Construction and Implementation of Innovation and Entrepreneurship Education System in Colleges and Universities under the Background of Information Technology

In the modern higher education system, innovation and entrepreneurship education plays an extremely important role, which focuses on the cultivation and enhancement of students' innovative thinking and entrepreneurial awareness, emphasizes the interdisciplinary integration and comprehensive ability cultivation, which helps to enhance the competence of modern college students in all kinds of jobs and their competitiveness in the fierce market environment, cultivate more high-end talents in line with the market development needs, and promote the economic and social development [1-6]. There is consistency between the concept of new liberal arts construction and the purpose of innovation and entrepreneurship education in colleges and universities, and strengthening the innovation and reform of innovation and entrepreneurship education system in colleges and universities under the background of informatization is of great significance for social progress and economic development.

Under the background of informationization, the innovation and reform of innovation and entrepreneurship education system in colleges and universities need to strengthen the combination of theory and practice, provide more learning channels and learning opportunities for students with the help of informationization education platform [7-9]. Firstly, innovation and entrepreneurship education in colleges and universities should not only stay in the teaching of superficial theoretical knowledge, but should strengthen the depth and breadth of theoretical education, take the form of building professional innovation and entrepreneurship courses to guide students to learn the relevant concepts, skills and methods, to understand the history and current situation of innovation and entrepreneurship, and to cultivate innovative and entrepreneurial ways of thinking and literacy [10-13]. Secondly, theoretical education is only the foundation of innovation and entrepreneurship education, and students still need to practice innovation and entrepreneurship through practice. Colleges can establish cooperation with enterprises and social organizations through the information education platform, establish innovation and entrepreneurship practice bases for students, and provide students with practical projects, internships and other resources according to their actual needs [14-16]. At the same time, colleges and universities can organize innovation and entrepreneurship competitions, entrepreneurship lectures and other activities with the support of information technology, so that students can understand the real entrepreneurial environment, motivate them to actively participate in all kinds of innovation and entrepreneurship activities, so that students can form a good teamwork ability in daily practice.

Based on the concept and development of "cloud" media platform, Literature [17] analyzes the advantages of “cloud" media platform in "broad-spectrum" innovation and entrepreneurship education by combining the connotation and mode of "broad-spectrum" innovation and entrepreneurship education, and promotes the effectiveness of college students' innovation and entrepreneurship education. Literature [18] analyzed the data collected from CSIPC participants. The results indicate that entrepreneurial alertness and opportunity perception mediate the relationship between personality traits and college students' willingness to engage in sustainable entrepreneurship. A theoretical basis is provided for sustainable entrepreneurship research and the practice of training sustainable entrepreneurs in universities. Literature [19] conducted a systematic literature review, and the systematization of the existing literature is highly relevant to future research aimed at understanding the interrelationships between entrepreneurial structural transformation within innovation systems and socio-economic systems in general. Such an understanding requires further extended research in areas related to methodology, content and theory. The aim of Literature [20] is to explore how to move from distinguishing between entrepreneurship and innovation to creating value from educators and policy makers. In practice, educators and policy makers have to tend to focus on the integration between the two. Research in Literature [21] examines the historical relationship between the adoption of compulsory education laws at the national level and measures of innovation and entrepreneurship. Literature [22] suggests good practices that could be replicated in other countries in the region and recognizes the immense value of higher education in creating synergies between actors in the innovation ecosystem, thereby contributing to social and economic growth.

This paper outlines an innovation and entrepreneurship education management platform. It mainly focuses on three parts: constructing the quality evaluation system of innovation and entrepreneurship training programs for college students, constructing the recommendation system of the information platform, and data visualization to study and analyze the data accumulated over the years of operation of the platform in colleges and universities, and explore the correlation between students' innovation and entrepreneurship practice ability and the function of the platform. students' satisfaction with the innovation and entrepreneurship education platform is investigated through the distribution of questionnaires through the online platform. Find the hierarchical relationship of each indicator according to the experience of experts, construct the hierarchical structure of the performance evaluation system of innovation and entrepreneurship education, embed the initiatives represented by each indicator into the process implementation of the innovation and entrepreneurship information management platform, quantify the weights of each indicator, and realize the distribution of scores according to the weights of the quantified indicators. Combining the natural language processing related technology with the collaborative filtering recommendation algorithm, the keywords of the recommended content are determined by subwording the recommended content, so as to realize the intelligent recommendation of learning materials related to innovation and entrepreneurship education.

2

Construction of Innovation and Entrepreneurship Education Platform System in Colleges and Universities

1)

By obtaining multi-dimensional innovation and entrepreneurship project process management data about on-campus students, on-campus and off-campus instructors, on-campus and off-campus review tutors, and school and college level administrators in the personnel management module recorded on the platform, we can track the whole process of the implementation of innovation and entrepreneurship projects and solve the problem of the lack of quality evaluation of the process of innovation and entrepreneurship projects by specifying the evaluation indexes during the process of the innovation and entrepreneurship project implementation. Thinking from the two dimensions of project management quality and project education quality, and starting from the four main subjects of managers, students, instructors, and evaluation tutors, we rely on the information management platform of innovation and entrepreneurship projects to increase the depth of mining process data, so that the process data provides support for quality evaluation.

2)

The implementation data on the platform are analyzed by means of mathematical analysis and other means, and data visualization is used to realize the data display of various aspects such as the results of the project theme classification, the results of the project source classification, the distribution of project faculties, the distribution of the number of participants' faculties, the change of the project scores, the distribution of the faculties of the excellent projects, and the ranking of the real-time excellent projects, and so on.

3)

Utilize Chinese word separation algorithm to complete keyword screening based on the title and introduction information of the project, and then realize the theme classification of innovation and entrepreneurship projects based on the screened keywords. According to the keyword classification results, the improved collaborative filtering algorithm is used to make recommendations for the new projects, including team recommendation module, project recommendation module, evaluation recommendation module, learning material pushing module, etc.

3

Realization of the innovative entrepreneurship education platform system

3.1

Intelligent Quality Evaluation of the Dual Innovation Project

Comprehensively analyzing the whole process of the implementation of innovation and entrepreneurship training program for college students, students, as one of the important subjects of the quality evaluation of the Da Venture program, and at the same time, the administrators involved in the management and guidance, the on- and off-campus instructors, the review tutors and other trainers should also be the main part of the construction. Considering the characteristics of multi-dimensional and multi-criteria of the quality evaluation system of the Da Chuang project, this paper, when constructing the hierarchical structure model for each subject, analyzes the accessibility of the evaluation indexes of the Da Chuang project by investigating the process data on the statistical platform, and breaks down the evaluation indexes into multi-level indexes such as the indexes of the target level, the indexes of the intermediate level, and the indexes of the program (bottom) level, etc., in order to go deeper and more detailed step by step. The hierarchical analysis method is used to obtain the relative weight values of the program (bottom) layer of the indicator evaluation system, and the allocation of indicator scores is realized on the platform. The final quality evaluation system is thought of in two dimensions: program management quality and student education quality.

Assuming that Indicator A is linked to the lower level Indicator B₁,B₂,B₃,……B_n, it is necessary to construct a B-level Importance Matrix to obtain the relative weights of each Indicator in Level B, and then to obtain the final weights of each Indicator in Level B based on the weights of Indicator A. The weights of the indicators at the program level are constructed layer by layer: (1) $B = {(b_{i j})}_{n * n} = [\begin{matrix} b_{11} & b_{12} & \dots & b_{1 n} \\ b_{21} & b_{22} & \dots & b_{2 n} \\ \dots & \dots & \dots & \dots \\ b_{n 1} & b_{n 2} & \dots & b_{n n} \end{matrix}]$

where b_ij*b_ji = 1.

The maximum eigenvalue and eigenvector of matrix B are solved by sum-product method. Where the eigenvector is the final relative weight value of the indicators of the layer, and the eigenvector is used for consistency test, the main steps of solving are as follows: 1)

Normalize the elements of judgment matrix B by columns to get matrix C = (c_ij)_n*n, where: (2) $c_{i j} = \frac{b_{i j}}{\sum_{i = 1}^{n} b_{i j}}, (i, j = 1, 2, \dots \dots, n)$

2)

Sum the elements of matrix C by rows to obtain vector D = (d₁,d₂,……, d_n)^T.

3)

Normalize vector D to obtain eigenvector W = (w₁,w₂,……,w_n)^T: (3) $w_{1} = \frac{d_{i}}{\sum_{k = 1}^{n} d_{k}}, (i = 1, 2, \dots \dots, n)$

4)

Find the largest characteristic root λ_max: (4) $λ_{\max} = \frac{1}{n} * \sum_{i = 1}^{n} \frac{{(B W)}_{i}}{w_{i}}$

5)

Consistency test index CR: (5) $C R = \frac{λ_{\max} - n}{R I (n - 1)} < 0.1$

where RI is the same-order average stochastic consistency indicator. If CR is less than 0.1, it can be regarded as passing the consistency test. If the consistency test cannot be passed, each relative importance ratio in matrix B needs to be adjusted until matrix B passes the consistency test.

3.2

Resource Recommendation Based on Collaborative Filtering Algorithm

In this paper, we use cosine similarity to calculate the similarity between the projects, after obtaining the similarity between the projects, the User-basedCF algorithm will recommend N most similar previous projects for each project, obtain the review mentors in these previous projects, and measure the degree of match between the review mentors and the projects to be recommended by using Equation (6): (6) $p (u, i) = \sum_{v \in s (u, K) \cap N (i)} w (u, v)$

where S(u,K) contains the K items that are most similar to the topic of item u, N(i) is the collection of previous items reviewed by review tutor i, and wuv is the topic similarity of items u and ν.

(7)

t f_{i, j} = \frac{n_{i, j}}{\sum_{k} n_{k, j}}

Where TF refers to the word frequency of the current document, the numerator represents the number of times the word occurs in a given document, the denominator represents the sum of the number of times all the keywords appear in the document, and IDF (Inverse Word Frequency) refers to the measure of the prevalence of a given word, which is calculated as shown below: (8) $i d f_{i} = \log \frac{| D |}{| {j t_{i} \in d_{j}} |}$

The fraction within log, the numerator represents the number of documents in the document set, and the denominator represents the number of documents that contain the current keyword, for which the fraction is taken logarithmically to get, the value of the IDF for the current vocabulary. The weight TF of the keyword is: (9) $T F - I D F (i, j) = t f_{i, j} \cdot i d f_{i} = w_{i, j}$

This algorithm evaluates the importance of the keywords in each item and is used to improve the similarity matrix in the collaborative filtering algorithm described above. Using the cosine formula on this similarity matrix, the similarity of two items p_m, p_n can be calculated: (10) $s i m (p_{m}, p_{n}) = \frac{p_{m} \cdot p_{n}}{‖ p_{m} ‖ \times ‖ p_{n} ‖} = \frac{\sum_{t = 1}^{j} w_{t, m} \cdot w_{t, n}}{\sqrt{\sum_{t = 1}^{j} w_{t, m}^{2}} \sqrt{\sum_{t = 1}^{j} w_{t, n}^{2}}} = w (u, v)$

The recommendation process is shown in Figure 1. After semantic enhancement of each project to obtain the keywords, when the new batch of DaCreative projects is opened, when recommending the evaluation tutors for the new projects, it is only necessary to obtain the K previous projects with the highest similarity of the current project according to the cosine similarity formula mentioned above, and then make the optimal tutor recommendation for the newly applied projects with the help of the improved collaborative filtering algorithm according to the evaluation tutor data of the previous projects.

3.3

Chinese Named Entity Recognition Study

3.3.1

Smoothing evaluation methods

One of the criteria for evaluating how good a model is the degree of confusion. First, look at the definition of the degree of confusion: (11) $p_{e} = 2^{H}$

where $H = - \frac{1}{L - n + 1} \sum_{i = n}^{L} \log_{2} P (x_{t} | x_{i - n + 1})$ .

x₁,x₂,……x_n is the sequence of words tested, n is the number of elements of the model, and L is the length of the test text. For the HMM model: (12) $H = - \frac{1}{L - 1} \sum_{i = 2}^{L} \log_{2} P (x_{t} | x_{i - 2})$

By mathematical derivation, the following equation is obtained: (13) $\frac{1}{\max (p (x_{t} | x_{i - 1}))} \leq p_{e} \leq \frac{1}{\min (p (x_{t} | x_{i - 1}))}$

It can be seen that if the data is smoother, the smaller the value of Chaos p_e will be and the model will be better. Thus there is a strong relationship between how good the model is and how smooth the data is.

3.3.2

Parametric Result Smoothing Methods

Since the lexical and lexical probabilities of the improved HMM model are both three-dimensional matrices and the data sparsity problem is more serious, this paper adopts a linear interpolation method with stable performance to smooth the lexical probability A and lexical probability B respectively.

For lexical transfer probability: (14) $a_{i j k} = p (o_{k} | o_{i} o_{j}) = λ * P (o_{k} | o_{j} o_{i}) + (1 - λ) P (o_{k} | o_{j}) = λ * \frac{N_{i j k}}{N_{i j}} + (1 - λ) * \frac{N_{j k}}{N_{j}}$

The sum of the interpolation coefficients is equal to 1, which ensures that the result is a valid probability value, and again it is clear that such a relationship exists $\sum_{o_{k} = O} a_{i j k} = 1$ . i.e.: (15) $b_{i j k} = p (x_{k} | o_{i} o_{j}) = λ * P (x_{k} | o_{j} o_{i}) + (1 - λ) P (x_{k} | o_{j}) = λ * \frac{N_{i j x_{k}}}{N_{i j}} + (1 - λ) * \frac{N_{j x_{k}}}{N_{j}}$

The basic idea of the algorithm is that the training corpus is divided into two parts, where the larger part is used to train the model parameters, which is called reservation data. And the remaining part is called the support data, which is used to estimate the interpolation coefficients and find the coefficient λ to maximize the probability of being generated from the support data by the corresponding interpolation model. The importance of cross entropy lies in the difficulty of text recognition using the model, or from a compression perspective, the average number of bits used to encode each word. In general, the higher the model's probability of predicting the data, the lower the cross-entropy. Using this principle, the cross-entropy of the supporting data is used as the optimization scale factor for the iterative calculation of the interpolation coefficient optimization, and the algorithm is as follows: 1)

Calculate the corresponding probabilities P(o_k|o_jo_i) and P(o_k|o_j) from the majority of the retained data.

2)

Count the corresponding probabilities of occurrence in the support data q = q(o_k|o_jo_i).

3)

Calculate the cross entropy for the support data: (16) $H_{0} = - \sum q \log p$

4)

Set the initial value 0≺λ≺1 for the interpolation coefficients and the initial value λ¹ = 0.97 for this paper.

5)

For the current interpolation coefficients λ^m(m≥1), calculate the cross entropy of the interpolation model for the supporting data H_k: (17) $H_{k} = - \sum q \log p = - \sum q \log (λ^{m} p (o_{k} | o_{j} o_{i}) + (1 - λ^{m}) p (o_{k} | o_{j}))$

6)

Predict the interpolation coefficients for the next iteration according to the following equation: (18) $λ^{m + 1} = \frac{H_{m}}{H_{m - 1}} λ^{m}$

7)

If $| \frac{λ^{m + 1} - λ^{m}}{λ^{m}} | \leq ε$ (ε = 0.0005 is a predetermined threshold), then λ^m is the desired value and the algorithm stops. Otherwise, repeat steps 5 through 7. Finally the interpolation coefficient λ = 0.974 is calculated on top of the set of markers.

3.3.3

Description of the Viterbi algorithm parameters

Let the given sentence S be cut into word strings S = x₁,x₂,……x_m(x_t ∈ S). For ease of processing, assume a first word x_h and a last word x_e before the beginning and at the end of the sentence, respectively, whose lexical ordinal numbers are set to 0. At this point the word strings: (19) $S = x_{h} x_{1} x_{2} \dots \dots x_{m} x_{e}$

Scan the syncopated sequence of sentence S word by word from left to right to find the subword strings containing entities: (20) $x t^{1}, x t^{2} \dots \dots x t^{v} 0 \leq v ≺ m$

Let any subword string xt^v be: $x_{0}^{v} x_{1}^{v} \dots \dots x_{m}^{v} \dots \dots x_{n}^{v}$ . where $x_{0}^{v}$ and $x_{n}^{v}$ are boundaries of the entity word string, both lexically determined words.

Of the various possible paths (i.e., the various possible strings of lexical tokens) from the lexeme of $x_{o}^{v}$ to the j th lexeme o_j of $x_{n}^{v}$ and the i th lexeme of $x_{m - 1}^{v}$ marking the lexeme o_i of $x_{n}^{v}$ , there must be a path that maximizes the probability of δ_m the lexical tokens, which can be notated in terms of Viterbi variables: (21) $δ_{m} (i, j) = \max_{t_{1} \dots \dots t_{m - 2}} p (o_{j}, o_{i}, x t^{1}, x t^{2} \dots \dots x t^{v}) 2 \leq m \leq n, 1 \leq i, j \leq N$

The probability of the entire path when the state of the HMM is transferred from $x_{n}^{v}$ to $x t_{m + 1}^{v}$ can be found by the maximum probability of the HMM at the previous state, i.e., the Viterbi algorithm can be valued recursively: (22) $δ_{m} (j, k) = \max_{1 \leq i \leq N} [δ_{m - 1} a_{i j k}] b_{i j k} (X_{m}^{v}) 2 \leq m \leq n 1 \leq i, j, k \leq N$

When scanning over $x_{n}^{v}$ state transfer to $x t_{m + 1}^{v}$ , a variable is needed to record which of the paths already traveled is the best. That is, the best lexical annotation for x_m is remembered, and this variable is denoted: (23) $Δ_{m} (i, j) = \arg \max_{1 \leq i \leq N} [δ_{m - 1} (i, j) a_{i j k} b_{i j k} (X_{m}^{v}) 2 \leq m \leq n 1 \leq i, j, k \leq N$

3.3.4

Implementation of Improved Viterbi Algorithm

1)

Initialization: (24) $δ_{1} (i, j) = π_{i} b_{i j k} (X_{1}^{v})$

where the ordinal number of the $X_{0}^{v}, X_{1}^{v}$ lexical is i,j,1≤j≤N, respectively: (25) $Δ_{1} (i, j) = 0, 1 \leq i, j \leq N$

2)

Recursive computation: (26) $δ_{m} (j, k) = \max_{1 \leq i \leq N} [δ_{m - 1} a_{i j k}] b_{i j k} (X_{m}^{v}) 2 \leq m \leq n 1 \leq i, j, k \leq N$ (27) $Δ_{m} (i, j) = \arg \max_{1 \leq i \leq N} [δ_{m - 1} (i, j) a_{i j k}] b_{i j k} (X_{m}^{v}) 2 \leq m \leq n 1 \leq i, j, k \leq N$

3)

When the last word (X_n) is reached, calculate the best lexical marker for this word: (28) $P^{*} = \arg \max_{1 \leq i, j \leq N} [δ_{m} (i, j)]$ (29) $T^{*} = \arg \max_{1 \leq i, j \leq N} [δ_{m} (i, j)]$ (30) $T_{n - 1}^{*} = \arg \max_{1 \leq i, j \leq N} [δ_{m} (i, j)]$

4)

Starting with the best lexical annotation for the last word X_n, the best lexical annotation for each word is obtained in turn: (31) $T_{n}^{*} = Δ_{n + 2} (T_{n + 1}^{*}, T_{n + 2}^{*}) n = n - 2, n - 3, \dots, 2, 1$

There are N lexical tokens in the lexical annotation and M words given. Considering the worst case, when scanning to each word, there are N³ paths from the individual lexical tokens of the two words preceding the current word to the individual lexical tokens of the current word, and the number of times the scanning of the true word string is computed is M N³ summed up, i.e. N³*M.

For a deterministic lexical labeling system, the set of lexical tokens is deterministic. Therefore, the computation time grows in a linear fashion as the M length increases. That is, the computational complexity is also linear for the improved Viterbi algorithm, which is due to the fact that this paper extends it for the change of model parameters and does not change the idea of dynamic programming of the Viterbi algorithm.

4

Implementation and analysis of the innovation and entrepreneurship practice platform system

4.1

Survey and analysis of innovation and entrepreneurship practice ability

A questionnaire survey was conducted among the students of University Z, mainly through questionnaire distribution. A total of 500 questionnaires were actually distributed, of which the number of valid questionnaires was 482, with an effective rate of 100%. The results of the questionnaire survey are shown in Table 1. In the importance score of innovation and entrepreneurship ability, personal quality score (4.105±0.852) is the highest, followed by team ability (4.105±0.877), entrepreneurial awareness (4.007±0.885), and personal ability (3.971±0.848).

Table 1

Mean index

Name	Sample size	Minimum value	Maximum value	Mean value	Standard deviation	Median
Entrepreneurial consciousness	500	1	5	4.007	0.885	4.148
Personal ability	500	1	5	3.971	0.848	4.005
Team ability	500	1	5	4.052	0.877	4.172
Personal quality	500	1	5	4.105	0.852	4.172

4.2

Functional elements of the education platform

The results of the analysis of innovation and entrepreneurship education platform factors are shown in Table 2, occupying the top five places are resource sharing (mean 4.144, standard deviation 0.922), entrepreneurial enthusiasm (mean 4.127, standard deviation 0.956), innovation and entrepreneurship guidance (mean 4.059, standard deviation 1.02), timeliness of innovation information (mean 4.048, standard deviation 1.009), and course training mode (mean 4.017, standard deviation 0.999).

Those occupying the bottom five places are: school-enterprise cooperation (mean value is 3.77, standard deviation is 1.213), cooperation mechanism (mean value is 3.779, standard deviation is 1.098), enterprise practice cooperation (mean value is 3.791, standard deviation is 1.216), safeguard conditions (mean value is 3.818, standard deviation is 1.111), entrepreneurship incubation base construction (mean value is 3.839, standard deviation is 1.125).

Table 2

Innovation and entrepreneurship platform factors

Name	Mean value	Standard deviation	Total
Resource sharing	4.144	0.922	1
Entrepreneurship	4.127	0.956	2
Innovative entrepreneurship guidance	4.059	1.02	3
Innovation information timeliness	4.048	1.009	4
Course training mode	4.017	0.999	5
Industry development demand	4.013	1.001	6
Information flow	4.005	1.05	7
Platform construction perfection	3.986	0.995	8
Process recording mode	3.961	1.091	9
Information update speed	3.959	1.068	10
Platform management rules and regulations	3.922	1.076	11
Message reminder mode	3.912	1.06	12
Innovative resources	3.905	1.129	13
Interactive frequency	3.872	1.139	14
External tutor cooperation	3.862	1.105	15
Information release	3.847	1.117	16
Incubation base construction	3.839	1.125	17
Guarantee condition	3.818	1.111	18
Enterprise practice cooperation	3.791	1.216	19
Cooperative mechanism	3.779	1.098	20
School cooperation	3.77	1.213	21

4.3

Correlation analysis between student competencies and platform functionality

Correlation analysis is used to study the correlation between entrepreneurial awareness, personal ability, team ability, personal quality and functional experience, platform management, innovation platform operation mechanism, user experience, and the correlation coefficient is used to indicate the strength of the correlation. Figure 2 displays the results of the coefficient analysis.

Entrepreneurial awareness and functional experience, platform management, innovative platform operation mechanism, user actual experience between a total of four items all show significance, the correlation coefficient values are 0.846, 0.557, 0.693, 0.813, and the correlation coefficient values are greater than 0, meaning that entrepreneurial awareness and functional experience, platform management, innovative platform operation mechanism, user actual experience between a total of four items have a positive correlation between entrepreneurial awareness and

The correlation coefficient values were 0.859, 0.557, 0.687, and 0.837, respectively, and the correlation coefficient values were greater than 0, which means that there was a positive correlation between personal ability and functional experience, platform management, innovative platform operation mechanism, and user actual experience.

The correlation coefficient values were 0.88, 0.596, 0.73, and 0.873, respectively, and the correlation coefficient values were greater than 0, which means that there was a positive correlation between team ability and functional experience, platform management, innovative platform operation mechanism, and user actual experience. The correlation coefficient values were 0.849, 0.572, 0.692, and 0.829, respectively, and the correlation coefficient values were greater than 0, which meant that there was a positive correlation between personal quality and functional experience, platform management, innovative platform operation mechanism, and user actual experience.

The relevant relationship shows that the various functions of the practice platform are closely related to the effects of the students' experience. Further improve the functional module and help the further development of innovation and entrepreneurship practice.

4.4

Platform performance evaluation indicator system

The performance evaluation index system of innovation and entrepreneurship education platform is shown in Table 3, which builds a relatively clear hierarchy and a performance evaluation index system of innovation and entrepreneurship education platform in colleges and universities with a clear structure. The first layer is the target layer of the final evaluation, i.e., the final result of the performance evaluation of university innovation and entrepreneurship platform, which reflects the role and interrelationship of the seven secondary indicators in the performance evaluation of university innovation and entrepreneurship education platform, as well as the degree of their influence. The first layer consists of internal factors (A1) and external factors (A2). The second level is the secondary indicator layer (B), with seven indicators of regional culture (B1), innovation and entrepreneurship instructors (B2), relying universities (B3), local government (B4), alumni resources (B5), service recipients of the platform (B6), and resources of science and technology park enterprises (B7). The third level is the tertiary indicator layer (C), which consists of 15 indicators representing influencing factors that specifically form the main part of the secondary indicator layer. The realization of students' innovative consciousness (C1), the degree of desire for life value (C2), the composition of teachers (C3), the proportion of the number of scientific research projects under the responsibility of teachers (C4), the number of innovation and entrepreneurship education courses (C5), the teaching environment supporting hardware (C6), financial support (C7), policy guidance (C8), the number of available alumni enterprises (C9), the number of alumni entrepreneurship mentors (C10), the regional number of majors per capita (C11), number of regional high-tech enterprises (C12), number of regional colleges and universities (C13), volume of innovative projects of industry-university-research cooperation (C14), and rate of transformation of scientific and technological achievements (C15). According to the data in the table, it can be seen that the top three weight coefficients of the secondary indicator layer are B6 (0.547), B2 (0.49) and B5 (0.29). In the innovation and entrepreneurship education platform of colleges and universities, each functional department can query the required knowledge and information according to the design of the authority, and the well-established network information platform can also provide specific targeted services.

Table 3

Innovation start-up platform performance evaluation index system

Primary indicator	Weight	Secondary indicator	Weight	Composite weight	Tertiary index	Weight	Composite weight	Order
A1	0.45	B1	0.16	0.072	C1	0.244	0.018	15
A1	0.45	B1	0.16	0.072	C2	0.756	0.054	7
		B2	0.49	0.2205	C3	0.760	0.168	1
					C4	0.240	0.053	8
		B3	0.22	0.099	C5	0.650	0.064	5
					C6	0.350	0.035	11
		B4	0.13	0.0585	C7	0.420	0.025	14
					C8	0.580	0.034	12
A2	0.55	B5	0.29	0.1595	C9	0.630	0.100	4
		B5	0.29	0.1595	C10	0.370	0.059	6
		B6	0.547	0.30085	C11	0.520	0.156	2
					C12	0.110	0.033	13
					C13	0.370	0.111	3
		B7	0.163	0.08965
		B7	0.163	0.08965	C14	0.470	0.042	10
					C15	0.530	0.048	9

4.5

Satisfaction with the platform system

Satisfaction with information-based innovation and entrepreneurship education is subjectively evaluated by evaluating the degree of satisfaction students receive from that education. The questionnaires were distributed through the online platform and the class group of current students of University Z. 1070 questionnaires were recovered, and the valid questionnaires were 1054 By analyzing the satisfaction survey on the satisfaction with the course quality, the satisfaction with the teachers, the satisfaction with the practice platform, the satisfaction with the policy environment and the overall satisfaction, the correlation coefficients were obtained as shown in Table 4, and the correlation coefficients between all the variables were in the range of 0.3-0.4 and show significance at the level of 0.01, and there is a significant positive correlation between overall satisfaction and satisfaction with course quality, faculty satisfaction, practice platform satisfaction, and policy environment satisfaction.

Table 4

Correlation coefficient

	Validity	Overall satisfaction
Quality satisfaction	0.933	0.371^***
The strength of the teacher's strength	0.898	0.306^***
Practice platform satisfaction	0.906	0.338^***
Policy atmosphere satisfaction	0.92	0.344^***

5

Conclusion

This paper applies the hierarchical analysis method, the collaborative filtering algorithm, and the Chinese named entity recognition algorithm to the innovation and entrepreneurship information management platform in universities, constructs the innovation and entrepreneurship project quality evaluation system, and realizes the visualization of the innovation and entrepreneurship education system. The results show that: 1)

The personal quality of college students' innovation and entrepreneurship practice ability has the highest rating, with an average value of 4.105.

2)

The top five functional elements of an innovation and entrepreneurship platform are resource sharing, entrepreneurial enthusiasm, innovation and entrepreneurship guidance, timeliness of innovation information, and course training mode. And there is a positive correlation between students' ability and the functional experience of the platform, platform management, and innovative platform operation mechanisms. The highest weight coefficient of the platform service object in the evaluation system is 0.547.

3)

The overall satisfaction of innovation and entrepreneurship education is high, showing significance at the 0.01 level.

Therefore, building a comprehensive and practical information-based education system for innovation and entrepreneurship is an effective way to cultivate professional and applied talents. The use of information technology as a platform for practice is a suitable option for the comprehensive development of talents in modernization.

Lingua:: Inglese

Frequenza di pubblicazione:: 1 volte all'anno
Argomenti della rivista:: Scienze biologiche, Scienze della vita, altro, Matematica, Matematica applicata, Matematica generale, Fisica, Fisica, altro

Feed RSS della rivista

Construction and Implementation of Innovation and Entrepreneurship Education System in Colleges and Universities under the Background of Information Technology

Hong Qian

Pubblicato online: 17 mar 2025

Ricevuto: 29 ott 2024

Accettato: 16 feb 2025

DOI: https://doi.org/10.2478/amns-2025-0243

Parole chiaveInnovation and entrepreneurship education platform, Evaluation index, Collaborative filtering, Hierarchical analysis

© 2025 Hong Qian, published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Parole chiave
Innovation and entrepreneurship education platform, Evaluation index, Collaborative filtering, Hierarchical analysis