Data monitoring in student psychological automatic evaluation system by using dynamic acquisition algorithm

The purpose of designing embedded Internet of things and studying the ad hoc technology of gateway is to meet the monitoring requirements brought by the continuous expansion of data center. Among them, in the aspect of dynamic data monitoring, this paper designs a dynamic acquisition algorithm, that is, by analyzing the characteristics of data center resources and the number of information acquisition methods, so as to adjust the data acquisition cycle and to reduce the monitoring workload and improve the monitoring flexibility. At the same time, this paper also studies the automatic psychological evaluation system of students. It is found that with the intensification and efficient operation of social competition, some special groups of college students are facing increasing pressure, and events caused by psychological problems emerge one after another. Even though there are many psychological evaluation systems on the network, a large number of psychological knowledge and mental health management on the Internet cannot effectively manage and improve the psychological status of college students. Therefore, the university needs to build a psychological evaluation system for college students in time and combine the automatic psychological evaluation system for college students with embedded Internet of things and cloud service technology to improve the accuracy of psychological evaluation. The application of the system, college students can independently evaluate their own mental state, if there are psychological problems can be timely counseling, not only can improve the level of mental health education in colleges and universities, but also can promote the development of physical and mental health.


Introduction
After the computer and Internet, the embedded Internet of things information and communication industry has made another major breakthrough and become an important foundation for the sustainable development of economy and society. The birth of the embedded Internet of things breaks the boundary between the virtual network world and the real physical world. It is the extension and integration of the Internet in the real world (Varaprasad et al. 2020). It can be considered that the embedded Internet of things is a wireless self-organizing LAN composed of different information devices (Manavalan and Jayakrishna 2019). It is the key equipment to realize the communication between heterogeneous networks. It can be connected to the Internet through various access networks. IOT gateway, as a bridge between sensing network and traditional network, is responsible for information exchange and transmission between networks (Ishaq et al. 2013).In recent years, the rapid development of Internet of things technology has promoted the practical application of various information technologies, and the dynamic data monitoring technology has also made great progress; dynamic data monitoring has become an indispensable part of the whole Internet system (Fu et al. 2018;Fathy et al. 2018). At present, with the high collection of information, even if many improvements are made in the management of information resources, the complexity of dynamic data such as information resources cannot be effectively guaranteed, which leads to the high concentration of risks of information resources in the process of dynamic supervision (Dong and Qin 2018). This is because there may be some failures in the infrastructure, so these storage devices, network devices, and cloud platforms will also have corresponding failures during the operation of the supervision system (Qin et al. 2020). These failures will inevitably affect the normal operation of the system or service, lead to the response delay to user requests, and affect the user experience. This paper also studies the automatic evaluation system of students' psychology. For college students in the most important and critical stage of life, although they are full of passion and vitality at this stage, their psychology and behavior are easy to change according to the external environment because of their poor ability to bear pressure and large psychological fluctuation (Sreeramareddy et al. 2007;Pedrelli et al. 2015). Based on network technology, modern information technology provides an advanced model for educational work, which can change the way of education, learning and university life. The informatization and interaction of the network have created a new model for college students' psychological education. It can integrate network technology into psychological education and realize network psychological education ). This educational model has been recognized and adopted by college educators.

Related work
A wireless ad hoc network on-demand multi-path distance vector routing protocol E-AOMDV was proposed based on node energy optimization. Energy consumption can be balanced by adding node energy model through energy control strategy (Sethuraman and Kannan 2017;Javadpour et al. 2023). The practical results show that network overhead and routing initiation frequency performance of E-AOMDV routing have been significantly improved. The on-demand remote vector multi-path AOMDV protocol based on AODV extension simulated the protocol. The AODV and AOMDV routing protocols are simulated on the network simulation platform. The simulation results show that compared with the AODV routing protocol, the AOMDV routing protocol has the characteristics of higher network delay and lower routing starting frequency (Asokan et al. 2007). In addition, when the network topology changes rapidly, the AOMDV routing protocol has obvious advantages in terms of data packet transmission rate. It puts forward the existing data center monitoring technology and the main data center monitoring system and analyzes the functional characteristics and shortcomings of the existing monitoring system (Saha and Majumdar 2017;Sangaiah et al. 2022Sangaiah et al. , 2023. On this basis, according to the specific needs of data center monitoring, research has been carried out from the two aspects of reducing the monitoring workload and studying the scalability of monitoring (Bondár et al. 2001). The literature shows the current background and environment of the development of health education and analyzes the system framework, that is, using the B/s and sqlserver structure as the database structure, and MyEclipse as the development environment, planning the structure of query management, mental health assessment management, and system management (Lianov et al. 2020). The literature first evaluates the psychological state of some college students and puts forward the strategies of psychological adjustment for college students combined with the research results. Using the fuzzy comprehensive evaluation method in fuzzy decision making, it solves the transformation of college students' mental health from qualitative analysis to quantitative analysis, so as to effectively improve and manage the psychological problems of college students (Braithwaite et al. 2010). The system using the J2EE technical architecture was introduced. The data layer of the system is responsible for managing the display of user interface data, and calls the corresponding modules and updates the module data based on user interaction data. The user response sent by the business logic layer of the system is returned to the presentation layer. The presentation layer reassembles the data sent back by the business logic layer and finally displays it to the user in a specific view style (Blanco et al. 2008).
3 Embedded IoT and dynamic data monitoring technology

Embedded IoT structure
According to the needs of all sectors of society, the application layer applies the Internet of things technology to modern society to realize the intelligent development of the industry. Its main function is to process and analyze the collected data to improve the level of information intelligence in the corresponding fields. The application layer is based on the information center and other departments to intelligently process a large amount of data obtained by the perception layer and the network layer and then apply it to related fields to realize an intelligent information application network. The network layer transmits and processes the information obtained by the perception layer. The communication network, information center, fusion network, and network management center jointly constitute the network layer of the Internet of things, which mainly realizes information transmission under different networks such as the Internet, mobile communication network, and wireless sensor network. By processing and analyzing a large amount of data acquired by the perception layer, the data becomes meaningful and valuable and is sent to the application layer for practical application. At the same time, through cloud technology, the network layer also supports big data storage to realize the portable access of data.
Two common perception layer technologies are the RFID-based IoT perception method and the self-organized multi-hop IoT perception method. The former is to realize the identification and collection of target information through the detection and analysis of RFID tags. The latter realizes the intelligent sensing and collection of various information through the IOT gateway.
The research of self-organizing wireless network based on embedded platform is not only a pre-verification and correction of actual network technology, but also an indepth exploration of the Internet of things technology. Nowadays, the research on realizing self-organizing network on embedded platform has achieved good results. In order to verify the new protocol and algorithm in the actual experiment, an AdHoc test platform based on RaspberryPi was designed, and the application experiment of the Internet of things and WSAN was carried out. In order to improve the video transmission in the centerless multipoint communication, an AdHoc transmission platform based on the ARM Cortex-A8 core embedded system is designed and implemented.

Dynamic data monitoring
According to these two characteristics of the monitoring data, the data collection system can adopt a non-uniform data collection algorithm to adapt to the value distribution of the monitoring data, as shown in Fig. 1.
If the data is close to the threshold and fluctuates greatly, the monitoring data acquisition interval can be appropriately shortened, the data acquisition frequency can be improved, and the characteristic state values such as violent fluctuation or near the threshold point can be retained as much as possible; If the data is far from the threshold and the fluctuation is relatively gentle, the data acquisition interval can be appropriately increased, the data acquisition frequency can be reduced, and the data points with unclear characteristics can be filtered out. Because the monitoring system is not sensitive to these data points, through the above idea of dynamic data acquisition, the data acquisition interval can be dynamically adjusted according to the characteristics of data fluctuation to meet the monitoring needs. Where the fluctuation is large, the value density is relatively high, while where the fluctuation is slow and far from the threshold, the value density is relatively low because the characteristics of the data itself are relatively concentrated. Under the same data acquisition conditions, more data points that can represent the overall characteristics of the monitoring data can be retained, so as to improve the value density of the collected data and reduce the burden of monitoring tasks and the cost of monitoring system.
We estimate the probability that the next acquisition value exceeds the threshold based on d value, as shown in Fig. 2.
The algorithm maintains a circular queue of length N to measure the recent fluctuation of the collected data. The circular queue is used to store the most recent N d values. The expected value of d and the variance are represented by l and r, respectively, v(t2) represents the collection value of i default time intervals after v(t1), and T represents the threshold. The calculation formula for calculating the probability that the next acquisition point is greater than the threshold T is as follows: Data monitoring in student psychological automatic evaluation system by using dynamic acquisition… 8461 In order to calculate the probability that the next collected value is greater than T through Chebyshev's inequality, the above formula is transformed into the following form: These inequalities are transformed into: The following results can be obtained: According to Chebyshev's inequality and formulas (3) and (4), the following relations can be obtained: The probability obtained by Eq. (5) is the probability that the next acquisition point at the default acquisition interval obtained by Eq. (2) exceeds the threshold T.
Before adjusting the data collection interval, the algorithm needs to calculate the probability that at least one collection point of the current collection interval I exceeds the threshold (I is based on the default collection interval and represents a multiple of the default collection interval), that is, the current collection interval is compared to the default collection. Threshold leakage rate is expressed by b(I): bðIÞ 4 Research on the design of students' mental state recognition and automatic evaluation system

Identification method of students' psychological state
Feature level fusion is to extract the features of multisource signals, respectively, then form a joint feature vector according to a certain fusion algorithm, and input the vector into the same classifier to produce the recognition results. Taking the psychological recognition layer as an example, the general process of feature layer fusion is shown in Fig. 3.
In the field of emotional psychological recognition, the existing feature level fusion methods include principal component analysis, linear discriminant analysis, and  Fig. 3 Feature layer fusion process genetic algorithm. Although the feature level fusion method makes effective use of the correlation and complementarity between different features, it also ignores the differences between different features and requires high synchronization between different modal features. However, it improves the classification accuracy of the classifier and is conducive to real-time processing.
Majority voting gives each classifier the right to vote. Each classifier classifies sample X and votes its classification results. The category with the most votes is the final classification result. The calculation formula of voting results of each category is Eq. (8), and the calculation formula of final classification function is Eq. (9): The summation method is to calculate the sum of the recognition probabilities of each class of classifiers, then compare the sum of the recognition probabilities of each class, and select the class corresponding to the maximum value as the final output recognition result. The calculation formula is (10): The product method first obtains the recognition probability product of each category and each classifier, then compares the products of all categories, and selects the maximum value as the final recognition result. The calculation formula is as follows (11): The maximum value is the final output.
The maximum value method is to calculate the discrimination probability of each classifier of each type of sample x, select the maximum value, and then select the maximum value of the behavior probability of all samples as the output result.
The minimum value method is to find the probability of each class of each pair of classifiers in sample X and select the minimum value. Then, select the minimum value of all sample behavior probabilities as the output result, and the calculation formula is as follows (14): The real-time and fault tolerance of the decision-making layer fusion, strong anti-interference ability, does not require similar sensors and makes full use of the differences between different characteristics. But the limitation is that it ignores the correlation between different features.
Assume that the intelligent answer link usage scale contains N factors. The Q&A record of the intelligent Q&A session is detected by a polygraph combined with facial expression recognition. If the facial expression meets the preset, the answer to the question will be recorded as valid. Calculate the answer record according to the quantitative scoring rule of the scale to which the question belongs, and obtain the score value of factor N, which is recorded as a column A vector. The formula is as follows: Concatenating the feature value of the expression and the physiological signal into a joint feature vector B, then B is expressed as (16): Concatenating the eigenvalues of various physiological signals into a joint eigenvector C, then C is expressed as (17): By weighting and fusing the combined feature vectors obtained in the three psychometric steps, the fusion feature vector p is obtained. If the dimension of vector B is larger, vector A and vector C must be filled with zeros (0), and vice versa. The calculation formula of P is as follows (19): Data monitoring in student psychological automatic evaluation system by using dynamic acquisition… 8463 P ¼ a Á Among them, the fusion weights corresponding to the three psychological measurement links of the joint feature vectors A, B, and C can be defined as a value or a vector containing a series of values according to actual needs.
Choose a facial expression image from each video image in the eNTERFACE database as the key image to identify the emotional state of the facial expression. Use ConvNet network to train emotion and mental recognition models based on facial expressions. The calculation formula is shown in formula (20); x i is the input sample.
i¼1 exp x i ð Þ ; n 2 f1; 2; . . .; 6g ð 20Þ The output of the emotional mental recognition model based on expression and the emotional mental recognition model based on speech signal is linearly fused and weighted. The output classification result is L, and the calculation formulas are (23) and (24): q fus;n ¼ a Á q fac;n þ ð1 À aÞ Á q spc;n ð21Þ Multi-angle training is performed on each sub-diagnosis model, and a severe mental illness diagnosis model based on multi-angle model decision fusion is obtained. The output classification result of the model is L, and the calculation formulas are (25) and (26): q fin;n ¼ a 1 Á q A:n þ a 2 Á q B;n þ a s Á q C;n þ a 4 Á q D:n þ a 5 Á q E;n þ a 6 Á q F;n ð23Þ L ¼ argmax q fus;n À Á ð24Þ Since most of the current research focuses on the correlation between mental illness and imaging, gene expression, and other data, it is impossible to determine the confidence of each sub-diagnostic model classifier.

System requirement analysis
In terms of hardware, the system requires a Pentium 4 series or higher CPU computer with at least 512 M memory and a hard disk capacity of more than 80 GB. As long as the computer meets the above configuration, the system can operate normally. The higher the computer configuration, the faster the system runs.
In terms of software, the database management system selected for this system is SQL Server 2008. SQLSer-ver2005/2008 occupies a large proportion of current software development and use, and there is no doubt in terms of security, reliability, and maintainability. In terms of development tools, the system uses MyEclipse 6.0. MyEclipse is an Eclipse plug-in, which can develop projects based on JavaEE enterprise B/S architecture. It integrates many third-party open source architectures, such as struts. MyEclipse IDE occupies most of the development of Java enterprise projects, so it is feasible to develop this system.
In terms of development language, this system uses Java language and is developed based on the Java EE enterpriselevel framework mvc. These three frameworks are the most popular technologies in javab/s development. These three frameworks implement three different functions, including user presentation layer, business logic layer, and data persistence layer. In terms of workflow engine, the system uses jBPM workflow engine. JBPM is the most popular workflow engine in java development. The work engine is powerful and can use custom JPDL language to create any complex process. Based on the above reasons, this system is also feasible in terms of language development.
The system describes the software development process according to the requirements of software engineering. The requirements analysis stage requires a comprehensive and detailed understanding of the objectives, users' business contents, and processes. This is a process of requirements collection and the basic preparation for requirements analysis. In the feasibility analysis stage, accurately locate the system objectives according to the specific project requirements, further understand the user requirements, and clarify the subsystems of the system. The system interface reflects the overall structural framework of the mental health evaluation system, such as psychological test, psychological counseling, and message module. At the same time, the system establishes the files of each student, including students' basic situation files, class files, students' psychological evaluation and analysis, psychological survey results, students' psychological counseling, etc. Manage the information of psychological test results, modify passwords, set evaluation indicators, calculate statistics, etc. Users can query various information in detail by name and department. The system mainly tests the personality of college students. College students enter the interface of personality psychological test subsystem, select test questions for test, and finally give the test results and corresponding solutions.
The interface style of the system is the same as that of Microsoft office, so it is simple and easy to use. The system uses network software and hardware to provide security measures to ensure the accuracy and security of users' shared data. At the same time, it makes full use of mature network technology and software technology. Taking MyEclipse as the development environment, the system has rich functions and wide support and is responsible for system software maintenance, network maintenance and hardware maintenance.

System module design
The system is divided into three functional modules. Although each module seems to be independent of each other, they are closely related in accessing the database. Each module accesses the same database, but accesses different tables. The functions of each module are based on data collected in previous surveys. The functional structure of the system is shown in Fig. 4.
According to the functional analysis of the above modules, the functional modules of the system include query management, mental health evaluation management, and system management. The system deploys general consultants, mental health teachers, and administrators to access the college students' fuzzy evaluation system on the application server through LAN, Wan, and other platforms through the firewall and call the data in the database server, as shown in Fig. 5.

Evaluation model design
This article conducts a psychological evaluation of college students and evaluates their psychological adaptability, frustration tolerance, emotional stability, temperament, and personality. Finally, the trend of changes in the proportion of each indicator to the overall mental state of students is summarized.

Fig. 5 System deployment diagram
Data monitoring in student psychological automatic evaluation system by using dynamic acquisition… 8465 The survey results of the five indicators of students are shown in Table 1.
The results of the students' emotional stability test are shown in Table 2.
The test results of students' frustration tolerance are shown in Table 3.
Through the analysis of Tables 2 and 3, the author has more cognition of the current college students' psychological state and psychological needs. These data can be used as an important foundation for the analysis of system needs and, as learning samples, to help the construction of relevant systems.

Database design
The student information table mainly describes student information, including student number, student name, gender, Department, professional class, contact information, and related data. The ''student number'' field is the primary key connecting other tables, as shown in Table 4.
The test volume information table mainly explains the psychological test information of college students. It is used to save test paper name, test time, test score, and other information. At the same time, according to the five factors that constitute the psychology of college students, five test papers are formed. Each test paper must undergo a certain number of selection tests, and the test paper score is determined according to the range of the total score of the test paper. Generally speaking, the score of a set of test questions does not exceed 5 ranges. Therefore, the test table defines four groups of test scores, including standard and test scores. Considering the time requirements for some tests, the test time of each group of test files is defined. The number field of the test paper is the primary key connected to the test table, as shown in Table 5.
The information table of test questions mainly explains the psychological test information of college students. The relationship between the test question table and the test paper table is one to many, that is, a record in the test paper  Table 7.   Data monitoring in student psychological automatic evaluation system by using dynamic acquisition… 8467

Conclusion
According to the actual situation of college students' psychological teaching, this paper proposes an automatic psychological evaluation system for students based on embedded Internet of things and dynamic data monitoring.
In the process of application, it can be found that there are still some difficulties in realizing fully dynamic data monitoring. Therefore, while ensuring the monitoring accuracy, an effective method is needed to reduce the burden of monitoring tasks and reduce the amount of monitoring data. With the continuous development of cloud computing, data center resources have been integrated and abstract managed to a certain extent. At the same time, the monitoring of virtual machines has become more and more important. However, due to the existence of various virtualization data technologies, virtual machine monitoring lacks a unified monitoring and management interface. Therefore, in order to improve the overall scalability of the system, it is necessary to define a monitoring consistency interface for the virtualization data platform. At the same time, this paper has done a lot of research work on the college students' psychological evaluation system. Setting the user login identity and authority in the psychological evaluation system can effectively prevent illegal users from deleting the evaluation data, improve the system security, facilitate the database management, provide convenient management methods for the database, and facilitate the management of the evaluation system. Thus, colleges and universities have promoted the online evaluation of mental health and optimized the management.
Funding The authors have not disclosed any funding.
Data availability Data will be made available on request.

Declarations
Conflict of interest The authors declare that they have no conflict of interests.
Ethical approval This article does not contain any studies with human participants performed by any of the authors.