Intelligent Design and Implementation of Talent Cultivation Models Related to Employment Innovation and Entrepreneurship in Big Data Environment

Employment, innovation and entrepreneurship are eternal topics in human society. Among them, innovation and entrepreneurship are independent of each other as well as intertwined with each other [1]. In the post-industrial era, the status of innovation and entrepreneurship has become more prominent, and the arrival of the networkization and informationization era has brought infinite space and possibilities for employment, innovation and entrepreneurship.

Employment talent cultivation mode mainly focuses on cultivating personnel’s practical operation ability and vocational technology, and is mainly characterized by providing professional knowledge and practical training opportunities that are closely integrated with occupational needs [2-4]. Innovation and entrepreneurship talent training mode mainly emphasizes the cultivation of personnel’s innovative thinking and entrepreneurial ability, and encourages personnel to transform their ideas into practice on campus or in social practice [5-8]. Schools and society should provide corresponding innovation and entrepreneurship platforms to guide personnel to cultivate the awareness and ability of innovation and entrepreneurship. Innovation and entrepreneurship is conducive to alleviating social employment pressure, increasing entrepreneurial enterprises and employment positions, and the talents cultivated through innovation and entrepreneurship have a certain degree of competitiveness in the job market [9-11]. Innovation and entrepreneurship training will reform “memory training”, promote “creativity training”, and reverse the situation of passive employment of talents. This will not only help the talents themselves to establish enterprises smoothly and realize their self-worth, but also increase social jobs through entrepreneurship, which is of great significance in alleviating the pressure of social employment and promoting the development of China’s entrepreneurial economy and the construction of an innovative country [12-13]. Cultivating innovative and entrepreneurial talents in terms of physical and mental, intellectual, sensitivity, aesthetic consciousness, personal responsibility, spiritual values and other aspects of the development of talents, so that through the cultivation of talents can form an independent, critical spirit of ideology, keen ability to judge [14-16].

As a new talent cultivation concept, innovation and entrepreneurship is the bridge of modern science and technology ultimately transformed into real productivity, and it is the choice of the times to cultivate and catalyze the new power of economic and social development [17-19]. Therefore, we should continuously strengthen the innovation and entrepreneurship consciousness and ability of talents through innovation-centered innovation and entrepreneurship training, make innovation and entrepreneurship become the common behavioral habits and value pursuit of the society, and continuously improve the ability and level of supporting China’s innovation-driven, transformation and development.

This paper proposes personalized recommendations of employment innovation and entrepreneurship resources using knowledge graphs for talent cultivation.Adopt the relationship extraction method based on machine learning to extract knowledge of employment innovation and entrepreneurship resources from the Internet.The triad vector and feature matrix of entities, attributes, and relationships are calculated, and the semantic matching matrix is constructed based on the knowledge graph.Based on the constructed semantic matching matrix, the employment innovation and entrepreneurship resource interest model has been established.The KG-NCF model contains a recommendation module, cross-connection module, and knowledge expression module.Personalized recommendations for employment innovation and entrepreneurship resources can be achieved by setting personalized recommendation parameters.

2

Intelligent design of talent training models related to employment innovation and entrepreneurship

2.1

Application of big data in the cultivation of employment innovation and entrepreneurship talents

In the era of big data, the application of big data technology to reform the employment innovation and entrepreneurship-related talent training mode can more accurately grasp the market demand, while combining the characteristics of the students, complete the reconfiguration of educational resources, and avoid the drawbacks of traditional education that focuses only on the transmission of theoretical knowledge. For example, with the use of big data analysis tools and online learning platforms, students’ learning progress can be tracked and fed back in real time, so that the application of teaching strategies can be effectively adjusted. In the use of artificial intelligence, virtual reality and other advanced technologies, a more realistic learning environment can be created for students, which can make the application of students’ professional knowledge and skills more effective in practice. It is worth mentioning that the use of big data technology can effectively manifest educational equity. Through the analysis of students’ background information and learning data, it is possible to recommend more personalized resources and learning programs for different students, which can ensure that all students can find suitable educational opportunities for themselves, but also allow students to identify their own strengths in what areas, so that students have a skill [20].

2.2

Intelligent design of talent training mode

This chapter uses big data technology to intelligently design the talent cultivation model, which focuses on personalized recommendations for employment innovation and entrepreneurship education resources.

2.2.1

Extracting knowledge of employment innovation and entrepreneurship resources

The study uses machine learning-based relationship extraction to extract employment innovation and entrepreneurship resource knowledge from the internet. Relevant lexical, syntactic or semantic features of different employment innovation and entrepreneurship directions are selected to describe various local and global features of the relationship instances, and the unstructured text is classified to complete the relationship extraction [21]. The specific extraction process is shown in Figure 1. Select the resource knowledge of different employment innovation and entrepreneurship directions, set the initial relationship extraction semantics, conduct a comprehensive search on the Internet, analyze the association between the employment innovation and entrepreneurship information and the initial semantics, and after determining the relationship, carry out the corresponding extraction to obtain the essential labels of the relationship that exists between the two. Due to the existence of many repetitive semantic relations in its content, when extracting the resource knowledge of employment innovation and entrepreneurship under each specific category, the phenomenon of the same resource knowledge being repeatedly extracted may occur.

2.2.2

Establishing semantic matching matrix based on knowledge map

Based on the knowledge of employment innovation and entrepreneurship resources extracted above, a semantic matching matrix based on a knowledge graph is established. Set the triad of knowledge graph, i.e. entity, attribute, and relationship. Let the entity of the knowledge graph triad be h, the attribute be r, and the relationship be t, then the vector representation of the knowledge entity in the semantic relationship space is h_r, and the calculation formula is: (1) $h_{r} = h M_{r}$ Where: M_r is the mapping matrix for mapping entities from entity space to relation space r . Adjusting in the relation space, the tail entity vector t_r can be obtained based on the calculated h_r. The loss function in performing semantic matching can be expressed as: (2) $f_{r} (h, t) = {‖ h_{r} + r - t_{r} ‖}_{2}$

Let the number of layers in the hidden layer of the knowledge graph be k, then h_k is the knowledge entity at layer k, which can be obtained by substituting the loss function: (3) $h_{k} = f_{r} (h, t) * k$

Calculated by Eq. (3), h_k . Let the dimension of the feature vector of the knowledge entity and semantic relationship be d. After transpose and multiply we can get the feature matrix T of d×d, and then bring in the number of hidden layers k to get the feature matrix T_k of the kth layer as: (4) $T_{k} = [\begin{matrix} M_{r}^{(1)} h_{k}^{(1)} & \dots & M_{r}^{(1)} h_{k}^{(d)} \\ ⋮ & ⋱ & ⋮ \\ M_{r}^{(d)} h_{k}^{(1)} & \dots & M_{r}^{(d)} h_{k}^{(d)} \end{matrix}]$ Where: the feature matrix T_k is multiplied with the mapping matrix to obtain the semantic matching matrix corresponding to the knowledge entities and semantic relations of dimension d and layer k.

2.2.3

Modeling student interest

Based on the semantic matching matrix, the students’ interest model for employment innovation and entrepreneurship resources is constructed.The model designed in this paper is the KG-NCF model, which takes the feature vectors of the semantic matching matrix as the objective vectors. The framework of the KG-NCF model is shown in Fig. 2. It includes three modules, which are recommendation module, cross-connection module, and knowledge expression module. Among them, the cross-connection module is responsible for extracting the corresponding knowledge of employment innovation and entrepreneurship resources from the database and delivering it to the knowledge expression module.

2.2.4

Setting personalized recommendation parameters

Students input the keywords of employment innovation and entrepreneurship directions they are interested in into the recommendation module, and the cross-linking module realizes the process of personalized recommendation. In this paper, the weighted fuzzy calculation is used to derive the recommendation level α of employment innovation and entrepreneurship resource knowledge, and the calculation formula is: (5) $α = \int_{z > 1}^{z} λ \times δ \times \frac{η}{2} d$ Where: z is the coefficient of weighted fuzzy. λ is the amount of access control. δ is the distribution location of this knowledge entity in space. η is the associated attribute of that resource. After calculating the recommended tier, in the space of this tier, the semantics of employment innovation and entrepreneurship input by students is brought into the semantic matching matrix for calculation to determine the corresponding knowledge of employment innovation and entrepreneurship resources. An affiliation assignment is performed, and the accuracy of the personalized recommendation is determined based on the size of the assignment μ , calculated as: (6) $μ = \sum_{q > 1}^{q} \frac{1}{ω} α$ Where: q is the number of unsolicited recommendation entries for employment innovation and entrepreneurship resource knowledge. ω is the boundary range of employment innovation and entrepreneurship resource knowledge recommendation. Determine the employment innovation and entrepreneurship resource knowledge to be recommended based on the calculated personalized recommendation accuracy. The recommended knowledge of employment innovation and entrepreneurship resources is adapted to the specific needs of students, and the adapted knowledge of employment innovation and entrepreneurship resources is transferred to the knowledge expression framework to complete the personalized recommendation of employment innovation and entrepreneurship resources [22].

3

Analysis of the impact of implementation

3.1

Comparative experiments

A regional pilot was chosen as the experimental site. Before the experiment, select the employment innovation and entrepreneurship education resources collated on campus in the past 6 months as test data, with multiple sources of data, including the development data of various majors and multiple industries in the market. The statistical data samples, which are divided into two categories, namely valuable resources and general resources, are counted and organized as test data in the comparison experiment. The sample data of the comparison experiment is shown in Table 1. The test data contains 130,000 items, while the control data contains 44,000 items.

Table 1.

Compare the sample size of the experiment

Data quantity/bar	Data attribute
174000	Total amount of data
130000	Valuable resource
440000	General resources
130000	Test data
44000	Control data

The probability distribution density of a user clicking or actively retrieving a recommended resource during the real-time recommendation process is calculated as: (7) $P = \sum_{i > 1}^{\infty} x (t) + \frac{1}{\sqrt{2} e^{R {(t + 1)}^{2}}}$ Where: P is the density of the probability distribution of users clicking or actively retrieving recommended resources. x is the number of times the user retrieved. t is the probability effective value interval. R is the number of iterations. i is the number of data mining. To quantify the calculation result P, the effective value range of P is defined as 0~1.0, when the calculation result tends to 1.0, it proves that the higher the probability of users clicking or actively retrieving the recommended resources, which indicates that the resources recommended by the method of this paper have a higher degree of fitness with the resources of the user’s needs.

A user is randomly selected to participate in this experiment, and according to the background record data, the change in the probability distribution density of the user’s active retrieval of recommended resources is shown in Table 2. From the experimental results, it can be seen that with the increase of user browsing time, the user behavior data obtained by the terminal increases, at this time the probability distribution density of the user clicking or actively retrieving the recommended resources increases, when the user browsing time reaches 6s, the probability distribution density of the user clicking or actively retrieving the recommended resources is 98.57%, i.e., the fitness between the recommended resources and the user’s demand for the resources begins to converge to 100%. It can be proved that the employment innovation and entrepreneurship education resources recommendation method based on knowledge graph designed in this paper is effective in practical application, which can improve the fitness between recommended resources and user resources, and provide users with high-quality employment innovation and entrepreneurship education resources.

Table 2.

Experimental results

Users browse web time	1s	2s	3s	4s	5s	6s	7s	8s
Click the recommended resource probability density	18.94%	30.54%	43.46%	57.37%	89.23%	98.57%	98.80%	99.47%

The evaluation index is determined by the degree of difference between recommended resources and user needs.Use the personalized recommendation method of educational resources based on knowledge graphs and the traditional method to recommend educational resources for employment innovation and entrepreneurship to different users. The statistical experiment results are shown in Figure 3. The degree of difference between the recommended resources and user needs of this paper’s method is small, with the degree of difference ranging from 1.41 to 2.63. It can be seen that: the design method of this paper can reduce the difference degree between the recommended resources and user needs, and ensure a high degree of fitness between the recommended resources and user demand resources.

3.2

Evaluation and analysis of learning tools

3.2.1

Survey on the effectiveness of the implementation of educational resource recommendation tools

1)

Purpose of the survey

The survey is to verify the effectiveness of the use of educational resource recommendation tools, and 30 subjects are selected to carry out a survey and research on personalized recommendation tools for educational resources based on knowledge graph.

2)

Subjects of the survey

Thirty learners from different backgrounds were selected as survey subjects for this teaching experiment. In order to improve the correctness and credibility of the results of the questionnaire, a balance was achieved in the selection of personnel from different perspectives. From the point of view of profession, the respondents were involved in computer-related fields, such as network engineering, communication engineering, and educational technology.From an occupational standpoint, there were both fresh graduates and those already employed. From a gender perspective, the ratio of men to women reached 1:1.

3.2.2

Application Effectiveness Survey Results and Analysis

The questionnaire mainly investigates whether the users are able to accept the recommendation of employment innovation and entrepreneurship resources based on knowledge mapping, and determines whether the tool is able to improve the learning efficiency of the learners, reduce the blindness in learning, and achieve the purpose of the study. The questionnaire produces a Likert scale for the learning effect agreement of the recommended tool, which considers three aspects: the effect of using employment innovation and entrepreneurship resources, the effect of recommending the resources, and the effect of using the tool in an integrated way. For each aspect, multiple questions were set for users to answer.

The questions related to the effectiveness of using employment innovation and entrepreneurship resources are as follows:

T1: Being able to accurately display linked learning resources and course system structure.

T2: Can clearly present key knowledge points.

T3: Be able to save learning time.

T4: Enhance learning motivation.

The following questions are set for the effectiveness of resource recommendation:

T5: Reduced time to find relevant resources.

T6: Willing to learn the recommended resources in the recommended list.

T7: Prefers the learning format under this recommendation method.

The questions set in terms of the effect of the combined use of the tool are as follows:

T8: The tool enhances the interest in learning.

T9: The tool is easy to use.

T10: The tool enhances learning.

A total of 30 questionnaires were distributed and all of them were collected.

The effectiveness of the use of knowledge graph-based resources for employment innovation and entrepreneurship is shown in Figure 4. The approval item generally becomes the highest value in this part, and the proportion of the number of people ranges from 50.5% to 53.3%, which shows that there is a high level of satisfaction with the effect of the use of employment innovation and entrepreneurship resources based on knowledge mapping.

The statistics of resource recommendation effect is shown in Figure 5. The experience of the resources is good, and the proportion of those who agree and above is extremely high. 87%, 84.9% and 82.9% of T5~T7 agree and above respectively.

The statistics on the effectiveness of the combined use of the tool are shown in Figure 6. This section examines the users’ evaluation of the overall effectiveness of the tool. The majority of users agreed that the tool was easy to use (87.9%). It increased learning efficiency (86.5%) and interest in learning (85.5%).

3.3

User capacity impact analysis

The results of the research on the impact of user capabilities are shown in Figure 7. 96.3% of the users believe that the personalized recommendation of educational resources based on the knowledge graph has a significant effect on the improvement of professional skills (Q3), and the users practice the experiments on the professional aspects over and over again until they are proficient, so that they can be familiar with them when they go to the enterprise, and they can be used for them.More than 82.5% of users said that they had improved their communication and communication skills (Q2), teamwork awareness and teamwork skills (Q8), and interest in learning and self-directed learning ability (Q10).The educational resources provided will have a great impact on their communication skills, teamwork, and learning abilities.

More than 61.6% of users believe that they have improved their "spirit of challenge and entrepreneurial awareness (Q1)", "organizational management ability (Q4)", "ability to understand the company and identify opportunities (Q6)", "professional ethics (Q9)", "time concept (Q11)", and "comprehensive quality (Q12)". More than 53.2% of users believe that they have improved their social resources and social relationships (Q5) and psychological tolerance (Q7). Among the above survey options, only 53.2%~69.3% of users believe that there is an improvement, which may be because the personalized recommendation of educational resources based on knowledge graph focuses on the cultivation of users’ professional skills, and the cultivation and guidance of other abilities are insufficient.

4

Construction of a guarantee mechanism for talent training in a big data environment

4.1

Building a sound policy support system

In the era of big data, it is necessary to build out a perfect policy support system for employment innovation and entrepreneurship, which has a crucial impact on improving students’ employment innovation and entrepreneurship.The government needs to introduce more specific employment innovation and entrepreneurship support policies, and clarify the objectives of these policies and the implementation path. For example, through the detailed formulation of the student entrepreneurship support plan, a series of preferential policies such as entrepreneurship subsidies and entrepreneurship loans are provided to students, while lowering the entrepreneurial threshold of students and assuming a part of the risk of student entrepreneurship, which is more conducive to radicalizing students to participate in entrepreneurial competition. The Government can set up an entrepreneurial project bank for young students, in order to provide students with a better choice of entrepreneurial projects.

4.2

Strengthen the training and introduction of big data talents

In the era of big data, the introduction and cultivation of big data talents can effectively improve the employment prospects and entrepreneurial skills of students. As the cradle of cultivating talents, only by closely combining the market demand, completing the optimization of employment innovation and entrepreneurship settings, and effectively deepening the development of practical teaching, can we cultivate high-quality talents with good big data thinking and mastering big data technology. In practical implementation, by strengthening the cultivation of big data talents, it can further promote the improvement of students’ employment innovations and entrepreneurship. In this process, all walks of life are actively involved in the cultivation of talents, and the optimization of students’ internship training can be realized under the cooperation of school-enterprise and school-society cooperation, which can make students come into contact with more opportunities for big data practice, and improve their employment innovation and entrepreneurship ability. At the same time, high-level big data talent with good practical experience and innovation ability can further lead and promote the application and development of big data technology. By vigorously introducing such talents, more high-level big data talents can be attracted to join the team of students’ employment innovation and entrepreneurship, so that the students can learn more knowledge and skills from these talents in employment innovation and entrepreneurship, and develop their thinking. In addition, strengthening the cultivation and introduction of big data talent requires the joint efforts of the government, enterprises, and all sectors of society. The government needs to introduce relevant policies to support the cultivation and introduction of big data talents, while it needs to cooperate with enterprises and all walks of life to complete the innovation of talent cultivation mode. Enterprises need to actively participate in the cultivation and introduction of talents to provide students with more practical opportunities and positions. And all sectors of society need to foster an atmosphere of innovation, so as to provide support for students’ employment in innovation and entrepreneurship.

4.3

Optimizing the environment for the application of big data technology in employment innovation and entrepreneurship

In the age of big data, it is necessary to improve students’ employment innovation and entrepreneurial abilities by ensuring that they use and master big data technology through continuous optimization. As a school, the effective construction of big data infrastructure needs to be strengthened, which includes improving the processing capacity of data and improving the security and storage capacity of data. The introduction of high-performance computing clusters and distributed storage systems, for example, can improve data processing and efficiency. At the same time, encryption and access control can ensure the privacy and security of such data. As for the promotion of the integration of big data technology and employment innovation and entrepreneurship education, it is possible to make students combine data to grasp the dynamics of market development as well as the trend of industry development with the references of big data analyzing tools and technologies, so as to further improve the employment innovation and entrepreneurship ability of students. Especially through the simulation of entrepreneurial scenarios, combined with the practice of data analysis, students’ data-driven thinking can be cultivated, and their innovation ability can be developed.In the practical application of employment innovation and entrepreneurship, the use of big data technology can help students effectively master it in practical exercises by building a practice platform and introducing actual projects.Like in the deep cooperation with enterprises, big data project practice can help students learn how to use big data technology to solve practical problems.Such practical experience can not only improve students’ skill level, but also accumulate experience in advance for their future employment in innovation and entrepreneurship.

5

Conclusion

This paper designs a personalized recommendation method of employment innovation and entrepreneurship resources based on knowledge graph to realize the intelligence of employment innovation and entrepreneurship related talent cultivation mode in big data environment.

The method can recommend educational resources with a high degree of adaptability to user demand resources in a short time. Compared with the traditional recommendation method, the difference between the resources recommended by this paper’s method and the user’s needs ranges from 1.41 to 2.91. It shows that the method of this paper can carry out efficient employment innovation and entrepreneurship education resources recommendations.

Most of the users who use this paper’s method for learning agree on the effect of using employment innovation and entrepreneurship resources, resource recommendations, and comprehensively using the tool. 96.3% of the users think that the personalized recommendation of educational resources based on knowledge graph has obvious effect on the improvement of professional skills, and can effectively improve their communication ability, team awareness and learning ability.

The cultivation of talents related to employment innovation and entrepreneurship requires the construction of a perfect policy support system, the strengthening of the training and introduction of big data talents and the optimization of the environment for the application of big data technology in employment innovation and entrepreneurship.

Język:: Angielski

Częstotliwość wydawania:: 1 razy w roku
Dziedziny czasopisma:: Nauki biologiczne, Nauki biologiczne, inne, Matematyka, Matematyka stosowana, Matematyka ogólna, Fizyka, Fizyka, inne

Kanał RSS czasopisma

Intelligent Design and Implementation of Talent Cultivation Models Related to Employment Innovation and Entrepreneurship in Big Data Environment

Guangjin Qin

Data publikacji: 21 mar 2025

Otrzymano: 09 paź 2024

Przyjęty: 03 lut 2025

DOI: https://doi.org/10.2478/amns-2025-0664

Słowa kluczoweBig Data Technology, Knowledge Graph, KG-NCF, Employment Innovation and Entrepreneurship

© 2025 Guangjin Qin, published by Sciendo

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

Słowa kluczowe
Big Data Technology, Knowledge Graph, KG-NCF, Employment Innovation and Entrepreneurship