Understanding the cultivation mechanism for mental health education of college students in campus culture construction from the perspective of deep learning

Abstract

Nowadays, there is an increase in attention to the college student’s mental health, and to enhance the awareness related to college students’ mental health, colleges and universities have executed an immense range of mental health publicity activities. In order to better combine deep learning with classroom teaching, this paper puts forward a deep learning algorithm formulated on convolutional neural networks. The purpose of this research is to investigate the development and use of a cultivation mechanism for mental health education of college students in campus culture creation from the perspective of deep learning. The study's primary goal is to comprehend college students' mental health training in campus culture creation. The study's objective is to develop experimental outcomes of college students utilizing mental health education courses as an optional or mandatory course. Finally, investigations related to college students' mental health from the current situation in China, the investigation, statistics and analysis related to the college students in China are carried out in this situation. The experimental results of this study show that 62 of the 156 schools and universities assessed provide courses on mental health education for college students that are both obligatory and optional. According to the students questionnaire survey, 86.7% of respondents believe that it is critical to establish mental health related educational courses, 61.9% believe that compulsory courses should be established, and students want to add group guidance or activities to the teaching process to improve their experience and participation.

Introduction

Nowadays, a number of colleges and universities provide psychological counseling for college students through the establishment of "mental health education for college students". However, due to the lack of unified regulations on the content, form, and implementation methods of the state, the teaching form, content and implementation are all groping, resulting in different situations and results. Through the research on the mental health education of college students, it can enrich the mental health education courses of college students, realize the mental health education of college students, promote the mental health of college students, and development of college students is promoted as all-round development. Judging from the research results obtained from home and abroad, scholars have already done considerable research on education regarding mental health. Therefore, to describe the urgent need for mental health education among obstetricians was the purpose of Jing (2022). Today's college students are facing increasingly fierce social competition. Wang (2020) took this as a starting point to conduct education based on mental health for college students to boost their continuous development. The main goal of Pottinger et al. (2021) was to evaluate the role of one-day short lectures in public mental health education. The overall goal of the Sholihat (2020) study was to describe, through a simulation model, the impact of mental health education on the family’s ability to recognize the disorder and care for family members having mental disorders. However, these scholars' exploration of mental health education lacks a certain technical demonstration, and it can be found that there is better research on mental health education based on deep learning. In this regard, we consulted the relevant literature on deep learning.

Some scholars also have some research on deep learning. Litjens et al. (2017) introduced important deep learning concepts for analysis of medical image and more than 300 contributions done in this area. O’shea and Hoydis (2017) proposes and discusses at the physical level some new applications which the communication system understands as an auto-encoding process and regards it as a new end-to-end reconstruction. Young et al. (2018) proposed many deep learning models and algorithms for NLP tasks, and gave the process of their evolution. However, from the perspective of deep learning, there are few studies on how to construct and apply the training mechanism for college students' mental health education in the construction of college campus culture. The goal of this paper is to explore the construction and application of the cultivation mechanism for mental health education of college students in campus culture construction from the deep learning perspective. The major objective of the study is to understand the mental health training of the college students in campus culture construction. For better performance, deep learning is combined with the classroom teaching which led this study to put forward a deep learning algorithm on CNN (Convolutional Neural Network). This study gathered data with the help of questionnaires to conduct surveys among the students from China. China is considered for the evaluation of mental health training mechanisms for college students in the development of campus culture in universities and colleges. The aim of the research is to come up with the experimental results of college students in implementing the mental health education courses as an elective or compulsory course.

This paper is organized as follows: the next section discusses the previous relevant research works, then in the succeeding section the campus culture construction from the deep learning perspective is described. The fourth section includes investigation and analysis of the educational courses on mental health for college students in the current scenario. The fifth section concludes the paper with results of the survey and the limitations occurred with the glimpse into deep learning.

Literature review

A web based mental health education system was proposed by Tian (2020) for the college counselors. The major objective of his work was to enhance the counselors’ psychological health education using a systematic approach. He used web technologies and computer networks to provide enhanced services of mental health education for the college counselors. His proposed system had two main subsystems namely; an online psychological counseling system and an online expert type psychological health knowledge system. As a result, the psychological health education ability of college counselors were improved. 150 college counselors were randomly selected for filling the questionnaire from which the 80% of college counselors rectified their own mental health issues using the proposed system. And the study limits the remaining 20% of counselors. The overall utility of this system resulted in college counselors' strong psychological common sense and it also promoted the ability of counselors in self evaluation, self balance and self cognition.

With psychological health education using back propagation neural network theory, Huang (2022) developed an entrepreneurship ability evaluation model to assess the innovation and entrepreneurship education in higher education depending upon the practical efficiency. The results referring to the students' assessment indexes are utilized as input parameters, and the students' innovative entrepreneurship ability is assessed. The amount of hidden neurons is computed utilizing the simulation results, and the results are evaluated using tests. After the completion of the upgraded back propagation neural network training, the results show that the real value is legitimate, the predicted value is appropriate, and the highest relative error among them is 1.6%, and that is virtually the optimal solution. The innovation and entrepreneurship education's standardized assessment score method has been demonstrated to be reasonable and capable of effectively capturing students' innovation and entrepreneurship education enthusiasm.

Shang (2022) put forward the cultivation path of psychological health education for college students. His paper not only explores the correlating factors among core value cultivation and mental health education but also analyzes current issues and scenarios. He has measured the conjunction among mental health education and the cultivation of socialist core values. This enables the students to gain knowledge and an overview on preferred socialist core value courses and psychology. He randomly chose 335 college students from a university. And found that most of the students have no willpower to be an entrepreneur and have hedonism. He concluded that students are not having fully matured and developed thoughts, rather they have inadequate emotional control, and are having weak psychological tolerance during difficulties.

Wang et al. (2021) investigated the undergraduate students to enhance their early childhood physical education. Under the concept of full practice, he explored the talent cultivation mechanism based on the preschooler's mental health. The report outlined existing challenges in university early childhood education systems. 110 questionnaires were circulated where 106 were being returned with the responses and 5 were incomplete. So, the total of 101 responses were collected and among them 91.8% were the responses. When the curriculum of early childhood education majors was examined, it was discovered that elective courses accounted for 18.8% of the total and required courses accounted for 81.2%. He concludes that it is essential to concentrate on the use of contextual teaching in the method of physical education practical training, which could not just permit university students to engage in education as frequently as necessary, and yet also develop their physical ability.

Fu et al. (2019) conducted a study based on the learning behavior of students based on deep learning. They have analyzed the automatic evaluation of teaching in the classroom. For this they have developed a multi-framework classroom learning behavior analysis system in which they have implemented convolutional neural networks (CNN) from deep learning context. In their execution they discovered that boys demonstrate the characteristics to overcome the limitations faced in the learning tasks and they are more active when compared to girls. Their system attains automatic, non-interference, classroom learning behaviors’ full-process analysis. The system also classifies the learning behaviors in five clusters such as read-write, listen, hand-up, fatigue, and sideways. Human postures’ uncertainties, misunderstandings, occlusion, influence of camera angle are some of the crucial limitations to be taken care of. Still the system is capable of analyzing the teaching process in the classroom.

Zhang et al. (2021) researched on the creative political and ideological curriculum of college students. They proposed an approach incorporating a teaching error feedback system linked to sources of knowledge as well as video and audio input data generated during the teaching process in the classroom. The methodology is a group strategy led by teachers wherein learners obtain and use skills and knowledge to discover logistical problems through a thorough method of learning. A structured method for measuring a teacher's performance in the classroom as well as an assessment for constructive advice on teachers' personal advancement describe teaching standards. By connecting subjective and objective data, the model generates reports of teaching anomalies. The technical assessment has an important role in the learning and teaching process.

Before heading to class, numerous students are excited over learning course material (Şen & Hava, 2020). Sometimes university graduates struggle to adjust to the variations in learning patterns between high school and college (Cosbey et al., 2019). Others experience learning difficulties and experience extreme anxiety and stress over studying, including exams. The college campus culture is the pinnacle of civilization. On campus, university students must deal with a variety of interaction setting demands. College students possess excellent communication and interpersonal abilities (Hu, 2022). A research of Kevser (2020) implemented a deep learning concept to study the flipped classroom among the undergraduates. The flipped classroom is more effective than the traditional classrooms and this method improves the emotional and cognitive engagements (Zainuddin et al., 2019).

Campus culture construction from the perspective of deep learning

Campus culture construction

Ancient and modern, at home and abroad, the definition of "culture" in academic circles is a matter of opinion. Different angles reflect human exploration of cultural issues, but in the school development process, culture of university campus has some commonalities in its values, ways of thinking, and codes of conduct. A good campus culture can cultivate students' sentiments, enlighten their minds, and fully develop their minds and bodies (Han, 2021). Figure 1 shows how the culture of campus is constructed.

Fig. 1.

Campus culture construction

The primary task of a school is to educate people, so the construction of a school’s campus culture is definitely student-centered. At the same time, forming a cultural group takes extracurricular cultural activities in the sense of main content and campus spirit as an important feature when the campus has the main space (Li & Zheng, 2021). The campus culture is constructed with an interactive, permeable and inherited environment. Its main functions are as follows: the first is to develop the scientific, cultural, moral and ideology-based abilities of teachers and students; the second is to establish noble morality; the third is to carry out various forms of literature and art, sports, military affairs, theoretical discussions, academic reports, etc., to create a vigorous and aggressive social atmosphere (Jia, 2021).

The long-term plan construction for campus culture characteristic with professional setting and construction of curriculum in a systematized and foresighted way of training teachers for it. Campus culture construction should involve the connotation and humanistic atmosphere that cultivates humanistic quality in students. As a result, the development of the college community should really have a strong foundation for scientific literacy and humanities integrity with the construction of humanities education edification has been highlighted (Li, 2020a).

Deep learning

Deep learning is a kind of continuous deep learning that is influenced by many factors such as social environment, school and individual, among which social environment and school are external influencing factors. Learning is the behavior directly issued by the individual itself, and the individual student is considered to be the most critical internal factor impacting deep learning, which directly affects the effect of deep learning (Li et al., 2021). The below Fig. (2) describes deep learning and its application in the teaching classroom.

Fig. 2.

Application of deep learning in the classroom

Deep learning is an algorithm set in machine learning that is able to learn important raw data features automatically as improved parallel processing abilities of hardware such as graphical processing units, increased data size substantially and implementation of advanced machine learning algorithms (Ahmad et al., 2019). Deep learning depends on the inherent motivation in students, which can make students have a strong interest and curiosity in learning, and can actively learn. At the same time, deep learning focuses on the internal connection and transfer application of knowledge, strives to explore its essential meaning, and pursues the development and advancement of higher order thinking skills which includes critical thinking and problem-solving skills. While shallow learning is a type of learning that replicates information. Learning is carried out under the external influence, and it is only to memorize scattered knowledge points in order to cope with the examination and the knowledge and skills have reached a certain standard to a certain extent (Li, 2020b).

Convolutional neural networks

Convolutional neural network is a deep learning algorithm based on multi-layer neural networks, and its training mainly relies on the back-propagation algorithm (BP algorithm), which is the same as other neural networks (Yang & Li, 2020). The training of the convolutional neural network is a typical supervised learning training. Because of the supervised training, the convolutional neural network needs to use a data set with sticky notes, so as to facilitate the training. At the beginning of training, the convolutional neural network needs to set these parameters to small random numbers. If the parameter setting is large, it is easy to saturate the network and lose the learning ability (Chen et al., 2014).

Convolutional neural networks are a very important algorithm in deep learning. It has many advantages that other algorithms do not have. Specifically, it can be divided into the following points: First, no other preprocessing is required when the image is input, and it is directly input to the convolutional neural network, which can simplify many complex operations; second, the function of features extracting is added to convolutional neural network, either explicit or implicit features can be easily and conveniently extracted (Codella et al., 2017) Making full use of the convolutional neural network algorithm in deep learning can better enable deep learning to run through the entire teaching classroom. Ogueji and Okoloba (2022) conducted a qualitative data analysis to determine the psychological issues which are a challenging factor of the target groups in the behavior of seeking professional help. A mixed method design was proposed by them which gathered the data of Nigerian black families and the United Kingdom through online social platforms such as facebook. As a result of their research they found that there is no noticeable statistical difference in the behavior of seeking professional help, where the majority of the respondents were ready to seek professional assistance. In the qualitative perspective they concluded that socio-cultural factors, accessibility, stigmatization, confidentiality concerns, infectious disease risks and lack of knowledge on psychological health are the challenges in seeking professional help.

Analysis of mental health education courses for investigating college students in current situations

The learning process of convolutional neural networks includes forward propagation and back propagation. The forward propagation stage is as follows: first input a sample in the network, namely (A, B), and then through the process of convolution and sampling and full connection, the final output U can be obtained. The above process can be expressed by the following formula:

$$U_j=G_p\left(\cdots ,\left(G_2,\left(G_1,\left(A_p,V_1\right),V_2\right)\right),\cdots ,V_p\right)$$ 1

The second is the back-propagation stage, which refers to the deviation based on supervised learning between the actual outputs received (U) and the expected output (B). In order to minimize the error, back-propagation is performed and the weights are gradually adjusted (Schirrmeister et al., 2017). The process of the training of entire convolutional neural network is as follows:

First, in the stage of forward propagation, the input parameters are convolutional and sampled respectively. Assuming that there are A categories and B training samples, there are:

$$R^c=\frac{1}{2}{\sum }_{c=1}^c{\sum }_{H=1}^A{\left(d_h^p,-,b_h^p\right)}^2$$ 2

Secondly, considering the error of the sample, the error for the p-th sample is expressed by the following formula:

$$R^p=\frac{1}{2}{\sum }_{h=1}^s{\left(d_h^p,-,b_h^p\right)}^2=\frac{1}{2}{\Vert d^p-b^p\Vert }_2^2$$ 3

Among them, the output of the current layer can be represented by the following function:

$$\begin{array}{cc}{a}^k=g\left(w^k\right)with& w^k=V^k{a}^{k-1}+\end{array}{y}^k$$ 4

Among them, g represents the activation function, V represents the matrix in which the weights are arranged, and y still represents the bias. In addition, in the backpropagation stage, the error in the backpropagation process is first considered, which is expressed by the following formula:

$$\frac{\partial R}{\partial y}=\frac{\partial R}{\partial W}\frac{\partial W}{\partial y}=\tau$$ 5

The back-propagation relationship can be expressed by the following relationship:

$${\tau }^k={\left(V^{k+1}\right)}^D{\tau }^{k+1}\ast g\left(W^k\right)$$ 6

The error of the neurons in the output layer can be expressed by the following relation:

$${\tau }^k=g\left(w^k\right)\ast \left(b^p,-,d^p\right)$$ 7

The next is to start to update each neuron, which is expressed by the following formula:

$$\frac{\partial R}{\partial V^k}=a^{k-1}{\left({\tau }^k\right)}^D$$ 8

In addition, we also need to understand the update of the layer of convolutional weights inside the convolutional neural network:

$$a_m^k=g\left({\sum }_{n\in Qm},a_m^{k-1},,\ast ,h_{\mathit{nm}}^k,+,y_m^k\right)$$ 9

As can be seen from the above formula, the characteristics of each convolution kernel are different, and although all the features come from the convolution of the same input, their cores are different (Zhu et al., 2017). As the downsampling follows the convolutional layer of the convolutional neural network, the update of the neuron weights on the convolutional kernel can only be known by obtaining the output error. Therefore, the error in the output layer of the neurons is obtained from the errors of the neurons in the output layer of the next layer (Wang et al., 2016). Even more difficult, due to the limitation of down sampling, only the error of the down sampled pixels can correspond to the output feature map pixels in the convolutional layer. Therefore, it is necessary to perform up-sampling on the down-sampling process first to obtain the corresponding error. After passing the above steps, the feature map with the same size as the convolutional layer can be obtained, and then the error of the convolutional base layer can be obtained. The specific formula is as follows:

$${\tau }_m^k={\phi }_m^{k+1}\left(g,\left(w_m^k\right),,\ast ,u,p,\left({\partial }_m^{k+1}\right)\right)$$ 10

Among them, up refers to the up-sampling process mentioned above, in which the Kronecker product can be used to achieve the above-mentioned up samplings

$$\tau up=a\otimes 1_{p\times p}$$ 11

Second, we need to find the gradient of the bias, the specific formula is:

$$\frac{\partial R}{\partial y_m}=\sum w,o\left({\tau }_m^n\right)wo$$ 12

Next, the obtained gradients need to be summed again, the specific formula is:

$$\frac{\partial R}{\partial h_{\mathit{nm}}^k}=\sum w,o\left({\tau }_m^n\right)wo\left(i_n^{k-1}\right)wo$$ 13

The above formula can be realized by Matlab programming, and its specific procedure is:

$$\frac{\partial R}{\partial h_{\mathit{nm}}^k}=rot180\left(c,o,n,v,2,\left(a_n^{k-1},,,r,o,t,180,\left({\tau }_m^n\right),,,{\mathit{valid}}^{{}^{\prime }}\right)\right)$$ 14

Among them, the function in Matlab is to rotate it, then perform the calculation, and finally flip it over (Goh et al., 2017). In addition to the convolutional layer, the weight update of the sampling also needs to be considered. The following is the calculation process of the down sampling layer:

$$a_n^k=g\left({\phi }_n^k,d,o,w,n,\left(a_n^{k-1}\right),+,y_m^k\right)$$ 15

In contrast to up, the down function represents a down-sampling function, which refers to the pixel sum operation of the feature map in the input layer (Majumder et al., 2017). For obtaining the error of the down sampling layer, the above formula is understood and the following formula is obtained:

$${\tau }_m^n=g\left(w_m^k\right)\ast conv2\left({\tau }_m^{n+1},,,r,o,t,\left(H_m^{k=1}\right),,,{\mathit{full}}^{{}^{\prime }}\right)$$ 16

Among them, $rot180\left(H_m^{k=1}\right)$ means that the convolution kernel is rotated. After that, the gradient of offset $\phi$ is calculated, which can be calculated by using the formula given below

$$\frac{\partial R}{\partial y_m}=\sum w,o\left({\tau }_m^k\right)wo$$ 17

Next, the down sampled map is represented by the following formula:

$$\frac{\partial R}{\partial {\phi }_m}=\sum w,o\left({\tau }_m^k,,\ast ,t_m^k\right)wo$$ 18

Survey objects

This paper takes China as a sample and uses the method of questionnaire survey to count and analyze the college student’s mental health status in China. The school questionnaire conducted a survey on the education institutions of various colleges and universities regarding mental health from the perspective of curriculum setting, and sent it to the education institutions of various universities and colleges related to mental health in writing and e-mail. On this basis, this study sent 64 questionnaires and collected 52, with a recovery rate of 81.3%. Among the 142 colleges and universities, there are 22 undergraduate colleges and 30 junior colleges; 46 public colleges, 6 private colleges, 15 comprehensive colleges, 16 science and engineering colleges, 7 teacher training colleges, and other 14 colleges and universities including agriculture, forestry, medicine and so on. A total of 30 multiple-choice questions, 5 multiple choice questions, and 1 open-ended question were set up in the questionnaire; it was mainly carried out from three aspects: the degree of attention, the attitude and the evaluation effect. A total of 1000 copies were distributed in this study, of which 832 copies were recovered, with a recovery rate of 83.2%. The sample data are: 410 undergraduates (see Table 1) and 801 junior college students (see Table 2).

Table 1.

Undergraduates students’ classification

Categories	Boys	Girls
Freshmen	61	119
Sophomores	45	72
Juniors	40	42
Seniors	14	17

Table 2.

Represent junior college student classification

Categories	Boys	Girls
Freshmen	115	105
Sophomores	120	193
Juniors	76	107
Seniors	47	38

Current situation of course offering

The school questionnaire surveys the current situation of curriculum offering from the aspects of curriculum offering form, class hours, class size, teaching materials, teaching staff, curriculum objectives, curriculum content, curriculum evaluation, and existing problems.

1. Course opening form

The survey shows that among the 62 universities, 48% of them offer compulsory courses for mental health education, and 10% of the universities not only offer compulsory courses, but also provide corresponding elective courses, and 39 universities offer only elective courses, and another 23 universities offer no courses. The specific statistics are depicted in Table 3:

Table 3.

Statistics related to mental health education courses offered by colleges and universities to the college students

Course form	Frequency	Percentage (%)
No class	8	15.3%
Electives	19	36.6%
Compulsory + Elective	10	19.1%
Required course	15	28.9%

According to Table 3, there are differences in the form of this course offered by universities China in terms of school level and school nature: at the school level, among undergraduate colleges, 50% of the schools offer compulsory courses and 50% of the schools offer elective courses; among the 6 private colleges and universities surveyed, 33.3% of the schools offered compulsory courses, and 66.7% of the schools did not offer courses. Generally speaking, in terms of offering courses, undergraduate colleges are better than junior colleges, and public colleges are better than private colleges. The statistical results are shown in Fig. 3 (a to d).

• (2)The size of the class

Fig. 3.

The classification statistics of mental health education courses related to college students in terms of school level and school nature, (a) Undergraduate, (b) Specialty, (c) Public and (d) Private

Course teaching is the main body of school education. The educational courses on mental health are mostly organized in class format and the size of the class has a great influence on the class teaching effect on college students. Statistics show that the class size of college students enrolled in educational courses for mental health is basically set at more than 80 people, which is a typical large class teaching. In the interviews with students and teachers, it is generally believed that large class teaching weakens the teaching effect of the course to a great extent.

• (3)The grade of the course

In the opening grade, schools that offer compulsory courses generally arrange the courses in the freshman year, and elective courses are generally arranged in each grade, but mainly in the sophomore year. In the student questionnaire, 51.9% of the respondents took the course in their freshman year, 36.5% of them took the course in their sophomore year, and 11.4% of them took the course in their third year.

• (4)Number of class hours

Colleges and universities vary in the setting of course hours. There are 13 colleges with 11–20 class hours, 18 with 30–40 class hours, and 9 of the 10 colleges and universities with less than 10 class hours are specialized colleges. The specific situation is shown in Fig. 4.

Fig. 4.

Survey statistics on the setting of class hours for college students' psychological education courses

In addition, in students' evaluation of class schedule, 45.4% of students from college believed that the number of class hours for mental health educational courses in their schools was just right, and 44.7% of college students believed that the courses were relatively few. It can be seen that the present number of class hours taken in most colleges and universities cannot meet the needs of students. The details can be observed in Table 4.

Table 4.

Statistics of evaluation of mental health education course hours done for college students

Number of course hours	Frequency	Percentage (%)
Very little	101	12.5%
Less	257	32.2%
Just right	364	45.4%
More	64	8.1%
A lot of	15	1.8%

The results of the comprehensive school questionnaire and student questionnaire show that most college students' mental health education courses offer less class hours, therefore, the students actual needs are not met. This paper believes that 24–36 class hours is more appropriate, because to build mental health knowledge, concept system and necessary skill training, less than 24 class hours cannot complete the teaching task.

• (5)Construction of teaching staff

Teachers are one of the subjects in the course teaching situation and an important part of course teaching research. Regarding the mental health education faculty assigned to college students, the statistical results are shown in Fig. 5.

Fig. 5.

(a, b) Survey statistics of full-time and part-time psychology teachers and college majors or universities majors

The overall 52 colleges and universities surveyed in this paper have a total of 314 teachers, 128 teachers in 30 junior colleges, and 186 teachers in 22 undergraduate colleges. Among them, 48.3% of the total number of teachers graduated from psychology majors, and 51.6% of teachers graduated from education, philosophy, ideological and political majors; full-time teachers account for 33.6% of the total number of teachers, and part-time teachers account for 66.4%. It is observed that the mental health education courses in colleges by teachers also conducted in universities are mainly part-time teachers with a small percentage of full-time teachers. Data obtained from student questionnaires show that part-time teachers are mainly composed of class counselors and ideological and political educators.

• (6)Survey on course implementation process

Through the statistics of student questionnaires, classroom discussion method, case explanation method and knowledge teaching method are the main methods used in the teaching process, while audio-visual material interspersed method, group coaching training, and special activity development method are used less frequently. The specific statistics are shown in Fig. 6.

Fig. 6.

Statistics of teaching methods used in college students’ mental health education courses

College students' evaluation of the course effect

1. Curriculum Effect Evaluation Survey

According to the data obtained from the student questionnaire, 4.4% of the respondents thought that the course had a good effect, 31.7% of the respondents thought it was relatively good, 51.3% thought the effect was average, and 12.8% thought it was relatively poor. The statistics are depicted in Table 5.

Table 5.

Survey of overall evaluation of the effect of mental health education courses on college students

Overall evaluation	Frequency	Percentage (%)
Very poor	26	4.1%
Relatively poor	81	12.8%
General	324	51.3%
Relatively good	172	27.3%
Very good	28	4.4%

The effectiveness of the course is evaluated for college students by the impact on mental health due to the specific content of the educational course. That is to say, as this course is evaluated from the aspects of developing college students’ curiosity in learning mental health related knowledge, mastering the mental health knowledge, improvement of academic performance, emotional management, improvement of interpersonal relationships, guidance of romantic relationships, self-evaluation, and the guidance of graduation. The specific situation is shown in Fig. 7.

Fig. 7.

Statistics on the impact of education courses of mental health on students of college

Based on Fig. 7, the cumulative percentage of students whose mental health education courses have a great and relatively large impact on their own mental health is over 40%. It proves once again that the education courses on mental health in colleges as well as universities have a certain influence on students, and the implementation of the courses has certain effects. However, it is undeniable that in the daily teaching process, due to the serious tendency of subjectivation and the over theoretical knowledge imparting, the practicality, operability and pertinence of the course are not strong, and there is still a certain gap between the actual effect of the course and the ideal effect.

• (2)Survey on the degree of emphasis and attitude of college students on the course

In terms of the degree of emphasis on the course and the understanding of the object of the course, 45.9% of the respondents believed that it was very necessary to set up the course, 40.3% of the respondents thought it was necessary, 10.5% did not care, and 4.4% of the students thought it was basically unnecessary or completely unnecessary. In general, college students basically recognize the importance of education courses based on mental health. For the understanding of the object regarding the course, 78.6% of the respondents believed that the object of the course was all students. It has been observed that the universality of mental health education has been basically accepted by everyone, and college students generally realize that mental health education is closely related to themselves. Table 6 contains the specific data.

Table 6.

Investigation regarding the awareness of necessity of mental health education courses by college students

Course Necessity	Frequency	Percentage (%)	Course object	Frequency	Percentage (%)
Very necessary	369	45.9%	all students	629	78.6%
Relatively necessary	319	39.2%	students with mental illness	24	3%
It doesn't matter	83	10.5%	Students with serious psychological problems	81	10.1%
Basically unnecessary	23	3.5%	students with mental problems	62	7.7%
Totally unnecessary	7	0.9%	other	5	0.6%

Results

For every survey there is a common implication of trust, during the period of data collection there exists a risk of consistency due to trust issues. To identify the kind and intensity of psychological disorders, persons with psychological problems might be asked to complete psychological scales or questionnaires (Tian, 2020). Deep learning provides a deeper model to enhance generalization and also a wide scale course (Huang, 2022; Zhang et al., 2021). The childhood physical education and mental health education is provided for the students to develop all the skills both mentally and physically (Hu, 2022; Shang, 2022; Wang et al., 2021). The flipped classroom is more productive than regular classes, because it increases mental and emotional engagement (Kevser, 2020; Zainuddin et al., 2019; Şen & Hava, 2020). The people may not respond the truth to the surveyor or the researcher may not get the appropriate data from the expected resource. These can be considered as the implications since these uncertain data are forced to implement on the analysis process (Kabir, 2016).

Taking China as an example, the method used by this paper is questionnaire survey to carry out the investigation and analysis in the current situation related to the mental health education courses of college students in China. The construction of educational courses in undergraduate colleges related to mental health paid more attention than junior colleges. According to the courses offered, all 22 undergraduate colleges offer courses, while the proportion of junior colleges offering such courses is only 73.3%. In terms of study time arrangement, 68.2% of the undergraduate colleges and 40.8% of the junior colleges have more than 12 credit hours, while only 65.4% of the junior colleges have more than 12 credit hours. These data clearly show that undergraduate colleges pay more attention to students' mental health education than junior colleges.

In terms of course offering and teaching methods, according to statistics, more than 92.6% of the respondents hope that the education courses on mental health for students in college will be offered in the form of compulsory or elective courses. Among them, 61.9% of the respondents asked for compulsory courses, 30.7% of the respondents believed that only elective courses were required, and only 7.3% of the respondents held an indifferent attitude. In terms of which teaching method to adopt, 59.2% of the respondents believe that it should be a mixture of teaching in the classroom and group activities. The specific statistics can be observed in Table 7.

Table 7.

Survey statistics of students demanding for courses in college

Course form	Frequency	Percentage (%)	Teaching methods	Frequency	Percentage (%)
Required course	239	29.9%	classroom lecture	41	5.1%
Electives	249	41%	group activity	89	11.1%
Compulsory + Elective	254	31.7%	Mainly lecture in class	95	11.8%
It doesn't matter	54	6.7%	Mainly group activities	103	12.9%
Other	5	0.6%	Class lectures + group activities	473	59%

In terms of class size and assessment methods, according to statistics, 33.8% of the respondents hope that the number of students should be less than 30 in the class, and 41.3% of the respondents believe that the number of students should be controlled between 30 and 50 in the class. This is in stark contrast to the large class teaching of more than 100 students in the current survey, thus indicating the desire of students to implement small class teaching in this course. At present, most of the examinations of education courses on mental health in colleges as well as universities are in the form of examinations: examinations and examinations plus homework. In the statistics on how students want to be assessed, only 26% of students want to take exams or exams plus homework. 55.3% of college students hope to take the examination method, because the examination method can reduce their academic burden. The specific statistics are shown in Fig. 8. In a previous study of Abbasi et al. (2020), a total of 382 students were considered as respondents to determine the view of students towards the online learning platform at the time of pandemic. Around 77% of students show a negative opinion on the online platform but as a contradiction 76% of students are accessing online platforms for learning. On the content of the education course on mental health, multiple-choice questions are designed in the student questionnaire for investigation, and the students are ranked according to the importance of the course content. The specific statistics are shown in Fig. 9.

Fig. 8.

Survey statistics of college students' hopes for class size

Fig. 9.

Survey statistics of content ranking that college students hope that mental health education courses should be taught

In terms of the teaching methods they hope to adopt, the order of the teaching methods that college students like to use by teachers in mental health education is role-playing, case explanation, and classroom discussion. In the study of Powell et al. (2022), a questionnaire based survey was conducted by 297 participants, from the psychological aspect, 19.7% were not having anxiety but 5.4% were at extreme to go back to university. 60.2% of students were anxious about the spreading of COVID-19. According to the current study, in the questionnaire survey, some students directly stated that they would like to teach in a group tutoring way, so as to develop the association of students. And one of the main methods in practical teaching, the knowledge teaching method, was ranked last. It is extremely important to get views regarding students' opinions toward the professor's teaching procedures and to incorporate such sentiments into a procedure of change (Zhou & Mann, 2021). Learners are happier and thus more supportive of the function it performs in developing educational performance, and the classroom has evolved toward a more mutually advantageous learning environment (Cosbey et al., 2019; Fu et al., 2019; Kevser, 2020). It can be seen that students reject traditional knowledge teaching methods and are eager for diverse teaching methods. The details are shown in Fig. 10.

Fig. 10.

Statistical expectations of students on the teachers teaching methods in college

According to Tartavulea et al., (2020), the recent technical developments have encouraged the colleges and universities across the globe to change the learning mode to online and developments in ICT (information and communication technology) has a larger contribution towards the developing innovative educational techniques (Kauppi et al., 2020). Electronic learning platform supports uncomplicated access to the learning environment, supports time flexibility so that learners can ignore the issues of time and space, and also provides new context for teaching which will lead to focus on possibilities and needs of each student (Huang et al., 2020). Although it was predicted that in the year of 2018, 15% of market share is owned by online educational platforms (Burquel & Busch, 2020).

Discussions

The negative psychological effects are mostly observed due to COVID-19 epidemic (Lai et al., 2020; Wang et al., 2020). Undergraduate colleges pay more attention to the construction of students' education courses on mental health in college than junior colleges. Higher grades, harmful habits like use of alcohol, sexual conduct, and gaming addiction, physical ailments, and familial variables all increased the likelihood of developing mental health problems (Luo et al., 2020). From the point of view of course offering, 22 undergraduate colleges and universities offer 100% courses, while only 73.3% of junior colleges offer this course; judging from the name of the institution, most of the undergraduate colleges are mental health education and counseling centers, and most of the junior colleges are psychological counseling centers; from the perspective of attribution, 72.7% of undergraduate colleges assign this institution to the Academic Affairs Office, 18.2% of independent institutions, 70% of junior colleges and 10% of independent institutions; in terms of class schedule, 68.2% of undergraduate colleges have more than 12 class hours, 40.8% have more than 30 class hours, and only 65.4% of junior colleges have more than 12 class hours.

The above data clearly show that undergraduate colleges specifically pay more attention to mental health education of students than junior colleges. Even after controlling for background factors, the findings show that adolescents who suffer basic necessities insecurity are meaningfully and considerably more probable than their self sufficient colleagues to report anxiety, depression, and planning, attempt or suicide ideation (Broton et al., 2022). Depression, stress, anxiety, fear, interpersonal sensitivity, psychological disorders, discomfort, excessive interpersonal sensitivity, suicidal ideation, sleeplessness, negative emotions, insecurity, and trauma were recognised as types of mental health difficulties (Shan et al., 2022). The construction of students' mental health in college education courses paid more attention in public colleges rather than private colleges. Graduates attending public universities who transfer to private universities are a distinct category because they are already a significant portion of the student population at private universities (Chen & Li, 2021).

Most of the recent researches states that the online courses has significant effect on the students mental health (Ajmal & Ahmad, 2019; Zhou et al., 2022)0.91.3% of public colleges as well as universities offer education courses related to mental health and 85.7% of colleges and universities have arranged the courses into their regular teaching plans; 62.2% of colleges and universities have a student–teacher ratio below 2000:1. According to the findings of a multicenter study of 1,563 medical staff members, the prevalence of anxiety and depression was 45% and 51% (Liu et al., 2020). However, due to the different purposes of establishing schools, private colleges generally do not connect students' mental health education and its importance, and only 33.3% of colleges and universities offer this course; in terms of faculty ownership, only 40% of students have a student–teacher ratio below 2000:1. Social interaction and outdoor activities got reduced during COVID-19 outbreak which caused depressive symptoms in children (Xie et al., 2020). As a bright spot of higher education in China, the work of education on mental health for college students should be strengthened.

Conclusions

Deep learning is a feature extraction technology based on unsupervised and semi-supervised, and adopts hierarchical feature extraction technology to replace artificially obtained information. Its purpose is to establish and simulate the neural network in the human brain. The aim of this paper is to explore campus culture construction from the deep learning perspective, the construction as well as application of the cultivation mechanism of mental health education for college students. 44 out of 52 colleges offer both compulsory and elective-based mental health education. In terms of the degree of emphasis, undergraduate colleges are the best, and basically all have relevant courses, followed by junior colleges and private colleges. In the student questionnaire survey, 86.7% of the respondents believe that it is necessary to set up mental health education courses, and 61.9% of the respondents hope that it can be conducted in the compulsory courses form. And students hope that the course can add group tutoring or activities in the teaching process to improve their experience and participation. In the course effect investigation, there are some problems, such as the students' evaluation of the actual effect of the course is not very high, and the present situation of construction of the course is unable to meet the students’ actual psychological needs.

The research is structured on the basis of 142 colleges and universities such as undergraduates colleges, junior colleges, public colleges, private colleges, comprehensive colleges, science and engineering colleges and teacher training colleges along with 14 other colleges and universities such as agriculture, forestry, medicine and so on. Colleges and universities are being determined on factors such as courses provided as elective and other courses, different class hours, size of class, construction of teaching staff with experience and abilities based on course implementation process take up. Effect of mental health education courses depending on their frequencies and impact of courses on emotional state and relatively future exceptions of students such as professional and personal life.

Teachers adopt methods for mental health education as class discussions, role play, case study related to particular topics and group coaching as per practical sessions. The study material implemented for mental health education is different types such as audiovisual material, thematic activities involving student, knowledge and other real time examples or case studies as assessment for students to learn in better way with grasping the knowledge in proper aspects related to subjects which are being selected for student academics.

Author contribution

Qingsong Gao is contributed on the text editing and writing of this paper, Yongxia Wei is responsible for the language proofreading, his is responsible for the data collation and analysis, contributed on the project implementation and organization.

Data availability

Datasets can be provided by corresponding authors upon request.

Code availability

Not applicable.

Declarations

Ethics statement

The studies were reviewed and approved by the Nantong University and Nantong University Xinglin College. The participants provided their written informed consent to participate in this study.

Ethics approval

This article does not contain any studies with human participants.

Human and animal rights

This article does not contain any studies with human or animal subjects performed by any of the authors.

Informed consent

Informed consent was obtained from all individual participants included in the study.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Conflict of interest

The authors state that the study was conducted without any potential conflict of interest.