Application of Interactive information system in College Personnel Management by using BP Neural Network Algorithm

The BP neural network algorithm has �exible modeling ability and data parallel processing ability. While processing data, it learns through self-learning and adaptive algorithms and associative functions. In the calculation mode of the bp neural network algorithm, the paper conducts forward conduction on the data signal, and conducts reverse conduction according to the obtained error value. The process of transferring the BP neural network algorithm content involved in the data from the input layer to the output layer under the data pattern formed by the forward guidance. And according to the analysis of BP neuron model data of three layers and above, the algorithm form of input layer and output layer is readjust, and the interval data is obtained and analyzed and summarized, so as to trigger the event research on the application of personnel management in colleges and universities. From the problem of university personnel management thoughts, to system �ow numerical analysis as a starting point, starting from the present situation, the form of the personnel management of colleges and universities, this paper analyzed the characteristics of the personnel management at this stage, and the personnel management of colleges and universities ever form the difference, and according to the principle of summed up the characteristics of modern personnel management model, and analyzing the theory of knowledge, From the concept, logic and physical structure of data analysis and design, in the personnel management system of colleges and universities form an important direction of development.


Introduction
Personnel management has become an important branch of management, gradually formed with the needs of management, and growing stronger with the development of management.In the early stage of the 20th century, many companies in China carried out a revolution in personnel management with the help of the wave of national reform, brought the "people-oriented" ideological wave into personnel management work, and gradually formed their own theories and systems, and innovated on the original theoretical basis to form a new management method [1][2].The development of the personnel management industry has gradually become a core part of the eld of management, which is constantly affecting the enthusiasm and creativity of employees [3].The development of the computer industry and the development of more scienti c elds have made management more perceptual.The rigid personnel management process has gradually become closer to the life of employees, and the more modern model has further developed personnel management [4].Under the wave of enterprise reform in our country, the personnel management model of colleges and universities has not changed, and it still continues the traditional management method of sticking to administrative orders [5].Under the reform and development of the college education industry, the content and form of personnel management in colleges and universities should conform to The Times, taking the content, connotation, core and form of work as the primary object of change, which is the development direction of personnel management in colleges and universities at the present stage, connotation, system and di culty, which has become the current stage of college personnel management focus [6][7].

Related Work
It has been proposed in the literature that the way of information transmission between neurons in the human brain is based on the evolution of arti cial neural network model.The similarities between it and the human brain are mainly re ected in two aspects: rst, they both acquire knowledge and learn from the external environment; second, each neuron can store knowledge internally, and then send it to the next neuron signal accordingly [8].In terms of algorithms, the literature has proposed a variety of algorithm models such as the neural network model based on genetic algorithm, the neural network model based on particle swarm algorithm, and the convolutional neural network model in recent years [9].The neural network model based on the genetic algorithm converts the problem domain into a string set (chromosome) through coding, and through the behaviors of replication, crossover, mutation, etc. in genetics, combined with the neural network model, the algorithm model has a good global search ability [10].The literature proposes to optimize the weight threshold and structure adjustment of neural network under the particle swarm optimization algorithm, and then iteratively update each parameter to obtain the steady state of the neural network model.This algorithm can speed up the neural network model to achieve steady state speed [11].The literature puts forward the function of the annual assessment of college teachers: personnel management is particularly important for the assessment of teachers in colleges and universities.It is not only a summary of the past year, but also an improvement for the future year [12].Adding annual assessment to the personnel management system of colleges and universities, and quantifying this part, can re ect the professional ability and level of college teachers in terms of moral education, intellectual education, and professional level [13].The literature suggests that the function of college recruitment is also one of the highlights of this system.College recruitment is an important means of delivering fresh blood to colleges and universities.Through this method, colleges and universities can absorb advanced teaching ideas and models in modern education construction [14].The recruitment function of colleges and universities is designed in strict accordance with the recruitment plan of the provincial organs, and every step is transparent and open, re ecting fairness, openness and justice [15].The literature proposes that the basic pension management of college teachers is a missing module in other college personnel management systems.This module is one of the characteristics of this system [16].This module is designed in combination with the basic pension calculation method of enterprise employees in Guizhou Province.Payment of endowment insurance funds, management of pension overdue nes, management of pension distribution, funeral subsidies for college teachers, etc.

The theoretical basis of bp neural network algorithm
One of the most widely used BP neural network algorithms is a forward neural network model, which integrates the results of the neural network model algorithm according to the input and output of data.Bp neural network algorithm is composed of more than multi-level neurons, each layer of neurons is independent, and only the upper and lower layers are connected with each other.In the application of the three-layer BP neural network algorithm model, the function form of the neurons formed in the hidden layer and output layer is the excitation function, also known as Sigmoid function.This function can be classi ed into log-sigmoid function and tan-sigmoid function.

BP neural network algorithm ow
The hidden layer of neurons in the bp neural network algorithm only includes one layer of neurons.The bp neural network model with one layer is used as an example to illustrate the process of the bp neural network algorithm.
(1) Con rm the number of neurons in the input layer, hidden layer and output layer.Where the number of neurons in the input layer represents the input parameters.The number of neurons in the output layer represents the output parameters.However, the number of hidden layer neurons is not available, and this number plays a key role.At the present stage, the maximum value can be estimated from the formula for the recognition of the number of neurons in the hidden layer, and the value can be substituted into the following equation to con rm the nal value, as shown in Eq. ( 1).

1
(2) Data preprocessing.Data preprocessing includes input vector, output vector determination and data normalization.The formula is as follows: (3) For the calculation of the output quantity of the hidden layer, the weighted quantity sum of the input layer is obtained rst, and the output layer is calculated by using the function.The formula is as follows: In the formula: Ki represents the output value of the ith hidden layer neuron; w i (i represents the weight of the kth input neuron to the ith hidden layer neuron, X k represents the input of the kth neuron in the input layer; w i0 represents the threshold of the ith neuron in the hidden layer. (4) The output quantity of the output layer is calculated, and the number of neurons in the output layer is obtained rst.The calculation formula is as follows: 6 (5) Calculation error.According to the error comparison between the output of the bp neural network model and the actual output value, the error value formula is calculated as follows: 7 (6) Adjust the weights between neurons and the thresholds of neurons in the opposite direction.The standard bp neural network algorithm adopts the method of gradient descent of the error function to learn.The partial derivatives of the error E to the weights and thresholds need to be calculated: 8 9 10

Problems existing in bp neural network algorithm
The traditional bp neural network algorithm is characterized by simple calculation, strong execution ability, relatively small amount of calculation, and can realize the calculation mode at the same time.
The fast descent method under nonlinear programming used by the bp neural network algorithm in the traditional algorithm can correct the weight value according to the value of the error function, so it is easy to form the following problems: The learning e ciency is reduced and the convergence speed is slowed down.The e ciency of learning is xed, so the speed of converging data will change relatively, and the time required will increase accordingly.On some complex problems, the bp neural network algorithm still takes a long time to analyze, because the size of the learning rate changes the calculation time.In the traditional bp neural network algorithm, the learning rate directly re ects the speed of convergence.If the learning rate becomes larger, it may cause the information of the network to change and break up the data form.On the contrary, if the learning rate becomes smaller, it will cause convergence.Slow down.It is easy to fall into the state of local minimum value.In the traditional bp neural network algorithm, the error derivative is used to calculate the value.In the form of sample input, the overall deviation from the larger value and some local minima are formed, and there are many minima points, so in this form the traditional bp neural network algorithm is easy to fall into the local minimum point and cannot ensure convergence to the global minimum point value.For the defect values generated in this form, deepen and improve while the learning e ciency and convergence speed are reduced at the same time.

BP neural network algorithm improvement
Initialize each weight and threshold, and set each weight and threshold to the random minimum value in the range of [-1.0 ~ 1.0], which is convenient for the subsequent changes of the algorithm.
Select the input sample and the expected sample, the input vector and the expected output vector from the samples.
Enter the hidden layer, calculate the output vector of each layer, and then calculate the actual output vector.
Adjust the learning rate: 16 Calculate the inverse transfer error of the output layer: 17

The improvement effect of BP neural network algorithm
The standard probability under the application of the traditional BP neural network algorithm and the improved BP neural network algorithm is analyzed and compared, and the actual output quantity and the expected output quantity are compared.The comparison results are shown in Table 1: It can be seen from Table 1 that under the accuracy comparison between the algorithms, the smaller the number of applications of the improved algorithm, the smaller the difference between the improvements, but there are still error values.When the number of applications increases, it can be found that the improved error value of the same type is reduced to a minimum value state, which ensures the accuracy of the nal algorithm data.From the conclusion, it can be concluded that under the improvement standard, the accuracy of the algorithm will gradually increase with the increase of the number of applications.
In this paper, the e ciency obtained by the improved BP neural network algorithm and the improved BP neural network algorithm is compared, and the comparison results are shown in Table 2.In the comparison of algorithm e ciency, it is found that under the same conditions, the numerical conclusions drawn from the algorithm e ciency for many times are used as comparison values.The number of times that the result is calculated rst is much higher than that of the bp neural network algorithm in the traditional form.Therefore, it is concluded that the improvement has advantages in algorithm e ciency.

System requirements analysis
The personnel management work in colleges and universities is still carried out by the combination of computer and manual management.The frequency of computer use is relatively high, but it is only used for le transmission and printing, and the informationization of personnel management in the form of computer has not been realized.Therefore, any employee cannot implement employee-to-employee cooperation and sharing within the scope of his own work.Establish a data information library, from the traditional single-line connection form to the form of shared data, making the data completer and more comprehensive.To allow business exchanges between various departments, employees can easily and quickly search for data through their own data accounts.Before the data sharing function was not achieved, many data were transmitted by copying, USB ash drives, etc., which greatly affected work.e ciency.Therefore, a university personnel management system that is convenient to improve work e ciency has been developed.The research shows that the main needs of users are as follows: (1) The system manages and establishes the personnel les transferred in and out of employees, which should include basic information such as the entry time of college teachers, colleges, professional titles, and administrative positions.
(2) The system needs to have the function of storing data and have a rationalized information structure, which can ensure the integrity and security of the information searched, and is more conducive to the sharing of data by employees.
(3) The system should be able to increase the data information of personnel management, making it more convenient to delete, add and search.
(4) The con dentiality of personnel system les should be strengthened, and functions such as identity veri cation should be used.Only when the authority is successfully obtained can the information be accessed.

System architecture design
The hardware of this system supports web server, application server, database server, router and related PC.The speci c hardware architecture diagram is shown in Fig. 2.
The personnel management system of colleges and universities uses a three-layer browser architecture.The rst layer is the interaction layer, which is mainly for human-computer interaction; the second layer is the business processing layer, which mainly completes system-related functional operations; the third layer is the data server, which is used to store and manage data.Each layer does not interfere with each other.The user sends an invitation to search for data through the browser, and displays the execution result.The structure diagram is shown in Fig. 3.

System function design
The personnel management system of colleges and universities should establish the following functions: maintain and build a data information database, realize the function of importing and exporting data, and gradually save the teacher information data in colleges and universities with changes, and there are traces to follow.At the same time, the functions of summarizing and searching required for personnel management can be exported and printed, realizing the form of template, compiling information according to template data, and inputting and outputting data according to requirements.A dictionary management module is established to facilitate personnel management personnel to maintain and upgrade the system itself.In the design of information security rights of the database, for the protection of information, in addition to the authorization in the form of ID card information veri cation, log management and other forms are also added to record the information of logged-in visitors.In addition, the system realizes the functions of automatic data backup and one-key recovery.The information classi cation of the personnel management system mainly includes basic information management, personnel management, establishment management, database management, system management and other main functions.

System security design
The subsequent adjustment of the system also has special requirements, including the user name and password on the login content.When the content is entered completely, the system can be o cially entered, and the error message will be logged in again.After the login is completed, the system will perform administrator rights according to the set template, use the menu form as the main guide, and perform system operations within the speci ed range.After the operation is completed, exit the system through the menu button.The system is roughly divided into ve modules.The buttons of the ve modules are used to enter the window.The staff can apply and select data according to different prompts.The single button can enter the next module operate.
System security refers to the application and management of the system and the identi cation of the source of danger, and effective measures are taken according to the source of danger to minimize the danger, so that the data system can achieve the maximum degree of security in terms of performance, cost, and time.The basic principle of system safety is to reconstruct the phased safety issues of the system and formulate a plan that conforms to safe work procedures.The system is designed to cultivate the symmetry and management of personnel information in colleges and universities, and the data itself has safety performance considerations.In order to ensure the further security of data, different rights level management settings are adopted for the personnel entering the system, and the basic authority is the second-level authority.The rst-level authority is the authority of the administrator.The operator with the administrator authority regularly manages the maintenance of the system, and can add, delete and modify data at any time.The second-level authority is user authority.Operators with user authority can preview and search information, except for the way of maintaining and building modules.For the system designed this time, three levels of user authority are established.

System development environment design
The and design of the system a way to maintain the program of the system on the basis of the superposition of software and hardware, which is called the software development environment.
The software development environment is formed under the combination of environment and development tools.

System test results
The test is shown in Table 3.

Function complete Yes
After the personnel management system has passed the functional test, its operation results are normal, the interface of the system is beautiful and generous, and the user operation is convenient and fast, which enables the personnel management personnel to complete the work better and improve the work e ciency.This system has fast access to data and strong practical ability.Personnel management personnel can assign management rights according to different needs, which facilitates the use and management of the system, and also increases the security performance of the system.The system also has functions such as statistical data and search data, which greatly meets the needs of employees in various departments.

Calculation of personnel assessment in colleges and universities
The annual assessment of college teachers is a comprehensive assessment of teachers' conduct, attitude, ability, performance quality and other aspects in the work process.In the traditional form of assessment, there are many drawbacks, such as relying on relationships to attract votes, relying on leadership in uence and other unfavorable factors that lead to unfair assessment.Whether it is the traditional form of voting or the nomination of leaders, it shows the omissions in the management of personnel in colleges and universities.The assessment itself is a means and form of motivating work.The results of the assessment are linked to the teacher's professional title level, which makes the assessment itself bene cial.There is no way to motivate teachers to work hard and make progress, and at the same time show their ability and quality.
Through the quantitative assessment mode, the formation of disadvantages can be reduced.The method is as follows: the assessment is divided into daily performance and work performance.Daily performance includes: moral education, ability, attendance; Work performance includes: service attitude, service quality, work e ciency, completion of tasks, policy level.

Application advantages the system in the personnel management
In the personnel management department, all personnel are assigned to the same oor, with a distance of tens to hundreds of meters, and there are special computers between departments.Energy, can greatly improve the e ciency of information transmission, and can ensure the consistency of information.In the process of transmitting information, each department constantly screens the information, making the format of the information more and more uni ed.If uni ed information is entered into a speci c shared management library, the time for delivery can be omitted, and the workload can be saved under uni ed collaboration.
In this mode, employees in various forms such as the application process and the post exchange process are all open and transparent, forming a sunshine work mode, which can avoid the consequences of black box operations.It is more liberalized in the mode of establishment and management, and it can give full play to what it is good at, and establish a mechanism, so that teaching, scienti c research, management, teaching assistance and other jobs can ow with each other, effectively assist and communicate, and return to work after the completion of the task.Inquiries and discussions between various departments can be carried out.For the system of separation of teaching and scienti c research, the title of teachers is established, so that more people are good at research and research.
As a result, the system under the new operating model can be deeply liked by personnel managers and teachers in colleges and universities, which further deepens and adjusts the system reform, plays a positive role, and is more practical and feasible.

Conclusion
When the autoregressive moving average model of the bp neural network algorithm for prediction, the complexity of the implementation of the system itself is considered, so the average model interval value is set, and the optimal parameter value is selected within a certain interval, and the algorithm is

15 ( 7 )
After the algorithm is nished, use the bp neural network model to solve the target problem.The ow chart of the use of the bp neural network algorithm model is shown in Fig.1:

Figures
Figures

Figure 1 Flow
Figure 1

Table 3
Test Cases for Teacher Information Management in Colleges and Universities