Practice and Effectiveness Evaluation of Personalized Training Program for Physical Education Teaching Based on Intelligent Algorithm

Physical education, as an important part of the educational field, not only contributes to students’ physical health and motor skill development, but also cultivates students’ teamwork, leadership and self-discipline [1-2]. Most of the current school sports training methods use centralized lectures, which can facilitate the management of students while saving teachers’ human resources, but this training method tends to ignore the diversity of students, resulting in a lack of targeted training and the traditional one-size-fits-all teaching method may not be applicable to every student [3-5]. Each student has a unique learning style, interest and ability level. Therefore, physical education requires more flexible and individualized teaching strategies to meet the needs and potential of different students [6-7].

A new model of physical education training has emerged in recent years, namely individualized physical education training. Physical education teachers develop personalized physical training programs for each individual or groups of students with similar physical qualities based on each student’s specific physical condition [8-10]. However, the implementation of this method does not cover a wide range, so it is important to assess the effect of physical training in order to promote the personalized physical training model, effectively improve the physical training method, and promote the improvement and development of personalized physical training, so as to improve the effect of physical training [11-13].

The current personalized sports training is a research hotspot for optimizing the teaching model, and has more worthwhile in-depth study of the exploratory, experiential, intuitive, interesting, and provides a superior teaching method for the overall development of students [14-16]. In order to study the current application effect under the personalized sports training model, it is proposed to assess the teaching effect. The traditional hierarchical assessment method, using the type of sport as the division level, assesses the teaching effect, but in the assessment process, it is found that the statistical results of the assessment of this method do not match with the students’ performance because it does not target the students’ own characteristics [17-19]. Therefore, it is necessary to explore a more reasonable and enjoyable personalized sports training assessment method, which can help improve the teaching mode of enhancing students’ interest and students’ performance, and provide a more systematic means of analysis [20-21].

The study first designs the exercise prescription generation model, which adopts a multifactor fusion approach to recommend personalized exercise programs based on the different exercise abilities, different physical conditions, and personal exercise preferences of the exercisers. Then the optimization process of the exercise parameters in the exercise prescription generation model is solved using an improved artificial raindrop algorithm expressed in discrete form. Finally, the decision-making process for prescribing exercise is proposed. In the experimental part, the validity of the method is verified and analyzed through relevant experiments, and the physical education teaching class of a school is used as an empirical research object to explore the practical application effect of the method of this paper from the changes of body morphology and physical function.

2

Personalized training program for physical education based on intelligent algorithm

2.1

Weight loss exercise prescription generation model based on artificial raindrop algorithm

2.1.1

Establishment of a standardized exercise prescription library

The structure of exercise prescription mainly includes exercise program, intensity, time, frequency, precautions, and exercise method. According to the basic parameters of exercise prescription, the exercise expectation, exercise process, effect evaluation, and risk prediction of the exerciser need to be considered when formulating exercise prescription, and some of the exercise prescriptions are shown in Table 1.

Table 1.

The display table of partial exercise prescription

The Core Parameters Of The Diet				Non-Core Parameters Of Weight Loss Exercise
Sports Project	Exercise Intensity (Times/Points)	Sports Time (Minute)	Motion Frequency (Secondary/Week)	Sports Method	Considerations	Taboo Disease
Walk Away	117~164	30~60	5~7	The Upper Body Is Upright In The Top 35 Degrees, And Head Up And Chest. The Two Arms Naturally Oscillate.	Choose The Appropriate Sports Environment And Relax.	Nothing
Stair	110~170	20~40	3~5	The Body Must Be Slightly Slightly Smaller Than The Stairs, And Keep The Flexibility Of The Lower Joints	Don’t Exercise This Exercise With The Ankle Joint.	Knee Joint, Ankle Disease, Asthma.
Cross Race	120~175	30~60	3~5	The Speed Of The Running Is Appropriate, And The Two Methods Are Practiced By Timing.	Choose The Appropriate Equipment And Relax After Exercise.	Heart Disease, Asthma, And Cardiopulmonary Dysfunction.
…	…	…	…	…	…	…
Aerobics	105~168	20~50	5~7	Choose The Appropriate Aerobic Rhythm Exercise.	Warm Up And Choose The Appropriate Environment.	Heart Disease, Asthma, And Cardiopulmonary Dysfunction.

2.1.2

Personalised exercise programme recommendations for weight loss exercise prescription

In this paper, the selection of exercise programs is based on three main considerations: the effect of exercise on similar users, the user’s preferences and diseases, and the comprehensive exercise ability of the exerciser.

1)

Exercise Effectiveness to Determine the Set of Programmes

The basic data information of eight dimensions of users (cardiac functional capacity, body size, fat percentage, body mass index BMI, lung capacity, step test, seated forward bending, and age) is used to normalise the data, and the processing is done in a min-max way, and the normalised data are used as input vectors for clustering, and the K-means algorithm is used for clustering the user data.

2)

Exercise preference to determine the collection of items

According to the personal information of sportsmen such as fitness venues, sports time, sports hobbies, sports habits and consumption level, sports preference rule processing is carried out to generate sports preference rule files and establish the rule base of corresponding sports items.

3)

Comprehensive sports ability to determine the collection of items

Calculate the comprehensive sports ability of sportsmen according to the results of physical fitness assessment, and the comprehensive sports ability is assessed in the way of 7A using the weighted summation, mainly through the indicators of cardiac functional capacity, BMI (body mass index), back strength, grip strength, and standing heel lift, etc. [22]. A fusion approach was applied to the three sport item sets to determine the final sport items. The specific process of exercise programme determination is shown in Figure 1. After the exercise program was determined, non-core parameters such as precautions and exercise methods were determined against a library of standard weight-loss exercise prescriptions.

2.1.3

Construction of a model for weight loss exercise prescription generation

The core parameters of weight loss exercise prescription exercise intensity, time, and frequency were determined by constructing a multi-objective optimization model and solving it using a multi-objective optimization algorithm. Based on the basic personal physique of the exerciser (e.g., height, weight, BMI, fat percentage, etc.), the negative balance of calories that the exerciser needs to achieve to reach a normal or ideal body shape is determined, and the multi-objective optimisation model is established under a variety of constraints to satisfy the exerciser’s physique, exercise goals, and other constraints.

1)

Determination of Expected Negative Balance for Exercisers

Negative balance to be achieved by the exerciser for weight loss (W_f) = Exercise consumption(W_c) + Daily consumption(W_B)–Intake(W_l) . When the negative balance of calories corresponds to the weekly weight to be lost, the desired effect of weight loss is achieved. Based on the BMI (body mass index), age and other information, the user’s weight to be lost and the weekly weight to be lost can be calculated [23].

The specific method of determining the ideal weight of an exerciser is shown in Equation (1), where H is the height of the exerciser, W_after is the ideal weight, and W_before is the current weight.

(1)

W_{a f t e r} = B M I \cdot H^{2}

Calculate the exerciser’s exercise goal, i.e., the ideal negative equilibrium calories W_f per week, as shown in equation (2).

(2)

W_{f} = \frac{W_{a f t e r} - W_{b e f o r e}}{12} \times 1840 k a l

In order to more accurately calculate the negative balance calories of the exerciser, the daily consumption of the exerciser is taken into account, and the daily activity coefficient of the exerciser, as well as the daily consumption of the exerciser, is calculated based on factors such as the intensity of the exerciser’s hours of work and the state of daily activity. Daily consumption W_B = Basic consumption BMR × Activity factor μ. where μ is determined as shown in equation (3). x_j is the intensity level of the influence factor, x is the highest level of the influence factor, and the weight of the influence factor is θ.

(3)

μ = 1 + \sum_{j = 1}^{n} θ_{j} \frac{x_{j}}{x}

2)

Determination of exercise optimisation parameters and their constraints

According to the core elements of weight loss exercise prescription can be determined that the parameters that need to be optimised and adjusted are intensity, time and frequency respectively. According to the process to determine the exercise items, according to the exercise items corresponding to the corresponding standard exercise prescription, the range of exercise parameters can be obtained. In this paper, taking into account the actual exercise process, the exercise state is divided into three parts, which are pre-exercise, mid-term and post-exercise, and the exercise process allocation scheme is shown in Table 2.

Table 2.

The distribution scheme of the exercise process

Motor Process	Motor Intensity	Motion Time
Presports	Lesser	Shorter (About 20%)
Medium Motion	Larger	Longer (About 60%)
Postmotion	Lesser	Shorter (About 20%)

(1) The range of exercise parameters can be obtained according to the sports corresponding to the standard exercise prescription, as in Equation (4), ν represents the frequency of exercise, s_i and t_i are the intensity and duration of exercise in the pre-, mid- and post-exercise periods, and ν is the frequency of exercise. (4) ${\begin{array}{l} s_{i \min} \leq s_{i} \leq s_{i \max} \\ t_{i \min} \leq t_{i} \leq t_{i \max} \\ 0 \leq v \leq 7 \end{array}$

(2) Due to the differences in the basic physical condition, overall exercise capacity and exercise expectations of different athletes, it is necessary to develop a personalised exercise intensity, duration and frequency for athletes, and the three exercise parameters affect and constrain each other. Exercise parameters determine the amount of exercise. The total exercise volume needs to be less than the maximum exercise volume N that the exerciser can tolerate, as shown in Equation (5). (5) $0 < \sum_{0}^{v} \sum_{i = 1}^{3} (ω_{i} \int_{0}^{t} s_{i} d t) < N$

(3) Introduce the user motion expectation into the model, and explore the optimal solution within the range of parameters acceptable to the user, and set the constraints that the distance between the motion parameters and the user motion expectation needs to be in a fixed interval, as shown in Equation (6). Calculate the gap between the motion parameters and the motion expectation in the middle of the motion process θ, where s_e, t_e are the user’s expected motion intensity and the user’s expected motion time, respectively. (6) $\frac{s_{2} s_{e} + t_{2} t_{e}}{\sqrt{s_{2}^{2} + t_{2}^{2}} \times \sqrt{s_{e}^{2} + t_{e}^{2}}} < θ$

3)

Determination of the motion objective function

According to the formula for calculating negative equilibrium calories in an exercise cycle, combined with the user’s exercise expectation, the exercise expectation satisfaction function was used to determine the objective function, and the expected achievement function of the negative equilibrium heat obtained by the model optimization parameters was f_w(x), the exercise intensity expectation achievement function was f_s(x), and the exercise time expectation achievement function was f_t(x), and the closer the objective function was to 0, the higher the desired achievement degree was achieved. The specific objective function construction is shown in equation (7): (7) $\begin{array}{l} f_{w} (x) = 1 - \frac{1}{1 + β {((v \sum_{i = 1}^{n} \int_{0}^{t_{i}} G (s_{i}) d t_{i} + (B M R \times (1 + \sum_{j = 1}^{n} θ_{j} \frac{x_{j}}{x}) - W_{l}) \times 7) - W_{f})}^{2}} \\ f_{s} (x) = 1 - \frac{1}{1 + β {(\frac{\sum_{i = 1}^{n} s_{i}}{n} - s_{e})}^{2}} \\ f_{t} (x) = 1 - \frac{1}{1 + β {(t - t_{e})}^{2}} \end{array}$

Where G(s_i) represents the exercise consumption per unit of time at an exercise intensity of s_i, W_t is the calorie intake of the exerciser, W_f is the calculated optimal negative balance calories of the exerciser. G(s) The calculation formula is shown in (8), where gender is the gender, gender = 1 when the exerciser is male, gender = 0 when the exerciser is female, weight is the weight of the exerciser, and age is the age.

(8)

\begin{matrix} G (s) = g e n d e r \\ \times (- 55.0969 + 0.6309 \times s + 0.1988 \times w e i g h t + 0.2017 \times a g e) \\ + (1 - g e n d e r) \\ \times (- 20.4022 + 0.4472 \times s - 0.1263 \times w e i g h t + 0.074 \times a g e) \end{matrix}

Based on the above conditions, it can be seen that there is a constraint relationship between the desired degree of achievement of exercise intensity, exercise time and the desired degree of achievement of negative equilibrium heat. Based on the determined constraints and the three desired degree of achievement objective functions, a multi-objective optimisation model can be established as shown in Equation (9): (9) $\begin{matrix} \min F (x) = (f_{w} (x), f_{s} (x), f_{t} (x)) \\ s . t . s_{i \min} \leq s_{i} \leq s_{i \max} \\ t_{i \min} \leq t_{i} \leq t_{i \max} \\ 0 < v \leq 7 \\ \sum_{0}^{v} \sum_{i = 1}^{3} ω_{i} \int_{0}^{t_{i}} s_{i} d t_{i} < N \\ \frac{s_{2} s_{e} + t_{2} t_{e}}{\sqrt{s_{2}^{2} + t_{2}^{2}} \times \sqrt{s_{e}^{2} + t_{e}^{2}}} < θ \end{matrix}$

4)

Constraint processing and objective function conversion

In order to more conveniently use the optimisation algorithm to solve the weight loss exercise prescription parameters, this paper adopts the weighting method to transform the weight loss exercise prescription parameter solving problem into a single-objective optimisation problem. The importance of the objective is considered according to the actual situation, and the negative balance of calories satisfaction needs to be higher in order to achieve the corresponding exercise effect.

2.2

Artificial Raindrop Algorithm

2.2.1

Basic Flow of Artificial Raindrop Algorithm

The artificial raindrop algorithm simulates the natural phenomenon by dividing the optimisation process of the algorithm into six stages, each of which corresponds to a different operator, namely, raindrop formation operator, raindrop descent operator, raindrop collision operator, raindrop flow operator, raindrop pool updating operator, and water vapour updating operator. The above six operators are described in detail below.

1)

Raindrop formation operator

In nature raindrops are formed by the continuous condensation of rising water vapour, the raindrop formation operator simulates the formation process of raindrops by defining the water vapour that forms raindrops as water vapour vectors, and N water vapour vector makes up the initial population size [24]. In this case, the i nd water vapour vector in the population is defined as shown in equation (10) on the following page.

(10)

V a p_{i} = [v a p_{1, i}, v a p_{2, i}, \dots, v a p_{D_{j}}], i = 1, 2, \dots, N

(11)

R a i n = (\frac{1}{N} \sum_{i = 1}^{N} V a p_{1, i}, \frac{1}{N} \sum_{i = 1}^{N} V a p_{2, i}, \dots, \frac{1}{N} \sum_{i = 1}^{N} V a p_{D, i})

2)

Raindrop falling operator

After a raindrop is formed and descends downwards due to gravitational factors, defining N _ Rain as a falling raindrop, the coordinates Rain_d of a dimension may change during the raindrop’s descent. The raindrop descent operator is defined as shown in equation (12).

(12)

N_R a i n_{i} = {\begin{array}{l} r a n d o m (F_{d}, C_{d}) & i = d \\ R a i n_{i} & O t h e r \end{array}

where random(Fd,Cd) is a random number between F_d and C_d.

3)

Raindrop collision operator

When a raindrop falls to the ground, it collides with the ground and is impacted into a number of small raindrops. In order to maximise the simulation of the raindrop collision scene, it is assumed that the raindrop collision produces small raindrops in the form of an approximate normal distribution falling around. The raindrop collision operator is defined as shown in Equation (13).

(13)

S_R a i n_{i} = N_R a i n + (N_R a i n - V a p_{k}) \cdot s i g n (a - 0.5) \cdot \log b

where a and b are denoted as any random numbers between [0,1] , sign(·) is a sign function, and Vap_k is any water vapour vector in the water vapour population.

4)

Raindrop flow operator

Small raindrops produced by the collision fall randomly at different locations and have a tendency to flow towards lower altitudes due to geographic location factors. The raindrop flow direction is determined by a linear combination of two sub-directions o1 and o2, as shown in Eq. (14).

(14)

\begin{array}{l} o 1_{i} = (L O C_{k_{i}} - S_R a i n_{i}) \cdot s i g n (f (L O C_{k_{1}}) - f (S_R a i n_{i})) \\ o 2_{i} = (L O C_{k_{2}} - S_R a i n_{i}) \cdot s i g n (f (L O C_{k_{2}}) - f (S_R a i n_{i})) \\ o_{i} = ζ \cdot r a n d o m 1_{i} \cdot o l_{i} + ζ \cdot r a n d o m 2_{i} \cdot o 2_{i} \end{array}

where ζ is the raindrop flow factor and LOC_k1 and LOC_k2 denote any two positions in the raindrop pool, respectively. The definition of raindrops after flow is shown in Eq. (15).

(15)

N_S_R a i n_{i} = S_R a i n_{i} + o_{i}, i = 1, 2, \dots, N

5)

Raindrop pool update operator

The artificial raindrop algorithm introduces the concept of raindrop pool to save the small raindrop Best^G with the lowest position information in each iteration, and sets the upper limit of the size of the raindrop pool to save raindrops to N. The raindrop pool updating operator is defined as shown in Eq. (16).

(16)

P_R a i n^{G + 1} = P_R a i n^{G} \cup B e s t^{G}

6)

Water vapour update operator

The artificial raindrop algorithm uses the water vapour update operator for the retention of good individuals. Taking the concatenation set of Vap^G and New^G, a sorting algorithm is applied to select the first N individuals with low elevation as the initial population for the next iteration, as shown in Eq. (17).

(17)

V a p^{G + 1} = V a p^{G} \cup N e w^{G}

2.2.2

Training programme for applying the artificial raindrop algorithm

In this paper, an artificial raindrop algorithm with improved discrete form expression is explored, and the original algorithm flow is modified for some operators, and the current algorithm mainly consists of raindrop descent operator, raindrop collision operator, raindrop flow operator, raindrop evaporation operator and raindrop migration operator [25]. The design of each operator is described in detail below. The general artificial raindrop algorithm simulation scenario diagram is shown in Fig. 2.

1)

Raindrop descent operator

Individual raindrops in the atmospheric raindrop swarm are individually descended to the ground by applying the raindrop descent operator, which completes the “perception” of the potential energy information at the current spatial location of the problem. The initialised raindrop population consists of a number of raindrops randomly generated according to specific rules, while the raindrop population in the loop iteration consists of raindrops evaporated from the ground to the atmosphere.

2)

Raindrop collision operator

The raindrop collision operator simulates the raindrop collision phenomenon in nature, the raindrops in the atmosphere fall to the ground after the collision phenomenon occurs, there is a certain chance of generating new raindrops through the collision of individuals, these new raindrops are scattered around the current raindrops individuals.

3)

Raindrop flow operator

The raindrop flow operator simulates the raindrop flow phenomenon in nature. When raindrops are in rugged terrain and not subject to external forces, raindrops always flow towards places with lower potential energy. The raindrop individual evaluation value is calculated as shown in Equation (18).

(18)

V = \frac{E_{\min}}{E_{c u r}}

Where V is the evaluation value of the current raindrop individual, E_cur is the potential energy value of the current raindrop individual, and E_min is the potential energy value of the historical optimal raindrop individual. The value of water molecules on the water molecule structure at position k after the flow of raindrops is shown in Equation (19) on the following page.

(19)

{\begin{array}{l} V_{k} = V_{k} + [Δ G] & (V > 0.5) \\ V_{k} = V_{k} + [Δ C] & (V \leq 0.5) \end{array}

4)

Raindrop Evaporation Operator

The raindrop evaporation operator will flow after the potential energy value of raindrops in the raindrop group in accordance with the sorting algorithm for ascending order processing, will be ranked in the N after a number of raindrops to apply the raindrop evaporation operator. The raindrops do not participate in the next iteration of the cycle after applying the raindrop evaporation operator. That is, some of the better raindrops in the current raindrop group will be selected to participate in the next iteration of the cycle. The raindrop individual evaporation simulation scenario diagram is shown in Figure 3.

5)

Raindrop migration operator

The raindrop migration operator simulates that during the evaporation of raindrops, some raindrops are deflected under the action of external forces, which makes the raindrops with higher potential energy values more likely to be deflected to the location where the raindrops with lower potential energy are located during the evaporation process.

In summary, the flow of the improved artificial raindrop algorithm is shown in Figure 4.

2.3

Decision-making process for generating exercise prescriptions

Exercise prescription in this study refers to an exercise programme based on students’ age, gender and diagnosis in the absence of other underlying diseases or movement disorders, and its structure includes seven parts: exercise goal, exercise content, exercise time, exercise intensity, organisational methods, exercise frequency and precautions:

1) Exercise goal refers to the exercise effect achieved by purposeful exercise based on the diagnostic results.

2) Exercise content refers to the means, methods and types of exercise used during exercise.

3) Exercise time is the sum of the time students last for a physical exercise session according to the exercise prescription, which was set at 50-60 minutes in this study.

4) Exercise intensity refers to the degree of physiological stimulation of the exercise to the students, using the average heart rate to measure the size of the exercise intensity, heart rate less than 120 beats/minute is low intensity, 120-150 beats/minute is medium intensity, 150-180 beats/minute or more than 180 beats/minute is high intensity.

5) Organisation refers to the number of repetitions, number of groups completed, and intervals for each action or exercise program.

6) The frequency of exercise refers to the number of times per week, generally 3-4 times per week, that is, every other day exercise, so that the body can get a “super recovery”, to get a better exercise effect.

7) Precautions refers to the safety precautions required to ensure exercise safety and avoid injury.

The decision-making process of generating exercise prescription on the platform is shown in Figure 5.

The platform stores the specific content of the exercise prescription in the database according to the correspondence of the diagnostic indicators. Based on the indicators of strengths and weaknesses and the order of development determined by the gap diagnosis, overall posture diagnosis and individual posture diagnosis of the students, the prescription data corresponding to the diagnostic indicators in the database are automatically selected in accordance with the structure of the exercise prescription, and a complete exercise prescription is generated.

In addition to using diagnostic results as the decision-making basis for exercise prescription, the platform also sets four decision-making factors: obesity, age, puberty, and gender. Among them, the obesity factor has the highest priority, and an obesity-specific aerobic exercise prescription is enabled when the student is judged to be obese. In the age factor, primary school is the low age group, middle school, high school and university is the high age group, taking into account that students’ bodies are not yet fully developed in the primary school stage, and setting up exercise prescriptions applicable to students in the low age group. In the puberty factor, taking into account the different growth and development characteristics of boys and girls, additional exercise prescriptions for puberty are set up separately according to the gender factor. The platform defaults to junior high school students as being in the puberty stage, but it can also be set by students or teachers to determine whether they are currently in the puberty stage or not.

3

Results and discussion

3.1

Experiments and analyses

In order to verify the reasonableness of the method and the effectiveness of the related improved algorithm, this paper carries out a large number of experiments and analyses. Using the existing data and simulation data under various motion conditions, the corresponding results are obtained through comparison experiments and verification experiments, and the results are analysed to draw the corresponding conclusions.

3.1.1

Data acquisition and correlation analysis

As the current experimental data related to sports courses cannot be obtained directly from other platform websites in a complete way, this paper uses the official sports-related competition data provided by Kaggle as the basic data for the experiments in this paper, and for this dataset, this paper also does the relevant processing, including screening and analysing, for the specific running data, in order to show the distribution of the data of the sports cases, the correlation matrix analysis of running sports data is shown in Fig. 6. 2000 running sports data were selected from the set to do the correlation analysis, and the correlation matrix analysis of running sports data is shown in Figure 6. From the figure, it can be seen that there is a strong correlation between the running sports data starting from 14 sports.

3.1.2

Parameter optimisation experiments for the combination of sports courses

In the simulation experiments comparing the artificial raindrop algorithm proposed in this paper with the ARA algorithm, the population size of both algorithms is set to 30, the size of the raindrop pool is set to 50, and similarly, the maximum number of iterations is set to 1500. For some parameters in ARA, the flow factor is set to 2, and the maximum number of flows is set to 3. For the parameters of this paper’s algorithm, the collision factor is α_e = 0.5 and α_s =-0.5, the flow factor is τ_g = 3 and τ_p = 1, and the maximum number of flows is 3. Specific cases are experimented using data from the case database, and data from the case database are used according to the following criteria: fat loss and slimming, improving aerobic endurance, improving anaerobic endurance, Muscle building and shaping, improve flexibility five exercise goals were randomly selected two, and in accordance with the male to female ratio of 1:1 for selection, a total of ten case data were selected. For clarity, this section chooses to provide an example of one of the case data in the form of a text description. The main attribute information of one of the cases is described as follows: Case 1 is a 25-year-old male, 173 cm in height, 132 kg in weight, no chronic diseases, with basic sports equipment and venues, the effect of the exercise prescription is to improve aerobic endurance, the intensity of the prescribed exercise is to keep the heart rate at 152 for the best results, the prescribed exercise time is 40 minutes for the best, and the frequency of the prescribed exercise is 4-5 times a week. The prescribed exercise frequency is 4-5 times a week. This section combined with ten specific cases were done single average optimisation process experiments, ten times the average optimisation process experiments, ten times the optimal average optimisation process experiments. The statistics of the calculation results of different values before and after the algorithm improvement are shown in Table 3. Shown in the table are the statistics of the calculation results of the different values for three experiments conducted by the improved artificial raindrop algorithm and the original artificial raindrop algorithm for ten cases respectively. From the table, it can be concluded that the improved artificial raindrop algorithm in this paper is significantly better than the original artificial raindrop algorithm in terms of optimisation results, both in terms of the optimal value and the average value.

Table 3.

The algorithm improves the calculation results of different values

Case	Different value	ARA	Ours
Case 1	Single average	1.49E+00	9.87E-01
	Ten mean	1.78E+00	1.07E+00
	The 10 optimal average	1.41E+00	6.83E-01
Case 2	Single average	2.27E+00	1.16E+00
	Ten mean	1.35E+00	0.74E+00
	The 10 optimal average	1.76E+00	6.67EE-01
Case 3	Single average	1.81E+00	1.5E+00
	Ten mean	2.15E+00	1.19E+00
	The 10 optimal average	1.23E+00	8.85E-01
Case 4	Single average	1.98E+00	10.23E+00
	Ten mean	2.05E+00	1.32E+00
	The 10 optimal average	1.52E+00	8.35E-01
Case 5	Single average	1.14E+00	8.67E-01
	Ten mean	0.88E+00	9.81E-01
	The 10 optimal average	7.74E-01	4.25E-01
Case 6	Single average	1.41E+00	8.93E-01
	Ten mean	1.18E+00	1.56E+00
	The 10 optimal average	7.67E-01	5.52E-01
Case 7	Single average	1.8E+00	1.89E+00
	Ten mean	1.92E+00	1.38E+00
	The 10 optimal average	1.26E+00	0.91E+00
Case 8	Single average	1.98E+00	1.86E+00
	Ten mean	1.65E+00	1.44E+00
	The 10 optimal average	1.25E+00	1.36E+00
Case 9	Single average	1.38E+00	0.83E+00
	Ten mean	1.84E+00	1.71E+00
	The 10 optimal average	1.23E+00	1.4E+00
Case 10	Single average	1.91E+00	2.08E+00
	Ten mean	2.14E+00	1.55E+00
	The 10 optimal average	1.25E+00	1.22E+00

In order to further confirm the above conclusions, this paper also conducts a comparison experiment on the specific optimised parameters. This experiment was conducted by running the data from Case 1 on ARA and this paper’s algorithm independently for 150 times, and storing the optimised parameter values for each time to be used in the calculation of the comparison experiment. Firstly, for the difference of exercise volume, this section calculates the cycle exercise volume of the exercise prescription in Case 1 and the cycle exercise volume optimised and calculated by the two algorithms according to the formulae, and then the cycle exercise volume optimised and calculated by different algorithms is used to do the difference calculation with the cycle exercise volume of the exercise prescription, and then the comparison experiment is carried out. It can be concluded from the line graph of Fig. 7(a) that the difference of cycle exercise volume is limited within 150 and the difference statistics of cycle exercise volume in Fig. 7(b) that the difference of 150 cycle exercise volumes optimally computed by the algorithm of this paper is distributed within 100 in the vast majority of the cycle exercise volumes, while the difference of 150 cycle exercise volumes optimally computed by the ARA is distributed above 100 in the vast majority of the cycle exercise volumes. The comparison and statistics of the difference in motion are shown in Fig. 7.

In addition, in this paper, the values of the four quantities of cycle motion, motion intensity, motion time and motion frequency optimised by ARA and the algorithm in this paper are normalised respectively, and then the average value of each quantity after normalisation is calculated for comparison. In addition, the values of the above four quantities optimally calculated by the algorithms of the ARA and this paper and the values of the specific case 1 are respectively differentiated and normalised, and then the average value of the difference of each quantity after normalisation is calculated and compared and plotted as a bar chart. A comparison of the mean value and the mean value of the difference is shown in Fig. 8 (Fig. a is a comparison of the mean value and Fig. b is a comparison of the mean value of the difference). According to the experimental results, it can be concluded that: this paper’s algorithm and ARA in the amount of exercise and exercise frequency of the two quantities of the comparison, whether it is the average value or the difference this paper’s algorithm is significantly better than ARA, this paper’s algorithm in terms of the intensity of the exercise, the average value and the difference is also slightly better than ARA, the difference of the performance of the time of the exercise is slightly lower than the RAR, and the average value is slightly better than the ARA.Comprehensive the above experiments found that. The algorithm in this paper can effectively optimise a combination of exercise course parameters that is closer to the effect of exercise prescription, and at the same time is more in line with the personalised needs of exercisers, which ensures the effectiveness of the application of exercise prescription and improves the effectiveness of the exercise combination programme. Additionally, the optimized set of parameters can match a variety of sports programs, which can ensure the effectiveness of sports while enhancing the diversity of sports.

3.2

Effectiveness of individualised programme practice in physical education

In this section, a 10-week experiment was conducted with students from two physical education classes in a school as the empirical research subjects.The experiment was set up with an experimental group (60 students) and a control group (60 students), with a male to female ratio of 1:1.The experimental group used the method proposed in this paper to assist in physical education teaching, while the control group continued to follow the traditional methods of physical education teaching. The experimental group used the method proposed in this paper to support physical education, while the control group continued to use the traditional physical education teaching method.

3.2.1

Experimental results and analyses in terms of body morphology

1)

Within-group comparison of body mass index of boys in the experimental and control groups before and after the experiment

The intra-group comparison of body mass index (in Kg/m²) between the boys in the experimental group and the control group is shown in Table 4. According to the table, after 10 weeks of aerobic endurance exercise prescription workout, the mean of the body mass index of the boys in the experimental group decreased by 1.08, which is a highly significant difference with P<0.01 using paired samples t-test. After 10 weeks of independent exercise, the mean body mass index of boys in the control group decreased by 0.09 using paired samples t-test, P>0.05, with no significant difference. This shows that aerobic endurance exercise prescribed exercise in the after physical education service had a very significant improvement in the body mass index of secondary school boys and autonomous exercise had no significant effect on the body mass index of boys.

Table 4.

The experimental group was compared with the control group

	Preexperiment	After the experiment	T	P
Experimental group n=30	21.29±4.04	20.21±3.24	5.075	<0.001
Control group n=30	20.27±4.09	20.18±3.59	2.063	0.058

2)

Intergroup comparison of body mass index between the experimental group and the control group of boys before and after the experiment

The intergroup comparison of body mass index (BMI) of boys in the experimental group and control group (unit: Kg/m²) is shown in Table 5. According to the table, the mean of body mass index of the boys in the experimental group before the experiment was 0.21 lower than that of the boys in the control group, and the mean of body mass index of the boys in the experimental group after the experiment was 1.06 lower than that of the boys in the control group.The body mass index of the boys in the experimental group after the experiment was subjected to an independent samples t-test, which showed that there was no significant difference at P>0.05. This shows that before and after the experiment, the body mass index of the boys in the experimental group decreased significantly compared to the boys in the control group, but there was no statistical difference between the two groups. This shows that the aerobic endurance exercise prescription workout in the after physical education service had a very significant improvement in the boys’ body mass index compared to the independent workout, but not enough to make a significant difference between the groups.

Table 5.

The experimental group was compared with the control group of the control group

	Experimental group n=30	Control group n=30	T	P
Preexperiment	20.34±4.17	20.55±3.89	-0.136	0.905
After the experiment	19.88±3.63	20.94±3.67	-0.533	0.592

3)

Intra-group comparison of body mass index of girls in the experimental group and the control group before and after the experiment

The intra-group comparison of body mass index (in Kg/m²) between the girls in the experimental group and the control group is shown in Table 6. According to the table, after 10 weeks of aerobic endurance exercise prescription exercise, the mean of the body mass index of the girls in the experimental group decreased by 0.26, which is a significant difference using paired samples t-test, P<0.05. After 10 weeks of independent exercise, the mean of the body mass index of the girls in the control group decreased by 0.18, using paired samples t-test, P>0.05, with no significant difference. This shows that aerobic endurance exercise prescribed exercise in the after-school physical education service had a significant improvement in the body mass index of secondary school girls and autonomous exercise had no significant effect on the body mass index of secondary school girls.

Table 6.

The experimental group was compared with the group of the control group

	Preexperiment	After the experiment	T	P
Experimental group n=30	18.47±2.49	18.21±2.34	2.721	0.017
Control group n=30	18.44±2.33	18.09±2.45	1.346	0.145

4)

Intergroup comparison of body mass index of girls in the experimental group and the control group before and after the experiment

The inter-group comparison of BMI (in Kg/m²) of girls in experimental group and control group is shown in Table 7. According to the table, the mean of body mass index of the girls in the experimental group before the experiment was 0.15 lower than that of the girls in the control group, and the mean of body mass index of the girls in the experimental group after the experiment was 0.23 lower than that of the girls in the control group.The body mass index of the girls in the experimental group after the experiment was subjected to an independent samples t-test, and the difference was not significant at P>0.05. This shows that before and after the experiment, the body mass index of the girls in the experimental group decreased significantly compared to the girls in the control group, but there was no statistical difference between the two groups. This shows that aerobic endurance exercise prescription exercise compared to independent exercise in the after physical education service has a significant improvement in the body mass index of secondary school girls, but not enough to make a significant difference between the groups.

Table 7.

The experimental group was compared with the control group of the control group

	Experimental group n=30	Control group n=30	T	P
Preexperiment	18.31±2.64	18.46±2.81	-0.166	0.492
After the experiment	18.85±1.86	19.08±1.96	-0.308	0.473

3.2.2

Experimental results and analyses in terms of physical functioning

1)

Intra-group comparison of spirometry between the experimental and control groups of boys before and after the experiment

The intra-group comparison of lung capacity (unit: ml) between the boys in the experimental group and the control group is shown in Table 8. According to the table, after 10 weeks of aerobic endurance exercise prescription workout, the mean lung capacity of the boys in the experimental group was improved by 128.05 ml, which is a highly significant difference using paired samples t-test, P<0.01. After 10 weeks of independent exercise, the mean lung capacity of boys in the control group increased by 69.73 ml, using paired samples t-test, P<0.01, with a highly significant difference. This shows that both aerobic endurance exercise prescription exercise and autonomous exercise in the after-school physical education service had a very significant effect on the lung capacity of the boys.

Table 8.

The experimental group was compared with the control group

	Preexperiment	After the experiment	T	P
Experimental group n=30	3225.07±793.09	3353.12±806.04	-5.136	<0.001
Control group n=30	3236.15±828.35	3305.88±811.88	-4.648	<0.001

2)

Intergroup comparison of spirometry between the experimental group and the control group of boys before and after the experiment

The intergroup comparison of spirometry (unit: ml) between the boys in the experimental group and the control group is shown in Table 9. According to the table, the mean of lung capacity of boys in the experimental group before the experiment was 12.36 ml lower than that of boys in the control group, and the mean of lung capacity of boys in the experimental group after the experiment was 45.68 ml higher than that of boys in the control group.An independent samples t-test was performed to compare the lung capacity of boys in the experimental group after the experiment with that of the boys in the control group and there was no significant difference at P>0.05. This indicates that the boys in the experimental group showed a greater improvement in lung capacity than the boys in the control group, but there was no statistical difference between the two groups. This shows that aerobic endurance exercise prescription exercise has a better improvement in lung capacity of boys than voluntary exercise, but there is no significant difference between the two.

Table 9.

The comparison between the experimental group and the control group was compared

	Experimental group n=30	Control group n=30	T	P
Preexperiment	3222.2±789.62	3234.56±829.88	-0.061	0.954
After the experiment	3353.75±802.6	3308.07±810.97	0.213	0.838

3)

Within-group comparisons of spirometry between girls in the experimental group and the control group before and after the experiment

The intra-group comparison of lung capacity (unit: ml) between the girls in the experimental group and the control group is shown in Table 10. According to the table, after 10 weeks of aerobic endurance exercise prescription workout, the mean of lung capacity of girls in the experimental group improved by 73.62 ml, which is a highly significant difference using paired samples t-test, P<0.01. After 10 weeks of independent exercise, the mean lung capacity of girls in the control group increased by 43.31 ml, with a highly significant difference of P<0.01 using paired samples t-test. This shows that both prescription and autonomous aerobic endurance exercise in the after-school physical education service have a very significant effect on the lung capacity of secondary school girls.

Table 10.

The experimental group was compared with the control group of the control group

	Preexperiment	After the experiment	T	P
Experimental group n=30	2433.49±666.02	2507.11±635.38	-6.747	<0.001
Control group n=30	2419.51±601.86	2462.82±603.48	-2.877	0.017

4)

Intergroup comparison of spirometry between girls in the experimental group and the control group before and after the experiment

The intergroup comparison of spirometry (unit: ml) between the girls in the experimental group and the control group is shown in Table 11. According to the table, the mean of lung capacity of girls in the experimental group was 12.63 ml higher than that of girls in the control group before the experiment, and the mean of lung capacity of girls in the experimental group was 44.7 ml higher than that of girls in the control group after the experiment.An independent samples t-test was performed on the lung capacity of girls in the experimental group and girls in the control group after the experiment and there was no significant difference at P>0.05. This indicates that the girls in the experimental group showed a greater improvement in lung capacity than the girls in the control group, but there was no statistical difference between the two groups. This shows that aerobic endurance exercise prescription exercise has a better improvement in lung capacity of girls than voluntary exercise, but there is no significant difference between the two.

Table 11.

The comparison was compared with the control group

	Experimental group n=30	Control group n=30	T	P
Preexperiment	2431.13±665.33	2418.5±603.28	0.102	0.91
After the experiment	2509.88±634.7	2465.18±603.48	0.272	0.793

4

Conclusion

The article takes the intelligent algorithm as the core, and constructs a personalized sports prescription generation model based on the artificial raindrop algorithm, which provides students with applicable sports training programs, and improves their enthusiasm for sports as well as their sports effects.

In order to verify the usability, reasonableness and relevance of this paper’s method, the article compares this paper’s algorithm with the ARA algorithm, and the experimental results show that this paper’s method significantly outperforms the ARA algorithm in both the optimal value and the average value of the optimization results.

After a 10-week experiment on students in a school, it is found that the exercise prescription designed in this paper can effectively improve the students’ body shape, and the improvement effect on boys’ body shape is more obvious, while the traditional method of physical education has no obvious effect on students’ body shape.

In summary, the exercise prescription generation model based on the artificial raindrop algorithm proposed in this paper provides a new idea for personalized training programs for physical education.

Funding:

In 2023, the Ministry of Education’s Humanities and Social Sciences Research Fund Planning Project “Research on the Mechanism and Path of Deep Integration of Digital Economy and China Sporting Goods Manufacturing Industry”.

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

Practice and Effectiveness Evaluation of Personalized Training Program for Physical Education Teaching Based on Intelligent Algorithm

Zhijin Tong

Xiao Zhao

Pubblicato online: 19 mar 2025

Ricevuto: 29 ott 2024

Accettato: 19 feb 2025

DOI: https://doi.org/10.2478/amns-2025-0542

Parole chiaveArtificial raindrop algorithm, Exercise prescription, Physical education teaching, Personalized training program

© 2025 Zhijin Tong et al., published by Sciendo

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

Parole chiave
Artificial raindrop algorithm, Exercise prescription, Physical education teaching, Personalized training program