Fuzzy evaluation model for physical education teaching methods in colleges and universities using artificial intelligence
The section proposes the creation of a multi-feature fuzzy evaluation model that utilizes artificial intelligence to facilitate the implementation of PET techniques. The multi-features of physical education are students’ motor abilities like strength, endurance, speed, flexibility, coordination, agility, and balance. This is achieved by integrating natural language with machine learning through fuzzy control instructions. Finally, this section outlines developing a teaching evaluation index system for PETCU that utilizes an AI-based approach. The system is constructed using an enhanced cuckoo search optimization technique, a student’s mobility mechanism, and movement vector deconstruction to establish the system architecture. The outcome is a framework for evaluating teaching quality at PETCU. The evaluation framework integrates a three-layered architecture with the ECSO algorithm to assess physical education teaching. It considers multiple features, uses fuzzy control, and evaluates from management, instructor, and student perspectives. The ECSO algorithm optimizes AI model parameters, contributing to intelligent assessments. The system ensures comprehensive evaluation, including data preprocessing, multi-feature assessment, and a structured ECSO workflow. The proposed MFEM-AI framework enhances the precision and consistency of instructional evaluations in physical education.
System model for accessing the quality of physical education teaching at PETCU
The basic system model for the proposed MFEM-AI for accessing the physical education quality at PETCU involves several processes. The MFEM-AI framework employs fuzzy logic, a rule-based system, and membership functions to handle imprecise linguistic variables in teaching evaluations. Fuzzy control instructions guide the evaluation process, allowing for adaptive decision-making and flexibility in interpreting natural language patterns. The approach enhances communication by enabling the system to understand and respond to the nuanced language used by teaching experts, making the evaluation process more intuitive and user-friendly.
Data acquisition and preprocessing
The system collects data related to physical education quality information on students’ motor abilities like strength, endurance, speed analysis, flexibility, coordination, agility, and balance. The data preprocessing technique involves data cleaning and normalization steps in subsequent steps.
Multi-feature assessment
The system assesses multiple features related to physical education, with a focus on student’s motor abilities. These identified features serve as the foundation for the evaluation process.
A fair evaluation of physical education quality necessitates establishing a systematic structure for evaluating indicators. The conventional method for assessing the quality of teaching includes two main categories, namely comprehensive teaching subject matter and comprehensive achievement in school, to establish the instructing index and assessment index. The fundamental aspects of the indications utilized in traditional assessment systems are their feasibility, representativeness, and autonomy. These factors serve as the foundational elements for the types above of traditional assessment system indications. The proposed MFEM-AI is executed through a process that involves integrating natural language and fuzzy logic. Teaching experts assign ratings to parameters, and the improved cuckoo search algorithm optimizes these ratings. The model’s efficiency is ensured through a carefully designed system architecture. MFEM-AI consistently outperforms traditional methods in evaluating physical education teaching, scoring high across various assessment categories, including skill performance, learning progress, physical fitness, participation rate, student satisfaction, and overall teaching efficiency.
System architecture
The system architecture is designed to manage data flow in three types of layers efficiently and executes the model with feedback. The architectural components include the data management layer, called the feature layer, the indication layer, and the AI evaluation layer, called the target layer. The ECSO is the core of the architecture for evaluating the optimization procedures for fine-tuning the fuzzy parameters for optimal solution of physical education quality.
The creation of the system architecture considers functional criteria related to evaluating physical education methods, including data acquisition and preprocessing, multi-feature assessment focusing on student’s motor abilities, a systematic structure for evaluating indicators, and a three-layered architecture comprising a data management layer (feature layer), an indication layer, and an AI evaluation layer (target layer). The architecture integrates the ECSO algorithm for fine-tuning fuzzy parameters and optimizing the assessment model. The system aims to enhance reliability and efficiency in evaluating teaching quality at colleges and universities.
This study constructs a multi-index evaluation system for PE in colleges using the analytic hierarchy process (AHP) methodology. The evaluation method depends on four dimensions: teaching method, teaching content, teaching mentality, and teaching impact. The layered architecture is displayed in Fig. 1.
The layered architecture of the proposed MFEM-AI.
The procedure for implementing the Enhanced Cuckoo Search Optimization (ECSO) based quality of instruction evaluation approach for sports departments in universities is illustrated in Fig. 2. Initially, the database sourced from the evaluation system for physical education (PE) teaching quality in tertiary institutions is partitioned into two distinct sets, namely the sets for training and testing, with a ratio of 4:1. The present study employs a joint nerve-tailored Enhanced Cuckoo Search Optimization (ECSO) approach, as proposed in the literature, to construct a model for evaluating physical education instruction in higher education institutions. The optimal setting of weights and limits is achieved through joint nerve optimization. To conduct testing, the ECSO model should be utilized for insertion. The subsequent text presents a comprehensive, sequential analysis of the execution procedure.
Cuckoo search optimization workflow.
-
1.
Stage 1: The data about the PE ratings of colleges and universities ought to be scrutinized academically. Specifically, it should be partitioned into learning and training sets, with a ratio of 4:1, and subsequently normalized using Eq. (1).
$$O_x+1=T_h+\fraco-o_mino_max-o_min+T_f-T_h$$
(1)
The available datasets consist of \(O_x+1\), which denotes the normalized information and \(o_min\) and \(o_max\), which respectively indicate the minimum and highest values visible in the \(O_x+1\) dataset. The dataset that has been normalized exhibits a range of values between the most negligible values (\(T_f\)) and the highest value (\(T_h\)), with \(T_h\) being assigned a value of -1 and \(T_f\) being assigned a value of 1.
-
2.
Stage 2: Initiate the primary variables of the joint nerve approach, including h and f, \(S_max\), P and Q. The first step is to set up the structure of the fuzzy system, followed by the initialization of the weighting and sensitivity.
-
3.
Stage ****3: Eq. (2) is utilized to initiate the joint nerve method group, followed by applying the Fuzzy model to obtain weights and limits. This ensures that all individuals in each population are initialized at a uniform starting point.
$$P_x=T_f(x)+rand\left(\mathrm0,1\right)\left(g_f(x)-T_f(x)\right)$$
(2)
The initial value of a person is denoted as \(P_x\), while the upper and reduced search boundaries of the x person are represented by \(T_f(x)\) and \(g_f(x)\), accordingly.
-
4.
Stage 4: Compute \(O_1\) and \(O_2\) for the joint section coefficients utilizing Eqs. (3) and (4) in the following manner:
$$O_1=c\left(1-\propto \right)+f\propto $$
(3)
$$O_2=c\propto +f\left(1-\propto \right)$$
(4)
The scaling factor is denoted \(\propto \), the coefficients are denoted \(c\), and the function is denoted \(f.\)
-
5.
Stage 5: Compute the optimal fitness value, denoted as ACC, for each population member using Eq. (5).
$$\textmax\left(y,v\right)A=\frac\sum_x=0^NA(x)I$$
(5)
\(A\) represents the mean value of K-fold cross-validation reliability. In contrast,\(A(x)\) denotes the mean accuracy values obtained from K-fold calculations.
-
6.
Stage 6: It is recommended that individual positions be modified to reflect current circumstances accurately. The unique position is denoted in Eq. (6).
$$S^u+1\left(x\right)=S^u\left(x\right)\left|\textsin\left(y_1|-y_2\textsin\left(y_1\right)\right)\right|\left|o_1T^u\left(x\right)-o_2S^u\left(x\right)\right|$$
(6)
The individual output is denoted \(y_1 and y_2\). The fitness function is denoted \(T^u\left(x\right)\), and the score is denoted \(S^u\left(x\right)\). The final updated score is denoted \(S^u+1\left(x\right)\).
-
7.
Stage 7: Compute the fitness value \(A_x+1\) of the calculators and juxtapose it with the optimal fitness value \(A_x+1\) of the preceding generation for the populace constituents whose locations have been altered. To obtain the most up-to-date fitness value, it is imperative to update the optimal fitness value to the present repetition if \(A_x+1\) surpasses \(A_best\). It is recommended that \(A_best\) be modified to accurately reflect the user’s present location while maintaining its static nature in all other aspects.
-
8.
Stage 8: To verify whether a method has satisfied its termination circumstance, it is necessary to examine the present cycle count t and compare it against the maximum allowable value \(R_max\). If the condition is met, it is recommended to cease all ongoing operations and produce an optimal fitness value and an ideal location.
-
9.
Stage 9: Conducting a quality evaluation of physical education pedagogy in higher education institutions, utilizing the results obtained from stage 8’s results.
College PE teaching evaluation index
To enhance the reliability of evaluation outcomes about the quality of physical education instruction in higher education institutions, it is imperative to undertake a comprehensive assessment and analysis of the instructional standards of PETCU. Therefore, this paper examines the PETCU from three evaluation perspectives: the management stage, instructors, and pupils. It is worth noting that each viewpoint emphasizes different aspects of the assessment process.
\(E_1\), from a managerial perspective, the assessment method for physical education instructional excellence in college can be categorized into sub-levels
The evaluation system denoted as \(E_1,\) primarily examines the influence of the fundamental physical education amenities of the college on the standard of instruction as viewed through the lens of the administration. This study undertakes an extensive assessment encompassing various aspects, including the professional competencies of physical education (PE) instructors, denoted as \(E_11\), the comprehensiveness of college-level PE courses, designated as \(E_12\), the capacity for reform in college-level PE instruction indicated as \(E_13\), the adequacy of software and hardware establishments and expenditures on PE instruction characterized as \(E_14\), and the ability to integrate PE instruction within the industry-college-research cooperation framework denoted as \(E_15\).
\(\textE_2\), the present study examines the sub-level assessment structure of college physical education teaching quality from the instructors’ perspective
The \(E_2\) assessment framework primarily examines the effects of implementing the PE instructional process on the overall quality of PE, as perceived by educational professionals. This study undertakes a comprehensive assessment encompassing several dimensions, including the originality and depth of the PE curriculum \(E_21\), the logical and empirical basis of PE lesson plans \(E_22\), the efficacy and flexibility of PE instructional techniques \(E_23\), the variety and cognitive complexity of PE pedagogies \(E_24\), the fulfillment of PE teaching objectives \(E_25\), and the instructional quality and impact of PE classes \(E_26\).
\(E_3\), the present study focuses on assessing college physical education teaching quality through a sub-level evaluation structure from the pupils’ perspective
The \(E_3\) assessment system primarily examines the outcomes of physical education instruction in higher education institutions, focusing on the student context. This study undertakes an in-depth evaluation of various factors, including the success rate of pupils in the \(E_31\) broad sports quality test, the unprecedented rate of pupils in the \(E_32 general\) sports quality examination, the creative capacity of pupils as measured by \(E_33\), the independent learning ability of pupils as measured by \(E_34\), the community serviceability of pupils as measured by \(E_35\), and the rewards received by students in competitive sports as measured by \(E_36\).
Enhanced cuckoo search optimization
A group of elements comprising n units is placed in a D-dimensional space to determine the most efficient solution. The i-th egg’s location is denoted as A, while its velocity is represented by B. Simultaneously, the trajectory of the egg’s motion ought to take into account both the egg’s historical optimal position (\(P_best\)) and the ECSO optimal location for the grouping (\(P_best\)). The ECSO optimizing issue can be solved by determining the feasible fix of the egg location.
The variable x is introduced into the given issue, and the outcomes are contrasted to assess the optimization level achieved by the viable solution. ECSO compatibility is a term commonly employed to describe the advantages and disadvantages of feasible solutions. In instances where the objective work necessitates minimal value, the ECSO optimal solution would be the one that is possible with the lowest egg appropriateness.
Eggs utilize an iterative calculation to modify their velocity and location based on all three tenets of motion. This ECSO process continues until the elements reach the optimum state under the circumstances at hand. The updated optimization result is denoted in Eq. (7):
$$s_x^t+1=w_x^t+k_1q_1\left(p_x^t-i_x^t\right)+k_2q_2\left(p_y^t-i_y^t\right)$$
(7)
The variables, \(y, w, t, k_1\), \(k_2\), \(q_1\), and \(q_2\) are utilized in ECSO. Specifically, \(x\) is a discrete parameter that denotes the number of elements in the swarm, while \(y\) reflects its spatial dimension. The momentum weight is represented by \(w, and t\) is an integer that signifies the number of repetitions. Additionally,\(k_1\) and \(k_2\) are learning variables, and \(q_1\) and \(q_2\) are arbitrary numbers that fall within the range of [0, 1].
The composition of ECSO egg velocity is primarily comprised of three distinct components. The initial component comprises the legacy velocity \(S_x\) from the preceding repetition, a historical account of the egg’s motion. The following element pertains to the ECSO egg’s ideal location \(V_p\), which serves as a point reference point for the egg’s self-experience and guides the optimization procedure. The final component pertains to the group’s ideal location S. The process of sharing ECSO information embodies the benefit of the population, which significantly aids individual eggs in locating the most effective approach.
ECSO stands out among other population methods and smart optimizing methods due to its notable advantage of requiring fewer variables to be selected while maintaining high result reliability. The rate of acceleration is comparatively higher due to the information exchange system. The computerized control system exhibits a higher level of ECSO responsiveness. The employed ECSO algorithm optimizes parameters for assessing students’ physical abilities in physical education. It fine-tunes the factors related to mobility mechanisms, including endurance, speed, flexibility, coordination, agility, and balance, to analyze how well students perform in terms of these physical attributes. Deconstructing the movement vector involves analyzing how each of the historical information, individual solution preferences, and overall population trends contributes to the changes in solution positions, providing insights for optimizing the evaluation process.
The present study outlines the key variables used in the ECSO method, which include the egg size denoted by n, the weight of the inertia represented by w, the learning variables \(k_1\), and \(k_2\), the highest speed \(S_max\), the slowest speed \(S_min\), the iterative number ‘t’, and the egg suitability indicated by \(\sigma \). Equation (8) is used for reducing inertia.
$$w\left(x+1\right)=w_init-\fracw_init-w_lastt_maxt$$
(8)
The variables under consideration include t, which denotes the present iteration count, \(t_max\), which represents the upper limit of the repetition count, \(w_init\), denoting the starting inertia weight; and \(w_last\), representing the ultimate inertia weight for the nest.
Learning factors \(\varveck_1,\varveck_2\)
The learning variables \(k_1\) and \(k_2\) is commonly called accelerated unchanged speed in academic literature. These factors signify the relative significance of the egg’s individual experience and collective participation in the investigation of motion.
A value of \(k_1\) equal to zero indicates that the egg’s personal experience does not contribute to the optimization process of identifying the best possible outcome. Currently, the egg swarm algorithm solely incorporates the notion of a collective, resulting in a quicker convergence rate than the conventional ECSO method. However, its efficacy in addressing intricate problems remains to be determined. Due to the lack of empirical evidence supporting specific eggs, there is a greater likelihood of observing a concentration of all fragments around one or a few extremes. The answer gathered can be a local extreme rather than a global optimum.
In the scenario where \(k_2\) equals zero, there exists a lack of communication regarding group data, and the collective experience of the entire population does not influence the egg’s motion. Currently, the benefits of the population method have yet to be manifested, as all eggs are undergoing individual action, resulting in a sluggish running pace for reaching the nest. Identifying the best answer for the goal function poses a challenging task.
Maximum speed \(\varvecS_\varvecm\varveca\varvecx\)
The variable \(S_max\) denotes the highest attainable velocity of an egg within a D-dimensional space during a single iteration, corresponding to the egg’s most significant possible displacement. If the value of \(S_max\) is excessively high, the ECSO egg should fail to converge on the most effective region. Conversely, if the value of \(S_max\) is excessively low, not only will the solution’s acceleration be impeded, but there is also a possibility of solely identifying the local extrema rather than the optimal global solution. In the D-dimensional area, it is common practice to establish the \(S_max\) value as equivalent to one-fifth of the disparity between the top and bottom limits. The mathematical equation is represented in Eq. (9).
$$S_max=\fracs_max-s_minn$$
(9)
The minimum and maximum solutions are denoted \(s_min\) and \(s_max\). The total number of solutions is denoted \(n\).
Minimum speed \(\varvecS_\varvecm\varveci\varvecn\)
The lowest velocity refers to the shortest distance each egg travels during its movement, guaranteeing the operational speed. Typically, the value assigned to \(S_min\) is the negation of the highest rate, denoted as \(S_max\), i.e., \(S_min=-S_max\).
The number of iterations t
The method of ECSO persistently endeavors to attain the most favorable solution using the iterative modification of eggs. It conventionally concludes the process by establishing the number of repetitions.
Optimization with ECSO
The ECSO algorithm is a metaheuristic optimization technique used to optimize AI model parameters. This involves adjusting parameters like weights, biases, or hyperparameters to improve the model’s performance. The system uses the ECSO technique to optimize and fine-tune the parameters of the AI models. This optimization step is crucial for achieving accurate assessments. Equation (7) shows how ECSO updates egg positions based on velocity and historical information.
ECSO suitability \(\varvec\sigma\)
The practicality of an answer can be succinctly expressed as the magnitude of the discrepancy between the real optimal solution and the idealized optimal solution. The attainment of the optimizing objective is evaluated through the computation of the suitability metric. The standardized variables for calculating nest variables using Eq. (10).
$$Q=\frac22-k-\sqrtk^2-4k\right$$
(10)
The egg velocity is denoted \(as k\). The development and analysis of an assessment system for performance for physical education instruction. The parameters of the ECSO algorithm involve population size \(P_x\) that gives the number of nests the maximum number of iterations, followed by the learning variables \(k_1\) and \(k_2\) to determine the relative importance of individual experience and group participation, the maximum attainable velocity of an egg \(S_max\). Likewise, the shortest distance each egg travels during its movement \(S_min\), the no. of iterations in the optimization process is given as \(t\), and the initial and last inertia weight is represented as \(w_init-w_last\), those matches with the termination criteria are analyzed for ECSO suitability factor \(\sigma \).
This study has constructed a physical education instructing assessment system utilizing the artificial intelligence fuzzy method with the methods and automobiles. The ECSO algorithm is shown in Table 1.
Before instruction, training students on properly utilizing said tools is imperative. Using instruction, learners can acquire comprehension and proficiency in the procedures and stages of the complete process of online course education, which encompasses downloading and setting up of the system, authorization, and authentication, selection of classes, viewing of videos, perusal, and retrieval of textual substances, utilization of message boards for debate, and fulfillment of online tasks and tests. In addition, these training sessions facilitate pupils adapting to teaching and establishing a solid groundwork for the seamless progression of their subsequent instructing endeavors.
ECSO algorithm workflow for optimizing the assessment model
The system utilizes the ECSO algorithm to optimize the assessment models. This includes setting parameters, evaluating fitness functions, and updating the next positions for optimal evaluation. The algorithm’s initialization process starts with the parameters like population size, maximum iterations, and learning variables. The velocity and position update is performed for each egg in the nest. It optimizes these parameters to approach the best solution. The algorithm also includes a termination criterion, such as reaching a maximum number of iterations or achieving a satisfactory solution. The best solutions found during the optimization process are recorded for further assessment of physical education.
The methodological analysis presented above indicates that using the ECSO model enhances the CS algorithm. The optimal input weights and limits are determined by manipulating the hidden layer node count in machine learning (ML). These are then contrasted with the most optimal location produced by the Cuckoo Search (CS) algorithm. This process enhances the precision and consistency of instructional evaluations while reducing the duration of training. The procedural steps for implementing the strategy above are as follows: initially, the Intelligent Instruction Effect Assessment Index is formulated by incorporating five key dimensions, namely fundamental quality, teaching demeanor, pedagogical approach, instructional proficiency, and teaching outcome. Subsequently, a professional evaluation determines each assessment index’s score and ultimate rating. Later, the evaluation indices’ scores are utilized as the input for MFEM-AI, with the resulting output being the top score. A model for evaluating the teaching effectiveness of MFEM-AI has been developed, demonstrating a high level of intelligence. Finally, the expert assessment technique is utilized to obtain the scores for every assessment index and the complete scores of the innovative teaching impact assessment. The evaluation indices have been assigned scores of 1, 0.7, 0.5, 0.3, and 0.1, which match the levels of outstanding, acceptable, medium, poor, and poor, accordingly. The structure of AI used in this research is not focused on the typical neuron structures; rather, it is used to access physical education teaching quality through fuzzy control instructions with the ECSO algorithm.
Initialization
The assessment data about the effectiveness of philosophical teaching will be read based on the index structure. Subsequently, the data is to be segregated into a set for training and a test set, followed by normalization. The parameters such as population size, maximum iterations, and learning variables are set. The range of egg velocities is defined with the initialized number of iterations.
Iterative optimization
The present study outlines the configuration of variables for the ECSO algorithm, which includes the number of structures denoted by N, the maximum amount of repetitions represented by M, and the likelihood of outside birds’ eggs being detected by the nest support, characterized by \(P_x\). Compute the values of all objective functions about the nests. The algorithm updates the velocity and position of each egg in the nest. Velocity is updated based on historical information, an egg’s ideal location, and the group’s ideal location. The fitness of each egg is calculated using a fitness function that quantifies the quality of evaluation of PE teaching.
Optimization process
The procedure involves updating the nest’s location, computing the unbiased function appreciation of the nest post-update, and subsequently comparing it with the neutral operate value before the revision. The nest with the prime objective operate value is then selected as the present place. The algorithm iteratively updates the egg positions, optimizing them to approach the best solution. A termination criterion is checked, such as reaching a maximum number of iterations or achieving a satisfactory solution.
Algorithm termination criteria
Produce a random variable, denoted as r, such that r belongs to the open interval (0,1), to conduct uniform analysis. It is assumed that r is more significant than \(P_x\). Revise the current nest position, evaluate the objective function parameters for all nests, and retain the nest position that yields the most effective accurate function value. After achieving the function criteria, the algorithm stops, and the optimal solution and its location are recorded. The stopping criteria of the proposed ECSO network stop when the following conditions are met that is a random variable \(r\) is generated in the open interval (0,1), and it is compared to a threshold \(P_x\). If \(r>P_x\), the algorithm revises the current position of the nest, and the algorithm stops after reaching the function criteria where an optimal solution with its location is recorded.
Revert to fitness evaluation
Ascertain the termination of the method. Upon satisfaction of the termination circumstance, the optimal solution from the historical data is duly documented. Alternatively, revert to Step 3. The fitness value for each egg in the population is determined with conditions and checks for teaching quality.
Evaluation result
The optimal positioning of the bird’s nest is contingent upon the optimal values of the starting input weight \(w_x\) and concealed layer bias \(b_x\) in the AI system. These optimized parameters are crucial for accurate assessments of PE quality. Similarly, the optimal values of the initial participation weight \(b_x\) and concealed layer bias \(w_x\) are utilized in the ECSO model to assess the efficacy of innovative classes.
The ML evaluation model, coupled with the ECSO, addresses parameters from PET experts by integrating fuzzy instructions. The ECSO optimizes these parameters through a comprehensive process, providing a systematic and unbiased evaluation of physical education teaching methods, ultimately leading to accurate and efficient assessment results.
The proposed model integrates natural and human language with machine learning techniques by implementing fuzzy control instructions. Additionally, the model employs an artificial intelligence-based evaluation index system to assess the effectiveness of the Pedagogical and Technological Competency Upgrading program. An analysis informed the development of the system architecture of the functional criteria. To construct a framework for evaluating teaching quality at PETCU, the proposed Multi-feature Fuzzy Evaluation Model based on Artificial Intelligence (MFEM-AI) model was utilized in conjunction with enhanced cuckoo search optimization. The multi-features of physical education are students’ motor abilities like strength, endurance, speed, flexibility, coordination, agility, and balance. The improved cuckoo search algorithm optimizes parameters in the MFEM-AI model for evaluating PET methods, enhancing precision, efficiency, and adaptability. Its advantages over traditional methods include global optimization, reduced sensitivity to initial parameters, and overall efficiency in handling the complexities of teaching method evaluation.
The importance and relationship of the obtained models MFEM-AI and ECSO algorithm, play a crucial role in enhancing the assessment of teaching quality in physical education. These models contribute to the precision and consistency of instructional evaluations, ensuring a more accurate and reliable assessment of various aspects of physical education, including student’s motor abilities and teaching effectiveness. The obtained models are interrelated through the integration of AI techniques, such as fuzzy logic and machine learning, with optimization methods like ECSO. This relationship enhances the capabilities of each model, combining the strengths of AI for intelligent decision-making and ECSO for parameter optimization. The integration of ANN with the clustering system further refines the analysis, providing a robust solution for assessing and understanding the complex dynamics of physical education instruction. The synergy between these components enhances the overall effectiveness and accuracy of the evaluation process.
link
