combining an anomaly behavior detection model and an anomaly content detection model. The method is different from the traditional salary accounting method, the salary is recalculated from the invalid working time angle of staff, and staff salary can be more accurately checked by the method, so that waste of financial resources of a company is avoided.
4. In the first embodiment, another salary accounting calculation method is designed, and salaries of staff are calculated through effective working time of the staff. As a supplement scheme of the invention, the method can be well combined with the calculation of staff salary in invalid time, achieves the effect of further accurately calculating the staff salary in effective work, can avoid the problem of overlarge estimated error in the salary calculation process, and can timely adjust the salary issuing scheme.
Drawings
FIG. 1 is a flowchart of a human resource management method with intelligent accounting salary according to an embodiment of the present invention;
FIG. 2 is a functional flowchart of an abnormal behavior detection model according to an embodiment of the present invention;
FIG. 3 is a network configuration diagram of an abnormal behavior detection model according to an embodiment of the present invention
FIG. 4 is a block diagram of a multi-scale information extraction module according to an embodiment of the present invention;
FIG. 5 is a functional flowchart of an abnormal content detection model according to an embodiment of the present invention;
FIG. 6 is a block diagram of a human resource management system with intelligent accounting salary according to an embodiment of the present invention;
FIG. 7 is a block diagram of a working state examining unit according to an embodiment of the present invention;
Fig. 8 is a block diagram of a salary statistics unit according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In the staff salary accounting process, most companies today use a fuzzy calculation mode, namely basic salary, performance assessment and other project synthesis. In the basic payroll part, the company calculates by counting the daily working time of staff, but certain disadvantages exist in the mode that the importance of the effective working time within the specified time is ignored in consideration of the relative working time of staff, so that the situation of inaccurate staff payroll calculation occurs.
In order to solve the problems, the invention provides a human resource management method and a human resource management system with intelligent accounting salary. The steps of the specific implementation of the present invention are included in fig. 1.
In order to illustrate that the invention can accurately calculate staff salary, staff working states of technical departments and personnel departments of the company A are adopted to calculate staff salary in the following embodiments.
Example 1
S10, acquiring working time of staff;
Acquiring a first check-in request and a first position of 6 staff of a company technical department at a first moment; the first moment refers to the card punching time before the business operating time specified by the company; recording the first moment as a first sign-in time;
Alternatively, in an embodiment of the present application, the first check-in request is executed by nailing software; in order to achieve the integrity of the scheme, the first sign-in request in this embodiment may also be implemented by face recognition sign-in, fingerprint sign-in, and the like; the first position is that check-in operation can be performed within a radius range of 10 meters of company A;
Further, a second check-in request and a second position of 6 staff of the technical department at a second moment are acquired; the second time refers to the card punching time of the company in a specified working time period; the second check-in request may be implemented by the first check-in request; the second position is also a position where the card punching and working can be performed only within 10 meters of the radius of the company A; and recording the second moment as a second check-in time. Further, working time lengths of 6 staff are calculated according to the first check-in time and the second check-in time respectively.
S20, staff work information in the work time is obtained;
further, referring to fig. 2, the behaviors of 6 employees are detected by adopting an abnormal behavior analysis detection model; the abnormal behavior detection analysis model acquires images of the staff in a working time period by using a camera;
Judging whether the staff is in a working area or not according to the image; when the staff is not in the working area, recording the dislocation time of the staff; and when the camera reacquires the image of the employee returned to the working area, recording the in-place time of the employee.
S30, detecting working behavior abnormality of staff;
Further, calculating the dislocation duration according to the dislocation time and the bit time ; Wherein the saidThe dislocation duration of the employee i at the j-th moment is represented;
Judging the said Whether or not it is greater than a preset first time length threshold/>; Wherein in the present embodiment the following will be describedSet to 25 minutes; when the dislocation time length is greater than the first time length threshold, marking the state of the staff as abnormal, and evaluating an abnormal grade according to the error of the dislocation time length and the first time length threshold; the abnormal grades are divided into six grades, namely G0, G1, G2, G3, G4 and G5; wherein the severity of the anomaly level increases stepwise upward, G0 indicates no anomaly, see table 1.
TABLE 16 personnel out of position in technical sector
The state of dislocation of 6 employees of technical sector at a certain moment in working period is shown in table 1, and it can be seen from the table that different levels of anomalies are generated when the dislocation duration exceeds the first duration threshold. The dislocation duration of employee 2 in the table was 4 minutes and no anomaly was generated.
In this part, calculating the abnormal value of the staff through the dislocation time of the staff, and setting the first time threshold as a judgment standard in the embodiment in order to not make misjudgment on the necessary dislocation situation of the staff; and when the dislocation duration of the staff exceeds the first time threshold, the staff is abnormal, and the abnormality appearing in the staff can be accurately classified by calculating the duration error.
Further, under the condition that the camera shoots that the staff is in a working area, analyzing the current working behavior of the staff through an abnormal behavior analysis network; the structure diagram of the abnormal behavior analysis network is shown in fig. 3. The specific steps of the abnormal behavior analysis on the behavior analysis of 6 staff comprise:
Step one, obtaining images of 6 staff at the ith moment;
Inputting the image into a second convolution layer to obtain a second characteristic diagram of staff;
step three, inputting the second characteristic diagram of the staff into a third rolling layer and a multi-scale information extraction module to obtain a third characteristic diagram of the staff and a first multi-scale characteristic diagram of the staff;
referring to fig. 4, the implementation process of the multi-scale information extraction module includes:
inputting the employee second feature map to a 1×1 convolution kernel to obtain an employee first convolution feature map; inputting the employee convolution feature map to a convolution kernel of 3×3 to obtain a second employee convolution feature map; the staff second convolution feature images are subjected to average pooling to obtain staff feature vectors;
inputting the employee second convolution feature map and the employee feature vector into a feature fusion layer, and outputting the employee first multi-scale feature map;
inputting the employee third feature map to a fourth convolution layer and the multi-scale information extraction module to obtain an employee fourth feature map and an employee second multi-scale feature map;
step five, inputting the fourth characteristic diagram of the staff into a fifth convolution layer to obtain a fifth characteristic diagram of the staff;
Step six, inputting the employee fifth feature map to a pool layer to obtain an employee pooling feature map;
And step seven, splicing the employee first multi-scale feature map, the employee second multi-scale feature map and the employee pooling feature map according to channel dimensions, and outputting the abnormal behavior type by a full-connection layer. Wherein, in the present embodiment, the abnormal behavior types are divided into gesture abnormality and activity track abnormality;
in the process, according to the images shot in the working time, carrying out behavior analysis on 6 staff on a station, wherein the accuracy of identifying the abnormal behaviors of the staff is improved by the abnormal behavior analysis network; in the representation 2, the detection results of the different detection models for 6 staff in technical sector during the current working period are shown.
TABLE 2 comparison of employee anomaly detection models
The experimental results in table 2 show that the accuracy and the calculated amount of the abnormal behavior analysis network designed by the invention are far better than those of ResNet network 152, and the accuracy is 5.1% higher than that of ResNet network, which fully demonstrates the effectiveness of the model for detecting abnormal behaviors by using multi-scale information.
In order to illustrate the detection effect of the abnormal behavior analysis detection model on the working states of the staff in the working time period, the working state detection results of 6 staff of the technical department in the current working day in two hours before work are selected in the table 3.
TABLE 3 detection results of abnormal behavior detection model by technical sector
From the above table 3, it can be seen that the abnormal behavior detection model outputs results for the abnormality of the staff at the work station and at the non-work station, respectively, and the corresponding abnormality durations are given in the table. The abnormal behavior detection model can better detect whether staff work seriously according to accurate abnormal information, and make corresponding penalties according to the abnormal information.
Dividing the behavior abnormality of the staff into dislocation abnormality and in-place abnormality in the abnormal behavior detection model, comparing the dislocation duration with a set first time threshold value through calculation when the staff is not in a working position, judging the behavior state of the staff, and evaluating the abnormality grade; the first time threshold limits the leaving time of the staff, so that the situation that the staff leaves the working position for a long time can be effectively solved, moreover, the staff leaves most of companies frequently, and malicious leaving actions are taken as a state that the staff does invalid work; when an employee is in a working position, the invention adopts an abnormal behavior analysis network to analyze the behavior of the employee image; the network can quickly identify abnormal behavior conditions of staff during work and record behavior duration, effectively detect invalid working time generated by abnormal actions, and enable final salary accounting to be more accurate.
Detecting employee behaviors by adopting the abnormal behavior detection model in the step S30; in order to further monitor whether the staff is in serious work or not more accurately, an abnormal content detection model is provided in the example to synchronously monitor the work content of the staff work equipment; the working device refers to a computer, a mouse and a keyboard in the embodiment.
S40, detecting abnormality of working contents of staff;
referring to fig. 5, the information of the current working equipment of 6 employees is obtained by adopting monitoring software inside a company;
Judging whether the input device of the current working device works or not; when the working equipment judges whether the input equipment of the current working equipment works or not, the non-working time of the equipment is recorded; judging whether the non-working time length is greater than a preset second time length threshold value or not; wherein the second duration threshold is set to 5 minutes;
When the non-working time exceeds 5 minutes, marking the current employee as an abnormal state, and evaluating an abnormal grade according to the error with the second time; when the staff reuses the working equipment within 5 minutes, the staff is in a normal state;
When the current working equipment is in a working state; calculating the correlation matching degree of the information and the working content of the staff; when the correlation matching degree is lower than a preset correlation matching degree threshold value, counting the operation duration of the staff; in the embodiment of the application, the correlation matching threshold is set to be 0.45;
When the correlation matching degree is higher than 0.45, the marking state is normal; when the operation time length is longer than the third threshold time length value, marking the abnormal state, and evaluating the abnormal grade according to the error between the operation time length and the third threshold time length value; when the correlation matching degree is lower than 0.45, comparing the duration of the operation with a third duration threshold; in order to avoid improper operation of staff during office work, performing non-image related work content operation, setting the third time period threshold to be 4 minutes; when the operation time length is longer than 4 minutes, marking the abnormal state, and evaluating the abnormal grade according to the error of the operation time length and the third threshold time length value; marking employee status as normal when the duration of operation is within 4 minutes. The device detection of 6 employees of the technical sector during the working time by the abnormal content detection model is given in table 4.
TABLE 4 abnormal content detection model detection results
Whether the equipment of the table 4 is in the operating state column or not is displayed as not operating by the equipment of two staff, and the corresponding equipment is displayed as a hovering state in the equipment activity state column; the hovering interface refers to that the computer is in a black screen or is in the same interface for a long time, the mouse hovers at the same position for a long time and the keyboard does not input for a long time; the hover state duration of employee 1 in the table is 3 minutes from the second duration threshold error, which is an acceptable time frame for the company, so that employee 1 is in a normal state; the staff 6 can see that the correlation is 20% from the work correlation column, the equipment activity state is in the chat interface, the operation time is 24 minutes, which is far longer than the third duration threshold, and the staff 6 is regarded as abnormal in size; the abnormality level of employee 6 may be determined based on the difference in the operation duration and the third duration threshold.
In order to further calculate the invalid working behavior of the 6 employees in the working period, in the above step S40, it is determined whether the office equipment of the employee performs the operation of the related office content by using the abnormal content detection model. The detection process of the model on office equipment is divided into two states, and a time length threshold value is set by collecting historical data through a technical department; the first state is that when the office equipment is not operated, and when the unoperated time period does not exceed a second time period threshold value, the office equipment is considered to be free of abnormality; when the unoperated duration exceeds a second duration threshold, the office equipment is considered to be abnormal; the second time period threshold value is set to 5 minutes because the office equipment is required to be used for operation in the case of discussion of work content by technical department staff and the like. The second state is that when the office equipment is operated, the operation of the office equipment is divided into irrelevant operation and relevant operation, so that the content of the operation of the office equipment is required to be subjected to relevance matching, and judgment is carried out through a relevance matching threshold; the threshold mainly measures the relevance to the work content, and also belongs to the content relevant to the work when the staff searches the material, so the relevance matching degree threshold is set to be 0.45. The related matching degree calculation is carried out, and meanwhile, the operation time is counted, so that staff is prevented from working with investigation data as a reference. For this purpose, a third time threshold is set in comparison with the operation time, the third time threshold is set to 4 minutes, and investigation of the related content can be completed within a prescribed time according to the attribute of the technical department work content. The scheme can effectively judge the operation behaviors of staff, calculate errors and classify the staff according to the errors.
S50, calculating overall abnormal values, and judging abnormal grades;
further, according to the detection results of the abnormal behavior detection model and the abnormal content detection model on 6 operators of the technical department in the working time period of the same day and at different time points, the total abnormal value of the same day can be calculated, and the abnormal level is judged;
The calculation formula of the total abnormal value is as follows:
；
Wherein, Indicated as overall outliers for employee i during the time of day; /(I)A weight value representing that the abnormal behavior detection model detects an nth abnormality; /(I)An abnormal value representing that the abnormal behavior detection model detects an nth abnormality; /(I)A weight value indicating that the m-th anomaly is detected for the anomaly content detection model; /(I)An abnormal value representing that the abnormal content detection model detects an mth abnormality; n represents the total number of anomalies detected by the abnormal behavior detection model during the working time period of the day of the employee i; m represents the total number of anomalies detected by the anomaly content detection model for employee i during the current day's work period. Referring to Table 5, detailed outliers for employee 1's time of day are shown in Table 5.
TABLE 5 staff 1 abnormality detection during working period
Calculating the overall outlier of employee 1 in Table 5 by the above formula; the abnormal behavior detection model and the abnormal content detection model are different in weight assignment for different abnormal grades; in this embodiment, the weights are assigned as follows:，/>。
；
according to the calculation formula of the total outlier, the total outlier of employee 1 on the same day is 0.903; from the calculated overall anomaly value, the anomaly level of employee 1 may be rated G3.
The calculation method of the staff overall abnormal value is given in the step S50; and comprehensively calculating according to the behavior abnormal data detected in the working time period of the employee 1 and the working equipment abnormal data. As the anomaly data shown in table 4, the anomaly data for each period was recorded, and the anomaly level was rated. In the total abnormal value calculation process, the abnormal of the staff is calculated from multiple aspects, the formula sets an abnormal weight value according to the abnormal of different forms, the abnormal weight value is set to account for different severity degrees corresponding to different abnormal grades, the formula can clearly induce the total abnormal result of the staff in the working days, and staff payroll is accurately calculated according to the total abnormal result.
S60, calculating salary according to the abnormal grade;
Further, calculating the amount of the fine on the same day of 6 staff according to the total abnormal value; the calculation formula of the fine amount of the current day is as follows:
；
Wherein, Representing the value of the fine amount of employee i on the same day; /(I)Penalty coefficients expressed as overall anomaly level k; represented as penalty amount corresponding to the anomaly level k.
The penalty amount when the rating is G3 is calculated by the total abnormal value calculated by the staff 1 in the embodiment;
；
from the calculation result, the penalty amount of employee 1 on the same day is 11.4 yuan.
In this section, total outliers in the working days of 6 employees in the technical department are counted by the calculation method in the above step S50. In order to further calculate the corresponding penalty amount, a corresponding calculation formula is given in S60, which is mainly based on the overall outlier and based on the penalty requirement set by company a, and which can be flexibly varied according to the company penalty specification, aiming at scientifically calculating the penalty of the staff.
Further, the calculation process is to calculate the fine amount of staff 1 in one day, and the salary is calculated as the month end of each month according to the actual salary issuing condition of company A; therefore, the calculation formula of the overall salary accounting of employee 1 is:
；
Wherein, Representing real payroll after staff 1 is checked by payroll monthly; /(I)Denoted as employee 1 basic payroll; /(I)Indicated as employee 1, amount t punishment meted out by God; /(I)Project prizes represented as employee 1; the project prize includes: performance prizes, project quality prizes, and project contribution prizes.
In the embodiment of the application, the working time length related to the staff is obtained by acquiring the working time and the working card punching time of the technical department staff. In order to achieve the purpose of accurately accounting staff salaries by utilizing invalid working time in staff working time, the invention establishes an abnormal behavior detection model and an abnormal content detection model to detect staff behaviors and staff office equipment respectively. The two models can accurately analyze whether the employee generates invalid working time and generate corresponding abnormal values. And carrying out corresponding punishment according to the abnormal conditions corresponding to the staff, and counting the punishment amount per day to realize accurate accounting of staff payouts in the current month.
Alternatively, in the embodiment of the application, the abnormality detection means is adopted to detect the abnormality of staff of technical departments in the working time period, so that salary accounting errors caused by lack of consideration of invalid working time factors in the staff salary accounting process are avoided. In addition, in the method for checking staff salary, the study can be performed by effective working time. For the integrity and accuracy of the salary accounting scheme, an alternative scheme for salary accounting of effective working time is provided in the embodiment, and accurate salary estimation is performed.
Taking 6 staff of a technical department as a research object, calculating actual issuing salaries in an effective working time mode to be expressed as: the calculation formula for actually issuing salaries is as follows:
；
Wherein, Actually issuing payroll for the t day of employee j; /(I)The effective working time of the t day expressed as employee j; /(I)The work cost per unit time on day t, denoted employee j; /(I)Bonus benefits expressed as the time of day t of employee j's effective work; wherein, the bonus benefit refers to project bonus of technical departments, and the method comprises the following steps: performance prizes, project quality prizes, and project contribution prizes;
according to the calculation formula for actually issuing salaries, the current month of staff in the technical department can be calculated to actually issue salaries; the calculating wages for actually issuing wages in the current month are as follows:
/>
Wherein, Representing the amount of payroll actually issued for employee j in the current month; in this example, the current day of the month was set to 24 days for calculation.
Further, to illustrate the effectiveness of the alternative for technical sector staff payroll accounting, here an effect demonstration is made by means of the detection data of technical sector staff 1 and staff 2 in the current month, see table 6.
TABLE 6 staff 1 and staff 2 salary accounting cases
From table 6, it is seen that employee 1 and employee 2 rated the anomaly at the same month as G3 and G2, respectively, and thus the payroll amounts deducted are also different. The effective working time periods of staff 1 and staff 2 differ by 39.6 hours, but the project prize of staff 2 under the project is 0.3 ten thousand yuan more than that of staff 1.
In the embodiment, the effect of staff fund accounting of technical departments is achieved through the two schemes of invalid working time and valid working time, and further, the problem of inaccurate salary in the background technology is effectively solved under interaction of the two schemes, the problem of overlarge estimated error in the salary accounting process can be avoided through the two schemes, and the scheme of salary issuing can be timely adjusted.
Further, in the staff salary accounting process, technology can be performed through two dimensions of invalid working time length and valid working time length, so that the relevant salary settlement departments can easily check the working benefit of staff and accurately calculate the salary of the staff. Furthermore, from the perspective of staff, the salary accounting method provided by the invention can supervise staff to carefully work in a working time period; meanwhile, staff can check wages by inquiring abnormal conditions occurring during work so as to obtain timely feedback when errors are found.
Example two
To illustrate that the invention is applicable to salary accounting for different departments of company a; in the second embodiment, the same salary accounting method is adopted to carry out salary accounting on personnel department staff of the A company; from the learned data, the personnel department of company a has 4 employees, and the main work responsibilities include: recruitment and recording, staff training and development, performance management, salary and benefit management, and the like; from the job responsibilities, it can be seen that the complexity of the personnel work has certain challenges for the practice of the present invention.
According to the first step of the embodiment, firstly, card punching information of personnel department staff is obtained, and working time is recorded;
In this embodiment, when the employee does not perform effective card punching within the time specified by company a, 3 card punching opportunities are provided per month, and if the card punching opportunity of the current month is used up, the prize of full work attendance of the current month is deducted if the card punching opportunity of the current month is performed again.
Acquiring the current day work information of personnel and staff in personnel departments; the work information includes: working state information and working content information;
further, the working state information is input into an abnormal behavior detection model, and an abnormal working behavior detection list of the staff is output;
in the detection of abnormal behaviors of staff of personnel departments, the corresponding threshold value is reset because of the particularity of the working content of the personnel departments;
When the staff is not at the station;
In an embodiment, a first time length threshold related to the dislocation time length can be set to 30-40 minutes; if the employee leaves for a time which does not reach the threshold value setting time, the employee is considered to be in a normal state; if so, evaluating the abnormal grade of the staff state according to the error between the dislocation time length of the staff and the first time length threshold value; the abnormality level adopts the division method in the first embodiment. See table 7 for personnel department employee dislocation anomalies.
TABLE 7 personnel department 4 personnel leaving situation
As can be seen from Table 7, the severity is different for different outlier levels corresponding to different outlier durations. The exception rank ordering case in table 7 is: g5> G2> G0;
When staff is on the station;
The camera on the station is adopted to collect the working image of each employee, and the behavior analysis is carried out on each employee through an abnormal behavior analysis network; wherein the process of analyzing comprises: step one, obtaining images of 4 staff at the ith moment;
Inputting the image into a second convolution layer to obtain a second characteristic diagram of staff;
step three, inputting the second characteristic diagram of the staff into a third rolling layer and a multi-scale information extraction module to obtain a third characteristic diagram of the staff and a first multi-scale characteristic diagram of the staff;
The implementation process of the multi-scale information extraction module comprises the following steps:
inputting the employee second feature map to a 1×1 convolution kernel to obtain an employee first convolution feature map; inputting the employee convolution feature map to a convolution kernel of 3×3 to obtain a second employee convolution feature map; the staff second convolution feature images are subjected to average pooling to obtain staff feature vectors;
inputting the employee second convolution feature map and the employee feature vector into a feature fusion layer, and outputting the employee first multi-scale feature map;
inputting the employee third feature map to a fourth convolution layer and the multi-scale information extraction module to obtain an employee fourth feature map and an employee second multi-scale feature map;
step five, inputting the fourth characteristic diagram of the staff into a fifth convolution layer to obtain a fifth characteristic diagram of the staff;
Step six, inputting the employee fifth feature map to a pool layer to obtain an employee pooling feature map;
And step seven, splicing the employee first multi-scale feature map, the employee second multi-scale feature map and the employee pooling feature map according to channel dimensions, and outputting the abnormal behavior type by a full-connection layer.
In the image detection process of personnel departments, because excessive communication is needed about the working content, the image detection result is further widened in the embodiment of the application, so that the application can be suitable for the working departments with different performances. The results of the behavioral abnormality detection of personnel staff are shown in table 8.
TABLE 8 detection results of personnel department abnormal behavior detection model
/>
Further, detecting staff work content abnormality;
acquiring information of current working equipment of 4 staff;
Judging whether the input device of the current working device works or not; when the working equipment judges whether the input equipment of the current working equipment works or not, the non-working time of the equipment is recorded; judging whether the non-working time length is greater than a preset second time length threshold value or not; wherein the second duration threshold is set to 15 minutes;
When the non-working time exceeds 15 minutes, marking the current employee as an abnormal state, and evaluating an abnormal grade according to the error with the second time; when the staff re-uses the working equipment within 18 minutes, the staff is in a normal state;
When the current working equipment is in a working state; calculating the correlation matching degree of the information and the working content of the staff; when the correlation matching degree is lower than a preset correlation matching degree threshold value, counting the operation duration of the staff; in the embodiment of the application, the correlation matching threshold is set to be 0.75;
When the correlation matching degree is higher than 0.75, the marking state is normal; when the operation time length is longer than the third threshold time length value, marking the abnormal state, and evaluating the abnormal grade according to the error between the operation time length and the third threshold time length value; comparing the duration of the operation with a third duration threshold when the correlation matching degree is lower than 0.75; in order to avoid improper operation of staff during office work, performing non-image related work content operation, setting the third time period threshold to be 2 minutes; when the operation time length is longer than 2 minutes, marking the abnormal state, and evaluating the abnormal grade according to the error of the operation time length and the third threshold time length value; marking employee status as normal when the duration of operation is within 2 minutes.
Further, calculating the overall abnormal value of staff and judging the abnormal level;
further, according to the detection results of the abnormal behavior detection model and the abnormal content detection model on 6 operators of the technical department in the working time period of the same day and at different time points, the total abnormal value of the same day can be calculated, and the abnormal level is judged;
The calculation formula of the total abnormal value is as follows:
；
Wherein, Indicated as overall outliers for employee i during the time of day; /(I)A weight value representing that the abnormal behavior detection model detects an nth abnormality; /(I)An abnormal value representing that the abnormal behavior detection model detects an nth abnormality; /(I)A weight value indicating that the m-th anomaly is detected for the anomaly content detection model; /(I)An abnormal value representing that the abnormal content detection model detects an mth abnormality; n represents the total number of anomalies detected by the abnormal behavior detection model during the working time period of the day of the employee i; m represents the total number of anomalies detected by the anomaly content detection model for employee i during the current day's work period.
Further, accounting payroll according to the abnormal grade;
；
Wherein, Representing the value of the fine amount of employee i on the same day; /(I)Penalty coefficients expressed as overall anomaly level k; represented as penalty amount corresponding to the anomaly level k.
And calculating overall salary accounting according to the punishment, wherein the calculation formula is as follows:
；/>
Wherein, Representing real payroll after staff 1 is checked by payroll monthly; /(I)Denoted as employee 1 basic payroll; /(I)Indicated as employee 1, amount t punishment meted out by God; /(I)Indicated as project prize for employee 1.
In the embodiment, 4 staff of personnel departments are calculated by adopting the accounting salary method provided by the invention; accounting for staff salaries can count the salaries through invalid working time of staff in a working time period; similarly, staff can verify the verified salary result through the method of the invention, and the specific process comprises that staff inquires the salary by using an internal system of the company A, and can acquire the detail of the salary about himself according to the internal system; during salary inquiry, abnormal history records stored in the system can be verified, and when errors exist in the history records, complaints are carried out to higher-level related departments, and salary is recalculated. The method can not only timely feed back wrong abnormal data, but also ensure the accuracy of staff salary accounting.
Example III
Referring to fig. 6, a human resource management system with intelligent accounting salary, the system comprising: the system comprises: the system comprises an employee data management unit, an attendance unit, a working state examination unit, a salary statistics unit, a salary display unit and an abnormal feedback unit;
the staff data management unit is used for storing staff data in a company, wherein the staff data management unit stores staff data of a plurality of departments such as a technical department, a marketing department, a personnel department, a logistics department and the like of the company;
The attendance checking unit is used for recording the daily attendance and attendance card punching conditions of staff;
Referring to fig. 7, the working state examining unit is used for examining whether the staff work seriously in the working time; wherein, the operating condition inspection unit includes: an abnormal behavior detection module and an abnormal work content detection module; the abnormal behavior detection module is used for detecting whether the limb state of the staff is in a working state or not;
The implementation process of the abnormal behavior detection model comprises the following steps: acquiring an office position image of the staff; judging whether the employee is in a working area at the current moment according to the image;
Recording as off-location time when the employee is not in the work area; acquiring time for the staff to return to the working area according to the image and recording the time as in-place time;
Calculating the dislocation duration according to the dislocation time and the in-place time;
Judging whether the dislocation duration is greater than a preset first duration threshold value or not; when the dislocation time length is larger than the first time length threshold value, marking abnormal states, and evaluating abnormal grades according to errors of the dislocation time length and the first time length threshold value; when the dislocation duration is equal to or less than the first time duration threshold, the marking state is normal;
inputting an employee image into an abnormal behavior analysis network when the employee is in the working area; outputting the abnormal behavior type and the abnormal behavior duration of the employee;
The implementation specific steps of the abnormal behavior analysis network comprise:
Step one, inputting the employee image into a first convolution layer to obtain a first characteristic diagram of the employee;
Step two, inputting the employee first feature map into a second convolution layer to obtain an employee second feature map;
step three, inputting the second characteristic diagram of the staff into a third rolling layer and a multi-scale information extraction module to obtain a third characteristic diagram of the staff and a first multi-scale characteristic diagram of the staff;
The implementation process of the multi-scale information extraction module comprises the following steps:
inputting the employee second feature map to a 1×1 convolution kernel to obtain an employee first convolution feature map; inputting the employee convolution feature map to a convolution kernel of 3×3 to obtain a second employee convolution feature map; the staff second convolution feature images are subjected to average pooling to obtain staff feature vectors;
inputting the employee second convolution feature map and the employee feature vector into a feature fusion layer, and outputting the employee first multi-scale feature map;
inputting the employee third feature map to a fourth convolution layer and the multi-scale information extraction module to obtain an employee fourth feature map and an employee second multi-scale feature map;
step five, inputting the fourth characteristic diagram of the staff into a fifth convolution layer to obtain a fifth characteristic diagram of the staff;
Step six, inputting the employee fifth feature map to a pool layer to obtain an employee pooling feature map;
And step seven, splicing the employee first multi-scale feature map, the employee second multi-scale feature map and the employee pooling feature map according to channel dimensions, and outputting the abnormal behavior type by a full-connection layer.
The abnormal grade of the staff is evaluated according to the abnormal behavior type and the abnormal behavior duration to obtain the abnormal working behavior detection list;
The abnormal work content detection module is used for detecting whether the current work equipment of the staff is used for the relevant work state; the implementation process of the abnormal content detection model comprises the following steps:
Acquiring current work content information of the staff;
judging whether the input device of the current working device works or not;
when the input equipment is in an inactive state;
recording the non-working time; whether the non-working time length is greater than a preset second time length threshold value; marking abnormal states when the non-working time length is larger than the second time length threshold value, and evaluating abnormal grades according to errors of the non-working time length threshold value; when the non-working time length is smaller than or equal to the second time length threshold value, the marking state is normal;
when the input equipment is in a working state;
Performing relevance matching on the current working content information and the operation of the input equipment; when the correlation matching degree is higher than a preset correlation matching degree threshold value, the marking state is normal; when the correlation matching degree is lower than the correlation matching degree threshold value, comparing the operation duration with a third duration threshold value;
When the operation time length is greater than the third time length threshold value, marking the abnormal state, and evaluating the abnormal grade according to the error between the operation time length and the third time length threshold value; when the operation duration is less than or equal to the third duration threshold, the marking state is normal;
and obtaining a normal work content detection list according to the content abnormality and the abnormality duration of the work equipment.
Referring to fig. 8, the salary statistics unit is configured to count the employee's payroll; wherein the salary statistics unit includes: a salary calculation module and the salary accounting module;
The salary calculation module is used for calculating salary of the staff based on the salary influence data corresponding to the staff according to the salary calculation rules corresponding to the staff, and the salary accounting unit is used for accounting the salary condition of the staff;
The salary accounting module is used for recalculating the salary calculating module according to the information of the working state examining unit and calculating the amount of the fine on the same day in the working time according to the total abnormal grade; the calculation formula of the fine amount on the same day is as follows: ；
Wherein, Representing the amount of the fine on the current day; /(I)The abnormal value corresponding to the overall abnormal level k is expressed, and k is a natural number greater than or equal to 0; /(I)Penalty coefficients expressed as overall anomaly level k; /(I)A penalty amount corresponding to the abnormal level k;
Calculating the accounting salary of the staff according to the punishment amount of the current day; the calculation formula of the accounting salary is as follows:
；
wherein ESlaray represents the payroll amount for the employee; BS represents the basic salary of the employee; The amount of fines expressed as t days; PB represents the project prize for the employee.
The salary display unit is used for displaying personal information of staff and salary;
The abnormal feedback unit is used for feeding back the abnormality detected by the system and objection of the salary accounting result by staff; in the unit, staff can apply to a financial auditing department for abnormality existence objection by inquiring the abnormal detection item of the working state auditing unit, if the abnormality exists, the error abnormality existing in the state auditing unit is canceled, and salary rechecking is carried out by a salary statistics unit; staff inquires payroll through a payroll display unit.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (10)
1. The human resource management method with intelligent accounting salary is characterized in that: the method comprises the following steps:
acquiring a first sign-in request and a first position of an employee at a first moment; the first position comprises position information in a preset area range;
Recording the first moment as a first sign-in time;
acquiring a second sign-in request and a second position of the employee at a second moment; wherein the second position comprises position information in a preset area range;
Recording the second moment as a second check-in time;
calculating the working time of the staff according to the first check-in time and the second check-in time;
Acquiring the working information of the staff in the unit working time of the working time; wherein, the work information includes: working state information and working content information;
inputting the working state information into an abnormal behavior detection model; outputting an abnormal working behavior detection list;
inputting the work content information into an abnormal content detection model; outputting an abnormal work content detection list;
Calculating overall abnormal values in the working time according to the abnormal working behavior detection list and the abnormal working content detection list, and judging overall abnormal grades;
Calculating the amount of the fine on the same day in the working time according to the total abnormal value;
Calculating the accounting salary of the staff according to the amount of the fine on the same day; the calculation formula of the accounting salary is as follows:
；
wherein ESlaray represents the payroll amount for the employee; BS represents the basic salary of the employee; The amount of fines expressed as t days; PB represents the project prize for the employee.
2. The human resource management method with intelligent accounting salary according to claim 1, wherein the calculation formula of the overall outlier is:
；
Wherein, Representing the total abnormal value of the staff in the working time period; /(I)A weight value expressed as an nth anomaly in the abnormal working behavior detection list; /(I)A value represented as an nth anomaly in the abnormal work behavior detection list; /(I)A weight value expressed as an mth abnormality in the abnormal work content detection list; /(I)And the value of the mth abnormality in the abnormal work content detection list is represented.
3. The human resource management method with intelligent accounting salary according to claim 1, wherein the abnormal behavior detection model is used for detecting whether the limb state of the staff is in a working state; the implementation process of the abnormal behavior detection model comprises the following steps:
Acquiring an office position image of the staff;
Judging whether the employee is in a working area at the current moment according to the image;
Recording as off-location time when the employee is not in the work area; acquiring time for the staff to return to the working area according to the image and recording the time as in-place time;
Calculating the dislocation duration according to the dislocation time and the in-place time;
Judging whether the dislocation duration is greater than a preset first duration threshold value or not; when the dislocation time length is larger than the first time length threshold value, marking abnormal states, and evaluating abnormal grades according to errors of the dislocation time length and the first time length threshold value; when the dislocation duration is equal to or less than the first time duration threshold, the marking state is normal;
inputting an employee image into an abnormal behavior analysis network when the employee is in the working area; outputting the abnormal behavior type and the abnormal behavior duration of the employee;
And obtaining the abnormal working behavior detection list according to the abnormal behavior type and the abnormal behavior duration.
4. The human resource management method with intelligent accounting salary according to claim 3, wherein the implementation specific steps of the abnormal behavior analysis network include:
Step one, inputting the employee image into a first convolution layer to obtain a first characteristic diagram of the employee;
Step two, inputting the employee first feature map into a second convolution layer to obtain an employee second feature map;
step three, inputting the second characteristic diagram of the staff into a third rolling layer and a multi-scale information extraction module to obtain a third characteristic diagram of the staff and a first multi-scale characteristic diagram of the staff;
The implementation process of the multi-scale information extraction module comprises the following steps:
inputting the employee second feature map to a 1×1 convolution kernel to obtain an employee first convolution feature map; inputting the employee convolution feature map to a convolution kernel of 3×3 to obtain a second employee convolution feature map; the staff second convolution feature images are subjected to average pooling to obtain staff feature vectors;
inputting the employee second convolution feature map and the employee feature vector into a feature fusion layer, and outputting the employee first multi-scale feature map;
inputting the employee third feature map to a fourth convolution layer and the multi-scale information extraction module to obtain an employee fourth feature map and an employee second multi-scale feature map;
step five, inputting the fourth characteristic diagram of the staff into a fifth convolution layer to obtain a fifth characteristic diagram of the staff;
Step six, inputting the employee fifth feature map to a pool layer to obtain an employee pooling feature map;
And step seven, splicing the employee first multi-scale feature map, the employee second multi-scale feature map and the employee pooling feature map according to channel dimensions, and outputting the abnormal behavior type by a full-connection layer.
5. The human resources management method with intelligent accounting salary according to claim 1, wherein the abnormal work content detection model is used to detect whether the employee's current work equipment is used for the relevant work status; the implementation process of the abnormal content detection model comprises the following steps:
Acquiring current work content information of the staff;
judging whether the input device of the current working device works or not;
When the input equipment is in an unoperated state, recording unoperated time; whether the non-working time length is greater than a preset second time length threshold value; when the non-working time length is greater than the second time length threshold, marking abnormal states, and evaluating abnormal grades according to errors of the non-working time length and the second time length threshold; when the non-working time length is smaller than or equal to the second time length threshold value, the marking state is normal;
when the input equipment is in a working state;
Performing relevance matching on the current working content information and the operation of the input equipment; when the correlation matching degree is larger than a preset correlation matching degree threshold value, the marking state is normal; when the correlation matching degree is lower than the correlation matching degree threshold value, comparing the operation duration with a third duration threshold value; when the operation time length is greater than the third time length threshold value, marking the abnormal state, and evaluating the abnormal grade according to the error between the operation time length and the third time length threshold value; when the operation duration is less than or equal to the third duration threshold, the marking state is normal;
And obtaining the abnormal working content detection list according to the content abnormality and the abnormality duration of the working equipment.
6. The human resource management method with intelligent accounting payroll as claimed in claim 1, wherein the calculation formula of the fine amount of the day is:
；
Wherein, Representing the amount of the fine on the current day; /(I)The abnormal value corresponding to the overall abnormal level k is expressed, and k is a natural number greater than or equal to 0; /(I)Penalty coefficients expressed as overall anomaly level k; /(I)Represented as penalty amount corresponding to the anomaly level k.
7. A human resource management system with intelligent accounting salary, the system comprising: the system comprises an employee data management unit, an attendance unit, a working state examination unit, a salary statistics unit, a salary display unit and an abnormal feedback unit;
the staff data management unit is used for storing staff data in a company;
The attendance checking unit is used for recording the daily attendance and attendance card punching conditions of staff;
The working state examination unit is used for examining whether the staff works seriously in the working time;
the salary statistics unit is used for counting the salary which the staff should send out;
the salary display unit is used for displaying personal information of staff and salary;
The abnormal feedback unit is used for feeding back error information of the inspection unit by staff.
8. The human resource management system with intelligent accounting salary of claim 7, wherein the operating state review unit includes: an abnormal behavior detection module and an abnormal work content detection module; the abnormal behavior detection module is used for detecting whether the limb state of the staff is in a working state or not; the implementation process of the abnormal behavior detection model comprises the following steps:
Acquiring an office position image of the staff;
Judging whether the employee is in a working area at the current moment according to the image;
Recording as off-location time when the employee is not in the work area; acquiring time for the staff to return to the working area according to the image and recording the time as in-place time;
Calculating the dislocation duration according to the dislocation time and the in-place time;
Judging whether the dislocation duration is greater than a preset first duration threshold value or not; when the dislocation time length is larger than the first time length threshold value, marking abnormal states, and evaluating abnormal grades according to errors of the dislocation time length and the first time length threshold value; when the dislocation duration is equal to or less than the first time duration threshold, the marking state is normal;
inputting an employee image into an abnormal behavior analysis network when the employee is in the working area; outputting the abnormal behavior type and the abnormal behavior duration of the employee;
The abnormal grade of the staff is evaluated according to the abnormal behavior type and the abnormal behavior duration to obtain the abnormal working behavior detection list;
The abnormal work content detection module is used for detecting whether the current work equipment of the staff is used for the relevant work state; the implementation process of the abnormal content detection model comprises the following steps: acquiring current work content information of the staff;
judging whether the input device of the current working device works or not;
when the input equipment is in an inactive state;
recording the non-working time; whether the non-working time length is greater than a preset second time length threshold value; marking abnormal states when the non-working time length is larger than the second time length threshold value, and evaluating abnormal grades according to errors of the non-working time length threshold value; when the non-working time length is smaller than or equal to the second time length threshold value, the marking state is normal;
when the input equipment is in a working state;
performing relevance matching on the current working content information and the operation of the input equipment; when the correlation matching degree is higher than a preset correlation matching degree threshold value, the marking state is normal; when the correlation matching degree is lower than the correlation matching degree threshold value, comparing the operation duration with a third duration threshold value; when the operation time length is greater than the third time length threshold value, marking the abnormal state, and evaluating the abnormal grade according to the error between the operation time length and the third time length threshold value; when the operation duration is less than or equal to the third duration threshold, the marking state is normal;
and obtaining a normal work content detection list according to the content abnormality and the abnormality duration of the work equipment.
9. The human resource management system with intelligent accounting salary of claim 8, wherein the implementation specific steps of the abnormal behavior analysis network include:
Step one, inputting the employee image into a first convolution layer to obtain a first characteristic diagram of the employee;
Step two, inputting the employee first feature map into a second convolution layer to obtain an employee second feature map;
step three, inputting the second characteristic diagram of the staff into a third rolling layer and a multi-scale information extraction module to obtain a third characteristic diagram of the staff and a first multi-scale characteristic diagram of the staff;
The implementation process of the multi-scale information extraction module comprises the following steps:
inputting the employee second feature map to a 1×1 convolution kernel to obtain an employee first convolution feature map; inputting the employee convolution feature map to a convolution kernel of 3×3 to obtain a second employee convolution feature map; the staff second convolution feature images are subjected to average pooling to obtain staff feature vectors;
inputting the employee second convolution feature map and the employee feature vector into a feature fusion layer, and outputting the employee first multi-scale feature map;
inputting the employee third feature map to a fourth convolution layer and the multi-scale information extraction module to obtain an employee fourth feature map and an employee second multi-scale feature map;
step five, inputting the fourth characteristic diagram of the staff into a fifth convolution layer to obtain a fifth characteristic diagram of the staff;
Step six, inputting the employee fifth feature map to a pool layer to obtain an employee pooling feature map;
And step seven, splicing the employee first multi-scale feature map, the employee second multi-scale feature map and the employee pooling feature map according to channel dimensions, and outputting the abnormal behavior type by a full-connection layer.
10. The human resource management system with intelligent accounting payroll of claim 7, wherein the payroll statistics unit comprises: a salary calculation module and the salary accounting module;
The salary calculation module is used for calculating salary of the staff based on the salary influence data corresponding to the staff according to the salary calculation rules corresponding to the staff, and the salary accounting unit is used for accounting the salary condition of the staff;
The salary accounting module is used for recalculating the salary calculating module according to the information of the working state examining unit and calculating the amount of the fine on the same day in the working time according to the total abnormal grade;
The calculation formula of the fine amount on the same day is as follows:
；
Wherein, Representing the amount of the fine on the current day; /(I)The abnormal value corresponding to the overall abnormal level k is expressed, and k is a natural number greater than or equal to 0; /(I)Penalty coefficients expressed as overall anomaly level k; /(I)A penalty amount corresponding to the abnormal level k;
Calculating the accounting salary of the staff according to the punishment amount of the current day; the calculation formula of the accounting salary is as follows:
；
wherein ESlaray represents the payroll amount for the employee; BS represents the basic salary of the employee; The amount of fines expressed as t days; PB represents the project prize for the employee.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410335821.4A CN118195561A (en) | 2024-03-22 | 2024-03-22 | Human resource management method and system with intelligent accounting salary |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410335821.4A CN118195561A (en) | 2024-03-22 | 2024-03-22 | Human resource management method and system with intelligent accounting salary |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118195561A true CN118195561A (en) | 2024-06-14 |
Family
ID=91392731
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410335821.4A Pending CN118195561A (en) | 2024-03-22 | 2024-03-22 | Human resource management method and system with intelligent accounting salary |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118195561A (en) |
-
2024
- 2024-03-22 CN CN202410335821.4A patent/CN118195561A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8073729B2 (en) | Forecasting discovery costs based on interpolation of historic event patterns | |
| Engelmann et al. | The Basel II risk parameters: estimation, validation, and stress testing | |
| EP1361526A1 (en) | Electronic data processing system and method of using an electronic processing system for automatically determining a risk indicator value | |
| CN107993143A (en) | A kind of Credit Risk Assessment method and system | |
| CN110246031A (en) | Appraisal procedure, system, equipment and the storage medium of business standing | |
| US20140324553A1 (en) | Computer-Assisted Method for Adaptive, Risk-Based Monitoring of Clinical Studies | |
| Bedard et al. | Evidential planning in auditing: A review of the empirical research | |
| CN112598184B (en) | Method and device for predicting repeated air suction risk of drug addict | |
| Wahyono et al. | The Effect of CSR Disclosure, Corporate Governance Mechanism, Auditor Independence, Auditor Quality, and Firm Size on Earning Management | |
| CN111090833A (en) | Data processing method, system and related equipment | |
| US8224690B2 (en) | Graphical risk-based performance measurement and benchmarking system and method | |
| CN113626335A (en) | Quality evaluation method and system for public security traffic management application software | |
| Herraiz et al. | Impact of installation counts on perceived quality: A case study on debian | |
| CN118195561A (en) | Human resource management method and system with intelligent accounting salary | |
| CN115375210A (en) | Occupational skill level identification management method and system | |
| Monoarfa et al. | The Influence of Time Budget Pressure and Dysfunctional Behavior to Audit Quality at Bawasda in Gorontalo Province | |
| CN108629506A (en) | Modeling method, device, computer equipment and the storage medium of air control model | |
| Kaymak | Quantifying the signaling role of education | |
| Torresetti et al. | Scaling operational loss data and its systemic risk implications | |
| KR102669893B1 (en) | Automatic calculation system for profit and loss by IT project using smart work log | |
| CN111861725B (en) | Three-party data source cost accounting method and system | |
| TWI809635B (en) | Insurance claims fraud detecting system and method for assessing the risk of insurance claims fraud using the same | |
| Albrecht | Estimation error in income determination | |
| CN117371783A (en) | Risk index positioning method of enterprise risk assessment model based on index monitoring | |
| Pata’dungan et al. | The Effect of Auditor Experience, Task Complexity and Expertise on Audit Judgment at the North Toraja Regency Inspectorate |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |