CN114037288A - Performance adjusting system and method based on machine learning - Google Patents
Performance adjusting system and method based on machine learning Download PDFInfo
- Publication number
- CN114037288A CN114037288A CN202111331877.5A CN202111331877A CN114037288A CN 114037288 A CN114037288 A CN 114037288A CN 202111331877 A CN202111331877 A CN 202111331877A CN 114037288 A CN114037288 A CN 114037288A
- Authority
- CN
- China
- Prior art keywords
- data
- performance
- employee
- machine
- learning
- 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
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000010801 machine learning Methods 0.000 title claims abstract description 36
- 238000011156 evaluation Methods 0.000 claims abstract description 24
- 239000011159 matrix material Substances 0.000 claims description 17
- 238000000605 extraction Methods 0.000 claims description 13
- 238000010276 construction Methods 0.000 claims description 10
- 230000008439 repair process Effects 0.000 claims description 9
- 238000013528 artificial neural network Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 7
- 238000012549 training Methods 0.000 claims description 7
- 230000004913 activation Effects 0.000 claims description 6
- 238000011157 data evaluation Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 5
- UBOXGVDOUJQMTN-UHFFFAOYSA-N 1,1,2-trichloroethane Chemical compound ClCC(Cl)Cl UBOXGVDOUJQMTN-UHFFFAOYSA-N 0.000 claims description 2
- 101000892269 Meleagris gallopavo Beta-1 adrenergic receptor Proteins 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 abstract description 7
- 239000000284 extract Substances 0.000 abstract 2
- 239000010410 layer Substances 0.000 description 27
- 230000002354 daily effect Effects 0.000 description 23
- 238000004140 cleaning Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 238000007726 management method Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000004899 motility Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 239000002356 single layer Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Economics (AREA)
- Medical Informatics (AREA)
- Game Theory and Decision Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a performance adjusting method based on machine learning, which extracts multi-dimensional characteristic information according to historical assessment condition data of employees to be used as sample input of machine learning, and constructs a dynamic adjusting model of an extreme learning machine based on the learning of the historical assessment condition data; acquiring relevant evaluation input data of the performance of the employee at this time, and inputting the data into the dynamic adjustment model of the extreme learning machine; and the extreme learning machine dynamic adjustment model adjusts the performance assessment result of the employee according to the current characteristic information and outputs the adjustment result. The invention can extract the staff performance information to the maximum extent and dynamically adjust the staff performance information, thereby really treating the performance score result fairly and preventing the problem that the data is too rigid according to the formula calculation.
Description
Technical Field
The invention relates to the technical field of employee management, in particular to a performance adjusting system and method based on machine learning.
Background
With the rapid development of economy in China, more and more companies, especially scientific and technical companies, have more and more measurement standards for the working capacity level and the productivity benefit of employees. The staff performance evaluation not only reflects the recognition degree of the company to the staff capability, but also can improve the enthusiasm, the initiative and the motility of the staff work, and combines the short-term income and the medium-term income with the continuous development to rationalize, standardize and systematize the salary management. The performance assessment does not aim at punishing and confining the staff, but is a means for exciting the staff, and increases the fighting capacity and cohesion of the team. A staff performance dynamic assessment method based on big data analysis has patent numbers: CN 110210762A relates to staff performance dynamic assessment method, which adopts data as benchmark, uses formula method to calculate, has no effective deviation correction, has less reference information dimension, only depends on the results of process performance assessment, mutual assessment performance assessment and staff workload assessment, has strict performance data, and can not completely satisfy the fairness and justice principle. With the appearance of machine learning and deep learning in the academic world, the attention and the application range are continuously expanded, the staff performance assessment theory is correspondingly and reasonably dynamically adjusted on the basis of the calculation of an original moving formula, the machine learning can report the working condition through the daily report of the staff at the past period, the daily report is reported and summarized, the rest time, the daily report is reported, the project on-term (namely the profit condition) time, the leadership performance evaluation and the final current-month performance assessment result, and a learning object is formed; after the monthly performance results appear, the dynamic adjustment is carried out through the preposed learning, so that the monthly performance results are fairer and fair, and the condition that the monthly performance results are too high or too low is prevented.
For example: a certain company sets a performance scoring system, and the normal performance appraisal score of the employee is 10 points;
7-9 (none), performance is 80%;
between 9-11 (no), 100% performance;
between 11 and 13 points, the performance is 120 percent;
above 13, performance development 150%;
however, the calculation of a too severe big data formula method inevitably occurs 10.9, and the next gear can not be entered only by the difference of 0.1.
Disclosure of Invention
The invention overcomes the defects of the prior art, provides a performance adjusting system and method based on machine learning, and aims to solve the technical problems that: the technical problem that performance scoring is too severe exists in current company performance management.
In view of the above problems of the prior art, according to one aspect of the present invention, to solve the above technical problems, the present invention adopts the following technical solutions:
a method of machine learning-based performance adjustment, comprising:
extracting multi-dimensional characteristic information according to the historical assessment condition data of the staff, inputting the multi-dimensional characteristic information as a sample for machine learning, and constructing a dynamic adjustment model of the extreme learning machine based on the learning of the historical assessment condition data;
acquiring relevant evaluation input data of the performance of the employee at this time, and inputting the data into the dynamic adjustment model of the extreme learning machine;
and the extreme learning machine dynamic adjustment model adjusts the performance assessment result of the employee according to the current characteristic information and outputs the adjustment result.
In order to better realize the invention, the further technical scheme is as follows:
and further, acquiring the relevant evaluation input data of the performance of the employee through an employee daily reporting system.
Further, the obtained input data of the related evaluation of the performance of the employee at this time comprises employee working condition report data, employee daily report reporting time report data, project time data, leader performance score evaluation data, jira and repair condition data, svn and git code submission condition data, code sonar scanning condition data or employee actual performance condition data in the current month.
Further, the method for extracting the multi-dimensional feature information comprises data evaluation value unification, data cleaning and data general feature extraction.
Further, the data form of the data evaluation value unification is a character string comma interval.
Further, the data cleaning is to realize non-number items in the data in a program programming mode so as to avoid mixing literal or character string information in the data.
Further, the data general feature extraction is performed by adopting a StandardScaler to perform data normalization processing.
Further, the construction process of the extreme learning machine dynamic adjustment model comprises the following steps:
sample data set x for acquiring employee performancej＝[xj1,xj2,...,xjn]T∈Rn，tj＝[tj1,tj2,...,tjm]T∈RmBuilding a regularized extreme learning machine network model, where xjTo input parameters, tjThe result is the performance score of the staff, n is the number of samples, m is the number of sample characteristics, and R is the set of real numbers;
and evaluating the dynamic adjustment model of the extreme learning machine by applying an evaluation index.
Further, the construction process of the extreme learning machine dynamic adjustment model comprises the following steps:
the network parameters of the regularized extreme learning machine network model are as follows: the number of nodes of an input layer, the number of nodes of an output layer and the number of nodes of a hidden layer are counted;
selecting an activation function;
for sample data set xj＝[xj1,xj2,...,xjn]T∈Rn，tj＝[tj1,tj2,...,tjm]T∈RmThe number of nodes of the input layer is TODO, the number of nodes of the output layer is 1, and the network model of the extreme learning machine with the number of nodes of the hidden layer as L is as follows:
∑i＝1Lβig(wi·xj+bi)＝oj,j＝1,2,...,N
wherein g (x) is an activation function, wi＝[wi1,wi2,...,win]T is the input weight of the ith hidden layer unit, biIs the offset of the ith hidden layer cell,βi＝[βi1,βi2,...,βim]t is the output weight of the ith hidden layer unit, wi·xjDenotes wiAnd xjInner product of (d);
the matrix form of the extreme learning machine network model is as follows:
Hβ＝T
wherein H is the output of the hidden node, β is the output weight, and T is the desired output;
obtaining wi, bi and beta i, so that | | | H (wi, bi ^ beta ^ -T | | minw, b, β | | | H (wi, bi) · beta-T |, and the method is used for training a single-hidden-layer neural network, wherein i ═ 1,2, ·, L;
by algorithm
When the input weight wi and the bias bi of the hidden layer are randomly determined, the output matrix H of the hidden layer is uniquely determined, and the training of the single hidden layer neural network is converted into the solving of a linear system: h · β ═ T, and the output weight is determined to be β ^ H + · T; and H + is a generalized inverse matrix of the matrix H to obtain an extreme learning machine network model H +, and the employee performance result is calculated through the generalized inverse matrix H + to obtain a solution through constructing the matrix H, so that the performance is dynamically adjusted.
The present invention may also be a performance adjustment system based on machine learning, including:
the front-end user display module is used for acquiring relevant evaluation input data of the performance of the employee and inputting the data into the dynamic adjustment model of the extreme learning machine;
the dynamic algorithm adjustment module is used for extracting multi-dimensional characteristic information according to the historical assessment condition data of the staff, taking the multi-dimensional characteristic information as sample input of machine learning, and constructing a dynamic extreme learning machine adjustment model based on the learning of the historical assessment condition data;
and the output module is used for adjusting the performance assessment result of the employee according to the current period characteristic information and outputting the adjustment result.
Compared with the prior art, the invention has the following beneficial effects:
the performance adjusting system and method based on machine learning are based on the application of the existing machine learning and deep learning in the academic world, and take the performance assessment theory of employees into consideration to carry out corresponding reasonable dynamic adjustment on the basis of the calculation of the original set of transfer formulas, and an extreme learning machine is used for carrying out dynamic adjustment on the performance data; the method is different from a formula calculation mode, known information is utilized in machine learning, data are processed in a characteristic extraction mode, employee performance information can be extracted to the maximum degree and dynamically adjusted, a performance score result is treated fairly, and the problem that the data are too rigid according to formula calculation is solved.
Drawings
For a clearer explanation of the embodiments or technical solutions in the prior art of the present application, the drawings used in the description of the embodiments or prior art will be briefly described below, it is obvious that the drawings in the following description are only references to some embodiments in the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a diagram of a machine learning-extreme learning method according to one embodiment of the invention.
Fig. 2 is a schematic diagram of training a model dynamically adjusted based on the performance of a machine learning employee according to an embodiment of the present invention.
Fig. 3 is a schematic diagram illustrating a process of adjusting employee performance according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to examples, but the embodiments of the present invention are not limited thereto.
Referring to fig. 1 to 3, a performance adjustment method based on machine learning includes:
extracting multi-dimensional characteristic information according to the historical assessment condition data of the staff, inputting the multi-dimensional characteristic information as a sample for machine learning, and constructing a dynamic adjustment model of the extreme learning machine based on the learning of the historical assessment condition data;
acquiring relevant evaluation input data of the performance of the employee at this time, and inputting the data into the dynamic adjustment model of the extreme learning machine;
and the extreme learning machine dynamic adjustment model adjusts the performance assessment result of the employee according to the current characteristic information and outputs the adjustment result.
Through the technical scheme of this embodiment, to staff's performance evaluation final score through machine learning, through the extraction of multidimension degree characteristic information, carry out machine learning sample input, carry out the performance dynamic adjustment through current period characteristic information and final appraisal result sample, can acquire staff's performance dynamic adjustment result, prevented that the data from calculating too hard according to the formula.
In the above embodiment, it is preferable that the employee performance related evaluation input data is acquired by an employee diary system. The input data may include: reporting the working condition of the staff, reporting the time (overtime time) of the staff in the same month, summarizing the project in-term (and profit condition) time, evaluating the leadership performance score, submitting the jira number and the repair condition, submitting the svn and git codes, scanning the codes sonar and actually performing the staff in the same month, and the like.
The method for extracting the multi-dimensional feature information in the implementation comprises data evaluation value unification, data cleaning and data general feature extraction.
The data evaluation value is unified as follows: reporting the working condition of the staff, summarizing the reporting time (overtime time) of the daily report of the staff in the same month, evaluating the project in term (namely the profit condition), evaluating the leadership performance score, judging the number of jira and the repair condition, submitting the svn and git codes, and replacing the scanning condition of the codes sonar and the actual performance condition of the staff in the same month by adopting a unified data form.
Wherein, staff's working condition reports: and acquiring the current working condition of the staff through a staff daily report system, collecting the progress of the completion condition of each project task, and unifying the data form into a character string comma interval. The shape is as follows: "'0', '1', '10'", where 0-10 indicates work completion, 0 indicates difference, 10 indicates excellent, intermediate data, and so on, each item representing progress of an item.
The reporting time (overtime time) of the daily report of the employee in the current month is summarized: extracting the daily newspaper extraction time of the current working day by an employee daily newspaper system, and calculating the difference value between the current working day and the normal working time, wherein the data form is unified into a character string comma interval, and the character string comma interval is as follows: "'6:30','2:30'".
Project on-term (i.e., profitability condition) time: the time and profit of the current staff in the project are derived through the summary of the staff daily report system, and the data form is unified as a character string comma interval: "'56-1', '32-0', where 56/32 is the day of the project due, 1/0 indicates profit/loss.
And (3) leader performance score evaluation: the score setting is the same as the above embodiment, i.e. 0-10 indicates the degree of completion of the job, 0 indicates the difference, 10 indicates the excellence, the intermediate data and so on, each item represents the progress of a project, and only one item, i.e. '4'.
Number of jira and repair: the data form is unified into a string comma interval as derived by the jira system. The shape is as follows: "'45-39'"; wherein 45 is the total number of jira in the month and 39 is the repaired number of jira.
svn and git code commit case: the data form is unified into string comma intervals, derived either svn or git. The shape is as follows: "'3048', '4509'; where 3048 is the total number of svn code submissions in the month and 4509 is the total number of gits.
Code sonar scan case: the data form is unified into a string comma interval as derived from the sound scan results. The shape is as follows: "'1-34-23-0-0', '0-56-32-0-0'; from left to right are: the quality valve (1 is normal, 0 is other), bugs, the number of leaks, newly added bugs and newly added leaks.
Actual performance condition of the employee in the current month: namely, the final performance result in the current month is used as real output data at the initial design stage of the dynamic adjustment model of the extreme learning machine. The shape is as follows: "'8.2'", where 8.2 represents the performance score for the month.
The data cleaning is to realize non-number items in the data in a program programming mode so as to avoid mixing literal or character string information in the data.
The data general feature extraction can be carried out by adopting a StandardScaler to carry out data normalization processing. The feature extraction is mean removal and variance normalization, and is performed for each feature dimension, not for the sample; the normalization processing method accelerates the speed of solving the optimal solution by gradient descent and has the possibility of improving the precision.
The construction process of the extreme learning machine dynamic adjustment model comprises the following steps:
sample data set x for acquiring employee performancej＝[xj1,xj2,...,xjn]T∈Rn，tj＝[tj1,tj2,...,tjm]T∈RmBuilding a regularized extreme learning machine network model, where xjTo input parameters, tjThe result is the performance score of the staff, n is the number of samples, m is the number of sample characteristics, and R is the set of real numbers;
and evaluating the dynamic adjustment model of the extreme learning machine by applying an evaluation index.
And the construction process of the extreme learning machine dynamic adjustment model further comprises the following steps:
the network parameters of the regularized extreme learning machine network model are as follows: the number of nodes of the input layer, the number of nodes of the output layer and the number of nodes of the hidden layer can be shown in FIG. 1;
selecting an activation function;
for sample data set xj＝[xj1,xj2,...,xjn]T∈Rn，tj＝[tj1,tj2,...,tjm]T∈RmThe number of nodes of the input layer is TODO, the number of nodes of the output layer is 1, and the network model of the extreme learning machine with the number of nodes of the hidden layer as L is as follows:
∑i＝1Lβig(wi·xj+bi)＝oj,j＝1,2,...,N
wherein g (x) is an activation function, wi＝[wi1,wi2,...,win]T is the input weight of the ith hidden layer unit, biIs the bias of the ith hidden layer unit, betai＝[βi1,βi2,...,βim]T is the output weight of the ith hidden layer unit, wi·xjDenotes wiAnd xjInner product of (d);
the matrix form of the extreme learning machine network model is as follows:
Hβ＝T
wherein H is the output of the hidden node, β is the output weight, and T is the desired output;
in order to be able to train a single-hidden-layer neural network, we want to get wi ^ bi ^ and β i ^ so that i | | | H (wi ^ bi ^) β ^ -T | | | | minw, b, β | | H (wi, bi) · β -T |, where i ═ 1,2,. said, L, is equivalent to the minimization loss function;
conventional algorithms based on gradient descent methods can be used to solve such problems, but the gradient-based learning algorithm requires that all parameters be adjusted in an iterative process. In the ELM algorithm, once the weight w is inputiAnd bias of hidden layer biRandomly determined, the output matrix H of the hidden layer is uniquely determined. Training the single-hidden-layer neural network can be converted into solving a linear system: h · β ═ T. And output weights may be determined
β^＝H+·T
Where H + is the Moore-Penrose generalized inverse of matrix H. And the norm of the found solution β ^ can be proven to be the smallest and unique. And obtaining an extreme learning machine network model H +, and obtaining a solution, namely dynamically adjusting the performance, by constructing a matrix H and calculating a generalized inverse matrix H + according to the employee performance result.
A machine learning based performance adjustment system, comprising:
the front-end user display module is used for acquiring relevant evaluation input data of the performance of the employee and inputting the data into the dynamic adjustment model of the extreme learning machine; the input data may include: reporting the working condition of the staff, reporting the time (overtime time) of the staff in the same month, summarizing the project in-term (and profit condition) time, evaluating the leadership performance score, submitting the jira number and the repair condition, submitting the svn and git codes, scanning the code sonar and scanning the actual performance condition of the staff in the same month.
And the algorithm dynamic adjustment module is used for extracting multi-dimensional characteristic information according to the historical assessment condition data of the staff, taking the multi-dimensional characteristic information as sample input of machine learning, and constructing a dynamic adjustment model of the extreme learning machine based on the learning of the historical assessment condition data.
And the output module is used for adjusting the performance assessment result of the employee according to the current period characteristic information and outputting the adjustment result, namely outputting dynamic prediction employee performance data according to the input data.
In the system of the above embodiment, or may be divided into a front-end user presentation end, a background and an algorithm part, specifically:
the front-end user display end relates to an employee daily reporting system, and the employee daily reporting system comprises an employee daily reporting and filling module, an employee performance management module, a project management module and a performance large screen module.
Background: setting an application server and a database server; each module of the user display end is respectively connected through an application server to realize data display and report processing; the application server is connected with the database server. The background contains application services and data storage services.
And an algorithm part: an extreme learning machine algorithm; wherein the extreme learning machine algorithm is a learning algorithm of simple single-layer feedforward neural network (SLFN) machine learning. Theoretically, extreme learning machine algorithms (ELMs) tend to provide good performance (learning speed is extremely fast); the extreme learning machine algorithm is applied to dynamic adjustment of staff performance.
Referring to fig. 1 again, fig. 1 is a schematic diagram of an Extreme Learning Machine (ELM) method, and the initial Extreme Learning Machine is a novel Learning algorithm proposed for single-hidden layer feed-forward neural networks (SLFNs). It randomly selects input weights and analyzes them to determine the output weights of the network. In this theory, such algorithms attempt to provide marginal performance in learning speed. The method has the fundamental effects that the existing complex relationships are known, and the relevance between the relationships is further deepened through data feature extraction. Specifically, as shown in the application of the embodiment shown in fig. 2, fig. 2 and fig. 3 are flowcharts for dynamically adjusting the performance of the employee based on machine learning, and the flow in the diagrams shows that the employee performs daily report filling operation through the employee daily report system every day and stores the daily report in the background server. And the project manager and the leaders at the end of the month evaluate project completion conditions and performance scores of the employees through the conventional employee daily reporting system.
The model building stage in fig. 2, comprises: 1) the daily report system is used for exporting the working condition report of the staff, the daily report reporting time (overtime time) of the staff in the same month is summarized, the time of the project in the same period (and the profit condition) is obtained, the performance score of the staff is evaluated, and the result of the performance appraisal in the same month is obtained. 2) The system comprises a jira system, an svn/git system and a sonar system. And respectively exporting the jira number and the repair condition, the svn and git code submission condition and the code sonar scanning condition. 3) And by describing the employee working condition report, the employee daily report reporting time summarization, the project on-date time, the leader performance score evaluation, the on-month performance evaluation result and the like, then carrying out data value unification, data cleaning and data feature extraction operations, completing the construction of input and output parameters based on a model, and carrying out dynamic model construction through an extreme learning machine.
The model usage phase in fig. 3, comprises: 1) the daily report system is used for exporting the working condition report of the staff, the daily report reporting time (overtime time) of the staff in the same month is summarized, the project in-date time, the leadership performance score evaluation and the performance evaluation result in the same month are obtained. 2) The system comprises a jira system, an svn/git system and a sonar system. And respectively exporting the jira number and the repair condition, the svn and git code submission condition and the code sonar scanning condition. 3) The input parameter construction based on the model is completed by scanning the relevant data of the embodiment, then performing data value unification, data cleaning and data feature extraction, and finally obtaining the dynamic adjustment performance result of the employee in the month through the construction model obtained in the step 2.
In conclusion, the method is different from the other existing patents in that the performance calculation is carried out in the form of a strong formula and few reference items; creatively proposes to adopt a machine learning mode and utilize multi-dimensional data to acquire performance, thereby really achieving fair and fair performance. In the model training parameters, the multidimensional data include: reporting the working condition of the staff, reporting the time (overtime time) of the staff in the same month, summarizing the project time (namely the profit condition), evaluating the leadership performance score, submitting the jira number and the repair condition, submitting the svn and git codes, scanning the code sonar and scanning the actual performance condition of the staff in the same month. And unifying the multidimensional data through data evaluation, cleaning the data and extracting the data characteristics.
The embodiments are described in a progressive manner in the specification, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
Reference throughout this specification to "one embodiment," "another embodiment," "an embodiment," etc., means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment described generally in this application. The appearances of the same phrase in various places in the specification are not necessarily all referring to the same embodiment. Further, when a particular feature, structure, or characteristic is described in connection with any embodiment, it is submitted that it is within the scope of the invention to effect such feature, structure, or characteristic in connection with other embodiments.
Although the invention has been described herein with reference to a number of illustrative embodiments thereof, it should be understood that numerous other modifications and embodiments can be devised by those skilled in the art that will fall within the spirit and scope of the principles of this disclosure. More specifically, various variations and modifications are possible in the component parts and/or arrangements of the subject combination arrangement within the scope of the disclosure, the drawings and the appended claims. In addition to variations and modifications in the component parts and/or arrangements, other uses will also be apparent to those skilled in the art.
Claims (10)
1. A performance adjustment method based on machine learning, characterized by comprising:
extracting multi-dimensional characteristic information according to the historical assessment condition data of the staff, inputting the multi-dimensional characteristic information as a sample for machine learning, and constructing a dynamic adjustment model of the extreme learning machine based on the learning of the historical assessment condition data;
acquiring relevant evaluation input data of the performance of the employee at this time, and inputting the data into the dynamic adjustment model of the extreme learning machine;
and the extreme learning machine dynamic adjustment model adjusts the performance assessment result of the employee according to the current characteristic information and outputs the adjustment result.
2. The machine learning-based performance adjustment method of claim 1, wherein the employee performance-related evaluation input data is obtained through an employee diary system.
3. The performance adjustment method based on machine learning of claim 1, characterized in that the obtained input data related to the performance of the employee at this time includes employee work condition report data, employee daily report reporting time report data, project time data, lead performance score evaluation data, jira and repair condition data, svn and git code submission condition data, code sonar scanning condition data or employee actual performance condition data in the same month.
4. The method of adjusting performance based on machine learning of claim 3, wherein the method of extracting the multi-dimensional feature information includes data evaluation value unification, data cleansing and data general feature extraction.
5. The method of adjusting performance based on machine learning of claim 4, wherein the data form of the data total unification is a character string comma interval.
6. The method of claim 4, wherein the data cleansing is implemented by programming non-numeric items in the data to avoid mixing text or character string information in the data.
7. The method of adjusting performance based on machine learning of claim 4, wherein the data generic feature extraction is processed by data normalization using StandardScaler.
8. The performance adjustment method based on machine learning according to claim 1, wherein the construction process of the extreme learning machine dynamic adjustment model comprises:
sample data set x for acquiring employee performancej＝[xj1,xj2,...,xjn]T∈Rn，tj＝[tj1,tj2,...,tjm]T∈RmBuilding a regularized extreme learning machine network model, where xjTo input parameters, tjThe result is the performance score of the staff, n is the number of samples, m is the number of sample characteristics, and R is the set of real numbers;
and evaluating the dynamic adjustment model of the extreme learning machine by applying an evaluation index.
9. The performance adjustment method based on machine learning according to claim 8, wherein the construction process of the extreme learning machine dynamic adjustment model comprises:
the network parameters of the regularized extreme learning machine network model are as follows: the number of nodes of an input layer, the number of nodes of an output layer and the number of nodes of a hidden layer are counted;
selecting an activation function;
for sample data set xj＝[xj1,xj2,...,xjn]T∈Rn，tj＝[tj1,tj2,...,tjm]T∈RmThe number of nodes of the input layer is TODO, the number of nodes of the output layer is 1, and the network model of the extreme learning machine with the number of nodes of the hidden layer as L is as follows:
∑i＝1Lβig(wi·xj+bi)＝oj,j＝1,2,...,N
wherein g (x) is an activation function, wi＝[wi1,wi2,...,win]T is the input weight of the ith hidden layer unit, biIs the bias of the ith hidden layer unit, betai＝[βi1,βi2,...,βim]T is the output weight of the ith hidden layer unit, wi·xjDenotes wiAnd xjInner product of (d);
the matrix form of the extreme learning machine network model is as follows:
Hβ＝T
wherein H is the output of the hidden node, β is the output weight, and T is the desired output;
obtaining wi, bi and beta i, so that | | | H (wi, bi ^ beta ^ -T | | minw, b, β | | | H (wi, bi) · beta-T |, and the method is used for training a single-hidden-layer neural network, wherein i ═ 1,2, ·, L;
by algorithm
When the input weight wi and the bias bi of the hidden layer are randomly determined, the output matrix H of the hidden layer is uniquely determined, and the training of the single hidden layer neural network is converted into the solving of a linear system: h · β ═ T, and the output weight is determined to be β ^ H + · T; and H + is a generalized inverse matrix of the matrix H to obtain an extreme learning machine network model H +, and the employee performance result is calculated through the generalized inverse matrix H + to obtain a solution through constructing the matrix H, so that the performance is dynamically adjusted.
10. A system for adjusting performance based on machine learning, comprising:
the front-end user display module is used for acquiring relevant evaluation input data of the performance of the employee and inputting the data into the dynamic adjustment model of the extreme learning machine;
the dynamic algorithm adjustment module is used for extracting multi-dimensional characteristic information according to the historical assessment condition data of the staff, taking the multi-dimensional characteristic information as sample input of machine learning, and constructing a dynamic extreme learning machine adjustment model based on the learning of the historical assessment condition data;
and the output module is used for adjusting the performance assessment result of the employee according to the current period characteristic information and outputting the adjustment result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111331877.5A CN114037288A (en) | 2021-11-11 | 2021-11-11 | Performance adjusting system and method based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111331877.5A CN114037288A (en) | 2021-11-11 | 2021-11-11 | Performance adjusting system and method based on machine learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114037288A true CN114037288A (en) | 2022-02-11 |
Family
ID=80137272
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111331877.5A Pending CN114037288A (en) | 2021-11-11 | 2021-11-11 | Performance adjusting system and method based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114037288A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115660508A (en) * | 2022-12-13 | 2023-01-31 | 湖南三湘银行股份有限公司 | Staff performance assessment and evaluation method based on BP neural network |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102129627A (en) * | 2011-03-24 | 2011-07-20 | 江苏电力信息技术有限公司 | Performance appraisal method based on weight adjustment |
| CN109582681A (en) * | 2018-10-26 | 2019-04-05 | 平安科技（深圳）有限公司 | Business data processing method, device, computer equipment and storage medium |
| US20190295012A1 (en) * | 2018-03-23 | 2019-09-26 | International Business Machines Corporation | Predicting employee performance metrics |
| CN110428128A (en) * | 2019-06-21 | 2019-11-08 | 泰康保险集团股份有限公司 | Performance data acquisition methods, device and storage medium |
| CN110675017A (en) * | 2019-08-13 | 2020-01-10 | 平安科技（深圳）有限公司 | Performance evaluation method and device based on artificial intelligence |
| CN111429090A (en) * | 2020-03-19 | 2020-07-17 | 李新华 | Salary management system and method based on artificial intelligence platform |
| CN111639861A (en) * | 2020-06-01 | 2020-09-08 | 上海大学 | Performance assessment method and system based on neural network |
| CN111754111A (en) * | 2020-06-24 | 2020-10-09 | 北京世纪好未来教育科技有限公司 | Office performance evaluation method, apparatus and computer storage medium |
| CN111881193A (en) * | 2020-06-12 | 2020-11-03 | 福建亿能达信息技术股份有限公司 | Performance scheme measuring and calculating system, device and medium based on machine learning |
| CN113159596A (en) * | 2021-04-28 | 2021-07-23 | 武汉增讯科技有限公司 | Big data configuration tuning system, method, device and medium based on deep learning |
-
2021
- 2021-11-11 CN CN202111331877.5A patent/CN114037288A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102129627A (en) * | 2011-03-24 | 2011-07-20 | 江苏电力信息技术有限公司 | Performance appraisal method based on weight adjustment |
| US20190295012A1 (en) * | 2018-03-23 | 2019-09-26 | International Business Machines Corporation | Predicting employee performance metrics |
| CN109582681A (en) * | 2018-10-26 | 2019-04-05 | 平安科技（深圳）有限公司 | Business data processing method, device, computer equipment and storage medium |
| CN110428128A (en) * | 2019-06-21 | 2019-11-08 | 泰康保险集团股份有限公司 | Performance data acquisition methods, device and storage medium |
| CN110675017A (en) * | 2019-08-13 | 2020-01-10 | 平安科技（深圳）有限公司 | Performance evaluation method and device based on artificial intelligence |
| CN111429090A (en) * | 2020-03-19 | 2020-07-17 | 李新华 | Salary management system and method based on artificial intelligence platform |
| CN111639861A (en) * | 2020-06-01 | 2020-09-08 | 上海大学 | Performance assessment method and system based on neural network |
| CN111881193A (en) * | 2020-06-12 | 2020-11-03 | 福建亿能达信息技术股份有限公司 | Performance scheme measuring and calculating system, device and medium based on machine learning |
| CN111754111A (en) * | 2020-06-24 | 2020-10-09 | 北京世纪好未来教育科技有限公司 | Office performance evaluation method, apparatus and computer storage medium |
| CN113159596A (en) * | 2021-04-28 | 2021-07-23 | 武汉增讯科技有限公司 | Big data configuration tuning system, method, device and medium based on deep learning |
Non-Patent Citations (2)
| Title |
|---|
| 李庆喜: "极限学习机(extreme learning machine)概述", 《HTTPS://WWW.CNBLOGS.COM/LIXUEBIN/P/10814857.HTML》 * |
| 范瑛: "基于ELM智能神经网络的公安民警绩效评估探究", 《开封教育学院学报》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115660508A (en) * | 2022-12-13 | 2023-01-31 | 湖南三湘银行股份有限公司 | Staff performance assessment and evaluation method based on BP neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8170841B2 (en) | Predictive model validation | |
| US7725300B2 (en) | Target profiling in predictive modeling | |
| US7730003B2 (en) | Predictive model augmentation by variable transformation | |
| US7499897B2 (en) | Predictive model variable management | |
| US8165853B2 (en) | Dimension reduction in predictive model development | |
| US7562058B2 (en) | Predictive model management using a re-entrant process | |
| US20050234688A1 (en) | Predictive model generation | |
| US20100010878A1 (en) | Predictive model development | |
| Vohra | Intelligent decision support systems for admission management in higher education institutes | |
| Hill et al. | Building an effective representation for dynamic networks | |
| CN107203822A (en) | Method and system based on the Logistic security customers attrition predictions returned | |
| Tuhaienko et al. | EFFECTIVE APPLICATION OF KNOWLEDGE MANAGEMENT IN CURRENT CRISYS CONDITIONS | |
| CN114037288A (en) | Performance adjusting system and method based on machine learning | |
| CN117522607A (en) | Enterprise financial management system | |
| CN117522132A (en) | Vendor risk assessment system and application method | |
| Ersoz et al. | Knowledge discovery and data mining techniques in textile industry | |
| Aikhuele et al. | Development of a TQM-based framework for product infant failure assessment | |
| Zhang | Research of boosting algorithm machine learning in logistics enterprise financial risk prediction | |
| Anggawidjaja et al. | WORKFORCE GROUPING IN COMPLETING PROJECTS WITH INTERN WORK ACTIVITY LOG DATA | |
| Lei et al. | Application of fuzzy association rules in the analysis on higher vocational college students’ performance | |
| Koç et al. | Fractional regression model for investigating the determinants of the unemployment rates in OECD countries | |
| Sterlin et al. | A study on lean management system indicators and its impact to waste minimization | |
| Li et al. | A grey method of prioritizing engineering characteristics in QFD | |
| Ahuja et al. | Business Intelligence in Management-Strategic Approach | |
| Sandefur | Explaining the Trend toward Informal Employment in Africa: Evidence from Ghanaian Manufacturing |
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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Liu Xiaojiang Inventor after: Ding Jicun Inventor after: Li Ning Inventor after: Xue Lingxiang Inventor after: Zhao Pengxiang Inventor after: Cao Zheng Inventor before: Li Ning Inventor before: Xue Lingxiang Inventor before: Zhao Pengxiang Inventor before: Cao Zheng |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20220211 |