Access to this full-text is provided by Wiley.
Content available from Journal of Environmental and Public Health
This content is subject to copyright. Terms and conditions apply.
Research Article
Research on Injury Causes and Prevention Effect of College
Rowing Athletes Based on Multiple Regression and
Residual Algorithm
Nan Mu
Tianjin University of Science and Technology, Tianjin 300457, China
Correspondence should be addressed to Nan Mu; mn@tust.edu.cn
Received 20 June 2022; Revised 21 July 2022; Accepted 21 September 2022; Published October 2022
Academic Editor: Hye-jin Kim
Copyright © 2022 Nan Mu. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Rowing competition in colleges and universities is an international competition, and it is also a favorite competition for college
students. However, in the course of rowing competition, the stability of athletes’injuries often occurs, which is difficult to solve
effectively. Aiming at the problem that the loss of athletes in rowing competition in colleges and universities cannot be
accurately prevented, this paper puts forward a multiple regression prevention effect model and makes a comprehensive
analysis combined with complex reasons. Through the integration of multiple regression and residual analysis, we can better
find out the influencing factors, aiming at finding out the causes of athletes’injuries and putting forward corresponding
countermeasures. First of all, analyze the causes of loss, establish a framework of injury prevention for college rowers, and the
overall diagnosis framework is reasonable. Then, according to the “University Rowing Prevention and Control Standards”
divided into various prevention measures, through the comprehensive prevention and control measure mechanism to get the
cause of injury, finally, the optimal combination of various control measures forms a control system. The results of MATLAB
show that the combination of multiple regression and residual analysis can improve the accuracy of athletes’injury prevention
and treatment, make the accuracy more than 90%, shorten the diagnosis time less than 10 minutes, and meet the requirements
of athletes’injury diagnosis under normal rowing competition.
1. Introduction
With the continuous improvement of rowing intensity and the
publication of “Rowing Standards in Colleges and Universi-
ties”and “Guidelines for Diagnosing Athletes’Injuries,”the
prevention and treatment of rowers’injuries are becoming
increasingly severe and showing a rapid development trend
[1]; so, it is particularly important to diagnose athletes’injuries
reasonably. Athletes’injury diagnosis is a comprehensive
judgment of competition level, competition requirements,
and competition evaluation. Multiple regression analysis can
not only judge the diagnostic characteristics of athletes’
injuries but also solve the problems of massive injuries and
complex injuries, which is the main method of comprehensive
judgment at present. Literature research shows that multiple
regression analysis can classify athletes’injuries, preliminarily
judge the relationship between different scenes, and improve
the overall diagnosis complexity of intelligent teachers. The
specific results are shown in Figure 1.
It can be seen from Figure 1 that the support rate of
athletes’injury prevention and control in 2022 is increasing
year by year, which is significantly better than that in 2019,
indicating that multiple regression has become the trend of
future research. However, in the process of multiple regres-
sion diagnosis of athletes’injuries, the complexity of massive
injury causes [2], the initial prevention and treatment of
injury causes, and the selection of related parameters will
have an impact on the diagnosis results of injury causes. At
the same time, the data in Figure 1 shows that the analysis
of athletes’injury causes increased year by year in 2019,
2020, and 2021, with the largest increase in 2022, and it
has become a hot research direction. In terms of the propor-
tion of support and opposition, the proportion will be 50%
in 2020 and will gradually increase in 2021 and 2022.
Hindawi
Journal of Environmental and Public Health
Volume 2022, Article ID 4896336, 12 pages
https://doi.org/10.1155/2022/4896336
Therefore, the research on the causes of athletes’injuries is
hotter in 2022, which also suggests that the prevention and
treatment of athletes’injuries have been paid attention to
by scholars at home and abroad. Some scholars put forward
the method of prevention and control effect and determined
the proportion of subjectivity and complexity by analyzing
the normality of residual error, so as to judge the feasibility
of multiple regression. Multivariate regression analysis
method can be used for comprehensive analysis of multiple
factors and get the influence of multiple factors on the
results. At the same time, the process of multiple regression
analysis is a progressive process, which can constantly adjust
the calculation index, find out many factors of sports loss,
and find out the main influencing factors. At present, there
are many reasons for athletes’losses, and the treatment
mechanism is complex; so, a single calculation method can-
not accurately identify them. In order to prevent and control
sports loss better, it is necessary to carry out the intelligent
comprehensive analysis method to improve the accuracy of
analysis results. The multivariate regression analysis method
has the problem of large transformation error in local and
global analysis; so, it needs to be combined with the residual
algorithm to realize the transformation between different
data analysis. The multivariate regression analysis method
can be used for comprehensive analysis of multiple factors,
but there are some errors in the transformation of multivar-
iate analysis and multivariate regression, which affect the
analysis results of causes. The residual algorithm determines
the error results of different factors and the degree of error
by analyzing the residual of data and eliminates the factors
that have great influence on the results. In addition, the
residual algorithm can find out the factors with large errors,
analyze the weights and thresholds of the factors, and finally
sort the influencing factors to get more accurate comprehen-
sive results. Therefore, the combination of multiple regres-
sion analysis and residual analysis can not only improve
the accuracy of calculation but also improve the credibility
of results, which is suitable for analyzing the causes of
athletes’injuries. At the same time, residual analysis can
integrate subjective judgment factors to avoid the misunder-
standing of overcalculation. At present, rower injury is a hot
research issue at home and abroad, but the reasons and
methods of its research have been controversial, and no
effective evaluation method has been found. Therefore, look-
ing for an effective analysis method is an urgent problem to
be solved at present, and finding a scientific and reasonable
analysis method is the trend of future research. However,
the feasibility of the control effect method is controversial
and lacks corresponding evidence [3]. In order to verify
the accuracy of the above analysis, this paper combines
multiple regression analysis with the residue optimization
model to optimize the cause of injury in rowing competition.
2. Related Concepts
2.1. Multiple Regression Analysis. Multiple regression is a
comprehensive analysis method, which mainly finds out
the main factors from complex factors and eliminates the
secondary factors. At the same time, multiple regression
realizes the adjustment and analysis of relevant indicators
through progressive regression. In the process of regression
analysis, the weights of different reasons are determined,
and the corresponding regression equations are constructed.
Multivariate regression is an analysis method based on uni-
tary regression, which can analyze more comprehensively
and improve the objectivity of analysis results. Compared
with ant colony algorithm and domino analysis method,
the multiple regression method is simpler and more effective
in multifactor analysis, which is suitable for simple speech
processing of complex problems. Multivariate regression
29
250
200
150
100 100
32
18
50
100
56
15
50
0
Summary in 2019
Support...
Opposition to...
Neutrality
Summary in 2022
Support...
Opposition to...
Neutrality
Figure 1: The prevention and treatment investigation of athletes’injuries (data source: CNKI, Economic Statistics Yearbook).
2 Journal of Environmental and Public Health
method is widely used in machinery, construction, economy,
and other fields but less applied in sports injuries. However,
multiple regression analysis can comprehensively judge the
causes of injury, infer the influence of different causes on
injury, provide theoretical support for injury prevention
and control, and make prevention and control measures
more reasonable. Therefore, the application of multiple
regression analysis in sports injuries is more reasonable.
Multiple regression analysis is a comprehensive analysis
method, which analyzes the structure of rowing competition
by using characteristic injury causes, integrates the injury
causes, and realizes the prevention and control system [4].
Multivariate regression is widely used in electric power,
computer, and other fields, which can accurately calculate
the relationship between rowing races, judge the logic
between diagnostic structures, and finally improve the accu-
racy of calculation results. In order to further explain the
diagnosis of rowing competition, the following hypotheses
are put forward.
S=λ⋅Lx
i,yi,zi
ðÞ
−w⋅Kd
i,qi,bi
ðÞ
:ð1Þ
Hypothesis 1. Preventive measures of rowing competition g
=xi, the relationship ribetween injury causes, the require-
ments qiof prevention and control of injury causes, and
the influencing factors biof rowing competition. The diag-
nostic function before the analysis of sports injury causes is
Kðdi,qi,biÞ, and the diagnostic function after optimization
is Lðxi,yi,ziÞ[5]. If the value of Sis larger, the optimization
effect is better, as shown in formula (1).
Among them, wis the design weight before optimization
and the standard after optimization λ. At that time S=1, the
diagnostic effect of prevention and treatment measures for
sports injuries was the best, and the difference between the
results before and after optimization was the largest [6]; At
that time S=0, the optimization results were the worst,
and the results before and after were the smallest.
Theorem 2. The prevention and treatment effect coefficient ξi
represent the influencing factors of sports injuries, which
shows that the prevention and treatment measures of sports
injuries are more influential than external factors. The smal-
lest value represents the smallest influence degree; otherwise,
it is the largest, as shown in Formula (2).
ξi+1=α⋅ξi+β⋅ξi:ð2Þ
Among them, α⋅ξiis subjective residual, and β⋅ξiis
objective residual.
Theorem 3. If the athlete is injured Kðdi,qi,biÞ=φðdiÞ′⋅φ
ðqiÞ′⋅φðbiÞ′, then formula (1) can be expressed as formula (3).
S=λ⋅Lx
i,yi,zi
ðÞ
−w⋅φdi
ðÞ
′⋅φqi
ðÞ
′⋅φbi
ðÞ
′,ð3Þ
where φðdiÞ′,φðqiÞ′,φðbiÞ′are derivatives of φðdiÞ,φ
ðqiÞ,φðbiÞ.
Equation Formula (3) realizes the projection of damage
cause diagnosis, as shown in Figure 2.
Figure 2 shows that the damage cause in formula (1) is
projected onto the plane by formula (3), which further sim-
plifies the related damage cause and achieves the preliminary
purpose of preventing and treating the damage cause. The
projection of points a, b, and c in two-dimensional space
in figure 2 represents the standardized processing of data.
In addition, the data in Figure 2 is discrete data, and the
projected data is directional and belongs to vector data.
However, the projected data is only the corresponding value,
and the attribute of the data itself is proposed, which belongs
to the data cleaning process.
2.2. Mathematical Description of Athletes’Injury Causes. The
prevention and control effect mainly analyzed the key indi-
cators, including competition reasons [5], prevention and
control methods, training time, and injury location [6]. In
the control system, the number of control measures is the
same in rowing competition [7], and different optimization
degrees represent the best measures. Firstly, the injury causes
and prevention measures of rowing competition are randomly
obtained, the optimization results of injury causes are opti-
mized among the measures with better matching degree, and
the final optimization results are determined by screening
one by one. Then, multiple regression adopts the combination
form of probability calculation to select the best measure and
give the best weight [8]. Finally, the optimization methods that
do not meet the standards are eliminated.
Assuming that the initial number of prevention
measures is nin rowing competition, the prevention
measure function L=ðxi,yi,ziÞof rowing competition xi,yi
represents plane coordinates, zirepresents influencing
factors, and the overall prevention measures of rowing com-
petition are shown in Formula (4).
Lixi,yi,zi
ðÞ
=w⋅Kd
i,qi,bi
ðÞ
+ rand 0, 1
ðÞ
Kd
i,qi,bi
ðÞ
:ð4Þ
Among them, xi,yi, and ziare arbitrary measures, and
rand ð0, 1Þis the random control measure selection func-
tion. Randomly select the causes of sports injure, carry out
crossanalysis [9], and update the analysis degree. Under
the constraint of matching degree, the optimal measure con-
forming to the weight is obtained through the combination
form, and the calculation process is shown in Formula (5).
Lixi,yi,zi
ðÞ
×Li‐1xi‐1,yi‐1,zi‐1
ðÞ
=w⋅Kd
i,qi,bi
ðÞ
2+w⋅λ⋅Kd
i,qi,bi
ðÞ
+λ⋅Kd
i,qi,bi
ðÞ
2
:
ð5Þ
Among them, the change process of formula (5) is
shown in Figure 3.
Multivariate regression is to realize the optimum of rowing
competition in the form of probability piand combination and
carry out neighborhood analysis on the optimal measures to
obtain the final measures that meet the standards [9]. The
calculation process is shown in Formula (6).
3Journal of Environmental and Public Health
Lixi,yi,zi
ðÞ
=pi⋅Kd
i,qi,bi
ðÞ
∑n
i,j,kKΔdio,Δqio ,Δbi
ðÞ
:ð6Þ
Among them, piis the occurrence probability of differ-
ent measures.
If the rowing has not been analyzed for many times, the
relevant preventive measures will be eliminated. Multivariate
regression will continue to analyze other prevention mea-
sures [10], and according to formulas (1)~(4), a new cause
analysis of sports injuries will be carried out [11].
The dynamic adjustment of analysis accuracy is as fol-
lows. When analyzing the causes of sports injuries, multiple
regression cannot guarantee the overall analysis, which may
lead to local analysis and reduce the overall performance of
rowing competition [12]. Therefore, in the process of ana-
lyzing the causes of sports injuries, we should try our best
to expand the scope of analysis, make in-depth analysis of
each prevention and control measure in the optimal mea-
sures, and constantly adjust the accuracy. Assuming that
the accuracy of the analysis is spiral, in order to reduce ran-
domness and locality, a three-dimensional analysis of the
accuracy should be carried out, and a dynamic adjustment
factor should be introduced for analysis. The process is
shown in formula (7).
J=k⋅line xi,yi,zi
ðÞ
:ð7Þ
Among them, kiis the I-order adjustment factor, and
lineðxi,yi,ziÞis a linear relationship function. The control
system method of residual error and multiple regression is
shown in Formula (8).
Lx
i,yi,zi
ðÞ
=k⋅line xi,yi,zi
ðÞ
+w⋅Kd
i,qi,bi
ðÞ
+k⋅ξ:ð8Þ
In the analysis of sports injury causes, the value ξis rela-
tively small, and the value kis relatively large, so as to expand
the analysis range [13] and keep the diversity of injury causes.
In the later stage of analysis, the value ξis relatively large, and
kis relatively small, which deepens the accuracy of analysis,
improves the mining ability of analysis measures, and
enhances the performance of multiple regression and control
effect. The whole process is shown in Figure 4.
It can be seen from Figure 4 that multiple regression and
control effect belong to a spiral adjustment process, which is
a dynamic adjustment of accuracy and obtains a global
extreme value. The analysis measures of injury causes in
rowing competition are in multiple regression model, and
the overall fusion effect is better. Therefore, the dynamic
adjustment of analysis accuracy can make athletes’injury
causes meet the requirements and improve the accuracy of
calculation results.
Introduce coordination factors for analysis. Rowing
competition is collected many times, and the collection of
injury causes will reach the limit. Multivariate regression
needs continuous injury causes. Therefore, it is necessary
to find the blank reasons for injury in rowing competition
and complete the analysis measures for the reasons of injury
in rowing competition [14]. Because of the control effect,
randomness, antiuncertain factors, and local extremum are
eliminated. Therefore, it is necessary to increase the
X-axis
Z-axis
Y- ax i s
a
b
c
Figure 2: Projection of injury causes in rowing competition.
X1
X2
X3
X4
..
..
Xn
y1
y2
y3
y4
..
..
yn
Figure 3: Crossrelationship of related damage causes.
4 Journal of Environmental and Public Health
−1−0.8 −0.6 −0.4
0.6 0.8
x
1
0.4
−0.2
0.2
0
−1
−0.5
0
0.5
1
–1
–0.5
0
0.5
1
z
y
Local search process
Final search process
Output optimization results
×
Figure 4: Dynamic adjustment process of accuracy.
80
60
VarName3VarName4
40
80
60
40
20
VarName5
80
60
40
20
51015
20 25
Optimization process
30 35 40 45 50
Figure 5: Stage analysis process of athletes’injury causes.
5Journal of Environmental and Public Health
coordination factor, realize the joint calculation of damage
cause collection, and realize the integrity of damage cause
calculation. The specific calculation is shown in Formula (9).
lx
i,yi,zi
ðÞ
=〠
n
t=1
line xi,yi,zi
ðÞ
+w⋅Kd
i,qi,bi
ðÞ
+λ⋅Kd
i,qi,bi
ðÞ½
2
:
ð9Þ
When lðxi,yi,ziÞis 1, it shows that the calculation results
of the whole function are relatively complete, and there is no
abnormality in the process of collecting damage causes.
When lðxi,yi,ziÞis 0, it shows that the damage cause has
reached the limit in the whole collection process, and the
coordination of the damage cause is better [15]. When lðxi
,yi,ziÞis 0.5, it shows that the coordination of damage
causes and the collection process are better, and the whole
damage causes are in the best. In order to fill in the blank
of the cause of injury, the cause of injury should be filled,
as shown in Formula (10).
ΔLi= Cauthy 0, 1
ðÞ
Kd
ik,qik ,bik
ðÞ
:ð10Þ
Among them, it is the proportion of random injury causes.
2.3. Mathematical Model of Multiple Regression. Use the
combined model of multiple regression and control effects
to analyze the reasonable selection of multiple regression
mechanisms and control effects. Competition content, com-
petition process, competition feedback, and other prevention
and control systems [16] can not only balance the relation-
ship between global analysis and local analysis but also
improve the arrangement ability of classroom injury causes.
From formula (10), we can see that when analyzing the
causes of sports injuries, we should pay attention to the
global analysis ability, and when analyzing later, we should
pay attention to the local analysis; so, multiple regression
should realize the combination selection according to the
prevention and control effect. The specific combination is
as follows.
Table 1: Collaborative combination table of rowing competition.
Multifeature Matching comparison Combination Parameter optimization Residual optimization
Causes of injury
{completely, incomplete,
unknown} {1, 2, 3, 5, 4} {optimization, not
optimized}
{subjectivity, objectivity,
semisubjective}
{completely, incomplete,
unknown} {1, 2, 3, 5, 4} {optimization, not
optimized}
{subjectivity, objectivity,
semisubjective}
Training time
{completely, incomplete,
unknown} {1, 2, 3, 5, 4} {optimization, not
optimized}
{subjectivity, objectivity,
semisubjective}
{completely, incomplete,
unknown} {1, 2, 3, 5, 4} {optimization, not
optimized}
{subjectivity, objectivity,
semisubjective}
Result
characteristics
{completely, incomplete,
unknown} {1, 3, 5, 2, 4} {optimization, not
optimized}
{subjectivity, objectivity,
semisubjective}
{completely, incomplete,
unknown} {1, 1, 3, 2, 4} {optimization, not
optimized}
{subjectivity, objectivity,
semisubjective}
Enter scene data
Calculate the
matching value of
scene data
Determine the
competition
structure
Multiple
optimization of
competition data
reshold for
matching
competition
Determine the local
optimal scheme
Get the global
optimal scheme
Whether the optimization
maximized
Output optimal
scheme
Data results for
training
Y
No
Whether the data structure
should be optimized
Y
N
Figure 6: Implementation process of multiple regression and
control effect.
6 Journal of Environmental and Public Health
100101102103
10−2
10−1
100
101
Number of iterations of parameter (times)
Optimization degree of parameters (%)
Figure 7: Parameter analysis process.
1
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
234
Number of sub-features
Feature fusion degree (100%)
56
Content characteristics
Process characteristics
Result characteristics
×
×
×
×
×
×
×
×
×
×
×
×
×
×
×
Figure 8: Multiple regression of injury causes.
7Journal of Environmental and Public Health
10
90
120
130
0.5
150
65
70
30
40
80
Semi-
structural Comprehensive
Other
properties Objectivity Subjectivity
Semi-
subjectivity
20 30 40 50
Optimization times
60 70 80 90 100
Figure 9: Residual analysis results.
Table 2: Result test of injury cause.
Damage content Algorithm Prevention and
control measures
Classification
of control
Control
time
Proportion
of control
Global
optimal
measure
Theoretical
optimal
measure
Training is not in
place
Multiple regression and
residual error algorithm 12 6 4 89 3 3
Multiple regression
algorithm 7525613
Poor recovery
activities
Multiple regression and
residual error algorithm 11 6 6 92 3 3
Multiple regression
algorithm 4217023
The training plan
is not rigorous
Multiple regression and
residual error algorithm 16 8 8 90 3 3
Multiple regression
algorithm 6336723
Unreasonable
policy
management
Multiple regression and
residual error algorithm 10 10 10 98 3 3
Multiple regression
algorithm 3528213
8 Journal of Environmental and Public Health
(1) The damage cause analysis measures of information
characteristics, as shown in Formula (11).
ΔLixi,yi,zi
ðÞ
= Cauthy 0, 1
ðÞ
⋅p⋅Kd
ik,qik ,bi−1k
ðÞ
:ð11Þ
(2) The damage cause analysis measures of content
characteristics, as shown in Formula (12).
ΔLixi,yi,zi
ðÞ
= Cauthy 0, 1
ðÞ
g⋅Kd
i−1k,qi−1k,bi−1k
ðÞ
:ð12Þ
(3) The damage cause analysis measures of content
characteristics, as shown in Formula (13).
ΔLixi,yi,zi
ðÞ
= Cauthy 0, 1
ðÞ
⋅g⋅Kd
i−1k,qi−1k,bi−1k
ðÞ
∀p
⋅Kd
i−1k,qi−1k,bi−1k
ðÞ
:
ð13Þ
Table 3: Accuracy of collecting damage cause judgment.
Classification Number of injury causes (number) Accuracy (%)
The preparation activities before training are not in place 43 98.05
The recovery activities after training are not done well 31 99.90
Life is not self-disciplined, which leads to physical fatigue 23 97.79
The training plan is not rigorous 98 98.26
Table 4: Contents of different research projects.
Research project Content
Cause of injury
The preparation activities before training are not in place
The recovery activities after training are not done well
Life is not self-disciplined, which leads to physical fatigue (staying up late, fatigue training,
irregular diet)
The training plan is not rigorous
The policy management is unreasonable. Colleges and universities have no special policies on
study, work, and rest for sports teams, which leads to insufficient training time
The level of coaches needs to be improved
Diet and nutrition cannot keep up after a lot of exercise
Control method
Precompetition preparation activities
Postmatch recovery, stretching, and rehabilitation exercises
Strict work and rest time and diet
Hire high-level coaches
Hire high-level coaches
Injury ratio of college rowers
The injury rate after three years of practice is 75%
The injury rate of rowing practice for two years is 60%
The injury rate of rowing practice for two years is 30%
Classification of injuries and injuries
of college rowers
Lumbar injuries, including sprain, strain, and lumbar disc herniation, are 80%
Knee injuries account for 10%
Table 5: Accuracy of judging athletes’injury prevention and treatment by different methods.
Prevention and control measures Multivariate regression and residual
model
Multiple
regression
Residual
error
Precompetition preparation activities 99.94 98.93 97.92
Postmatch recovery, stretching, and rehabilitation exercises 99.97 98.92 98.97
Strict work and rest time and diet 99.91 95.94 95.93
Hire high-level coaches 99.94 97.92 94.93
The policy of colleges and universities is inclined, leaving enough
training time 98.21 97.23 96.36
9Journal of Environmental and Public Health
In this paper, there are two main improvements in the
prevention and treatment system: at the same time, the pro-
cess of prevention and control system should be continu-
ously standardized to improve the overall analysis ability
[17]. On the other hand, the coordination factor is used to
realize the balance between global analysis and local analysis,
and through the matching function, the athletes’injury
causes and prevention measures are analyzed to improve
the accuracy of the final results of injury causes. The results
are shown in Figure 5.
The synergistic combination of athletes’injury causes is
as follows: collaboration is the main way to realize dynamic
analysis in rowing competition. The algorithm in this paper
is based on the prevention effect of multiple regression and
diagnoses the dynamic distributed cooperative combination
form [18]. Different submultiple regressions use different
cooperative combination forms, complex parameters, and
operations. Multiple regression was randomly divided into
five characteristics, each of which represented a series of
injury causes. In each analysis process, features will ran-
domly select collaborative measures. After each athlete’s
injury cause is collected, compare the matching degree
between different characteristics and injury cause analysis,
analyze the complexity, and record the global optimal mea-
sures; Other sub-multifeatures use optimal measures to
compare the analysis degree of different measures, and the
results are shown in Table 1.
2.4. Multiple Regression and Prevention and Control Effect on
Prevention and Control Measures and Steps of Athletes’
Injury Causes. The basic idea of multiple regression and
prevention effect is to use different synergy and combination
forms to realize damage cause analysis [18], obtain global
optimal measures, and improve the accuracy of measure
formulation. The implementation steps of this algorithm
are shown in Figure 6.
Step 1. Determine the structure and complexity of damage
cause analysis. According to the characteristics of damage
causes, the analysis measures are determined.
Step 2. In the prevention and control system, determine the
relevant parameters, analyze that the prevention and control
measures are consistent with the number of damage causes,
and set up functions.
Step 3. Set the match function. By means of multiple regres-
sion and prevention effect, this paper analyzes the causes of
athletes’injuries. Determine the weight wand standard λfor
later comparison. At the same time, the matching measures
of rowing competition are compared to determine the local
optimal matching measures.
Step 4. Calculate the global and local optimal measures. Mul-
tivariate regression is divided into five features, the matching
degree is obtained, and the optimal global measures and
local optimal measures are compared.
Step 5. Comparison of analysis degree is as follows.
According to the change of rowing competition, the five
characteristics dynamically adjust the analysis factors and
randomly select the combination forms to realize the
stability of local analysis.
Step 6. Dynamic coevolution of each Dortmund is as follows.
After the control system measures are determined, the opti-
mal global measures are selected, different measures are
compared, and irrelevant measures are eliminated.
Step 7. Judge whether the injury cause of rowing competition
reaches the maximum. If it does not reach the maximum,
repeat steps 1-5; otherwise, stop analysis and return to the
global best measure.
3. Empirical Analysis of Prevention and
Control System
3.1. Analysis Results of Injury Causes in Rowing Competition
in Colleges and Universities. The analysis results of injury
causes in rowing competition in colleges and universities
are as follows:
(1) In analysis of parameters, the results are shown in
formula (14).
fx
ðÞ
=〠
i=1
k−Δk+ξ:ð14Þ
Among them, kis analysis parameters, ξis preven-
tion and control effects, and the analysis process is
shown in Figure 7.
As can be seen from Figure 7, the relationship between
the analysis times and the analysis degree of parameters is
basically linear, which shows that the parameters of injury
causes in rowing competition are in line with positivity
and can be analyzed later.
(2) In multiple regression analysis, the result is as shown
in formula (15).
fx
ðÞ
=〠
i=1
mi2−M:ð15Þ
Among them, mi2is the characteristic index, Mis the
characteristic index of local optimal measures, and the
results are shown in Figure 8.
As can be seen from Figure 8, in the process of damage
cause analysis, the fusion trend of multiple features is
basically the same, showing ups and downs. Among them,
negative change is reverse fusion, and positive change is
positive fusion. However, on the whole, the radian of multi-
ple regression is consistent, which shows that the effect of
multiple regression is better [19].
(3) In the control effect, the result is shown in
formula (16).
10 Journal of Environmental and Public Health
fx
ðÞ
= line + Δξ:ð16Þ
Among them, it is a linear function and a residual adjust-
ment, and the result is shown in Figure 9.
As can be seen from Figure 9, the overall analysis result
of residual error is good. The results of subjective and objec-
tive analysis are the best, followed by other nature, compre-
hensive, and semisubjective analysis. All the control effects
showed ups and downs, and there was no biased result.
The results are as shown in Table 2 in the process of 50
analysis of damage causes.
It can be seen from Table 2 that the multiple regression
and residual model are superior to the control effect, and
its global optimal measures are closer to the theoretical opti-
mal measures. Moreover, the range of measures and average
measures and calculation errors of multiple regression and
residual models are less than the control effect.
3.2. Prevention and Treatment of Athletes’Injuries in Rowing
in Colleges and Universities. In this paper, 12 colleges and
universities were selected as the research objects. After pre-
liminary arrangement of injury causes, 6 injury causes and
4 types of athletes’injury causes were obtained. According
to the diagnostic criteria of athletes’injuries, the judgment
results are shown in Table 3.
3.3. Accuracy of Different Damage Analyses. According to
the experiment, Xi’an JiaoTong University, Tianjin Univer-
sity of Science and Technology, Henan University of Finance
and Economics, Tongji University, Shanghai JiaoTong
University, North China University of Water Resources
and Hydropower, Nanchang University, Jiangxi Yichun
University, Jimei University, Tianjin Normal University,
Shanghai Ocean University, Shanghai Lixin School of
Accounting and Finance, Zhejiang University, and Peking
University are selected as the research objects. The analysis
is shown in Table 4.
In order to further prove the effectiveness of the model
proposed in this paper, other comparative models are intro-
duced for comparative analysis: (1) multivariate regression,
(2) multivariate regression combined residual, (3) residual
algorithm, (4) multivariate regression, and residual model
in this paper [20]. Methods 2 and 4 are not an algorithm.
Method 4 is based on method 2 and makes corresponding
improvement and perfection. The accuracy of different algo-
rithms is analyzed from the aspect of diagnosing the types of
athletes’injury causes from complex athletes’injury causes,
and the results are shown in Table 5.
It can be seen from the above table that the prevention
and treatment accuracy of multiple regression and residual
model is higher [21], and the accuracy does not change with
the change of athletes’injury cause types. The main reason is
that the dynamic analysis of residual factor on the preven-
tion measures of rowing competition makes its continuous
calculation time shorter and can change the types of athletes’
injury causes more flexibly [22]. Therefore, the residual
factor can not only reduce the impact of complexity on the
results but also quickly realize the accurate analysis of differ-
ent athletes’injury causes.
4. Conclusion
Regarding rowing athletes in the course of the process of
injury, how to effectively prevent injury is an urgent problem
to be solved. However, there are many reasons for rowers’
injuries [23], and it is difficult to obtain results through simple
calculation. In this paper, based on multiple regression, com-
bined with multiple regression method, a multiple regression
model is proposed. By setting standards, weights, and dynamic
cooperative combination forms, the causes of athletes’injuries
are classified, and finally, the regression results are obtained
[24]. At the same time, in order to avoid the excessive error
of multiple regression, the residual analysis method is inte-
grated to find out the causes of athletes’injuries and put
forward reasonable prevention measures [25]. MATLAB
simulation results show that the accuracy of the proposed
diagnosis algorithm is more than 90%, the convergence time
is less than 10 minutes, and the overall convergence is good,
converging at 10 iterations, which is significantly better than
other algorithms. The algorithm proposed in this paper adapts
to different types of athletes’injury causes and has high global
analysis ability. However, the research on the internal relation-
ship between injury causes is insufficient, and future research
will pay attention to the analysis of internal causes in order
to improve the algorithm in this paper.
Data Availability
The data used to support the findings of this study are
available from the corresponding author upon request.
Conflicts of Interest
The author declares no conflicts of interest.
Acknowledgments
This work was supported by the Tianjin University of
Science and Technology.
References
[1] J. Zhang, Q. Ding, L. Xiu-Liang, H. Yi-Wei, and Y. Yang, “Sup-
port vector machine versus multiple logistic regression for pre-
diction of postherpetic neuralgia in outpatients with herpes
zoster,”Pain Physician, vol. 25, no. 3, pp. E481–E488, 2022.
[2] N. Vora, R. Parikh, V. Chaudhari et al., “Optimization of bead
morphology for GMAW-based wire-arc additive manufactur-
ing of 2.25 Cr-1.0 Mo steel using metal-cored wires,”Applied
Sciences-Basel, vol. 12, no. 10, p. 5060, 2022.
[3] Z. C. Ru, J. P. Liu, M. Kadzinski, and X. W. Liao, “Bayesian
ordinal regression for multiple criteria choice and ranking,”
European Journal of Operational Research, vol. 299, no. 2,
pp. 600–620, 2022.
[4] C. Ricciardi, A. M. Ponsiglione, A. Scala et al., “Machine learn-
ing and regression analysis to model the length of hospital stay
in patients with femur fracture,”Bioengineering-Basel, vol. 9,
no. 4, p. 172, 2022.
[5] S. Ozlem and O. F. Tan, “Predicting cash holdings using super-
vised machine learning algorithms,”Financial Innovation,
vol. 8, no. 1, p. 44, 2022.
11Journal of Environmental and Public Health
[6] D. Nishioka, S. Kino, K. Ueno, and N. Kondo, “Risk profiles of
frequent outpatients among public assistance recipients in
Japan: a retrospective cohort study using a classification and
regression trees algorithm,”Bmj Open, vol. 12, no. 5,
p. e054035, 2022.
[7] M. Nabati, S. A. Ghorashi, and R. Shahbazian, “JGPR: a com-
putationally efficient multi-target Gaussian process regression
algorithm,”Machine Learning, vol. 111, no. 6, pp. 1987–2010,
2022.
[8] A. Mohsenijam, M. Lu, and S. Naumets, “Integrating model
tree and modified stepwise regression in concrete slump pre-
diction and steel fabrication estimating,”Canadian Journal of
Civil Engineering, vol. 49, no. 4, pp. 478–486, 2022.
[9] E. Malina, B. Veihelmann, M. Buschmann, N. M. Deutscher,
D. G. Feist, and I. Morino, “On the consistency of methane
retrievals using the Total Carbon Column Observing Network
(TCCON) and multiple spectroscopic databases,”Atmospheric
Measurement Techniques, vol. 15, no. 8, pp. 2377–2406, 2022.
[10] S. L. Liu, L. Yang, C. J. Zhu et al., “A method of identifying cell
suspension concentration based on bioimpedance spectros-
copy,”Acta Physica Sinica, vol. 71, no. 7, p. 078701, 2022.
[11] D. H. Liang, D. A. Frederick, E. E. Lledo et al., “Examining the
utility of nonlinear machine learning approaches versus linear
regression for predicting body image outcomes: The U.S. Body
Project I,”Body Image, vol. 41, pp. 32–45, 2022.
[12] L. X. Li, G. W. Huang, S. X. Xi, Z. G. Wang, and C. Z. Zhou,
“Application of fuzzy probability factor superposition algo-
rithm in nuclide identification,”Journal of Radioanalytical
and Nuclear Chemistry, vol. 331, no. 5, pp. 2261–2271, 2022.
[13] E. Kocak, E. Ayli, and H. Turkoglu, “A comparative study of
multiple regression and machine learning techniques for pre-
diction of nanofluid heat transfer,”Journal of Thermal Science
and Engineering Applications, vol. 14, no. 6, 2022.
[14] B. Ibrahim, A. Ewusi, I. Ahenkorah, and Y. Y. Ziggah, “Model-
ling of arsenic concentration in multiple water sources: a com-
parison of different machine learning methods,”Groundwater
for Sustainable Development, vol. 17, p. 100745, 2022.
[15] S. Haddad, A. Boukhayma, and A. Caizzone, “Continuous
PPG-based blood pressure monitoring using multi-linear
regression,”IEEE Journal of Biomedical and Health Informat-
ics, vol. 26, no. 5, pp. 2096–2105, 2022.
[16] Y. E. Guzelel, U. Olmus, K. N. Cerci, and O. Buyukalaca, “New
multiple regression and machine learning models of rotary
desiccant wheel for unbalanced flow conditions,”International
Communications in Heat and Mass Transfer, vol. 134,
p. 106006, 2022.
[17] M. R. Elkharbotly, M. Seddik, and A. Khalifa, “Toward sus-
tainable water: prediction of non-revenue water via artificial
neural network and multiple linear regression modelling
approach in Egypt,”Ain Shams Engineering Journal, vol. 13,
no. 5, p. 101673, 2022.
[18] M. K. Devi, V. P. Vemuri, M. Arumugam et al., “Design and
implementation of advanced machine learning management
and its impact on better healthcare services: a Multiple Regres-
sion Analysis Approach (MRAA),”Computational and Math-
ematical Methods in Medicine, vol. 2022, Article ID 2489116, 7
pages, 2022.
[19] S. Contant, G. Klein, J. Morelli, and T. W. Krause, “Inverse
algorithm for extraction of multiple parameters using analyti-
cal model of eddy current response,”Journal of Applied Phys-
ics, vol. 131, no. 20, p. 205103, 2022.
[20] C. Belei, J. Joeressen, and S. T. Amancio, “Fused-filament fab-
rication of short carbon fiber-reinforced polyamide: parameter
optimization for improved performance under uniaxial tensile
loading,”Polymers, vol. 14, no. 7, p. 1292, 2022.
[21] M. A. Ahmad, M. Elloumi, A. H. Samak et al., “Hiding
patients' medical reports using an enhanced wavelet steganog-
raphy algorithm in DICOM images,”Alexandria Engineering
Journal, vol. 61, no. 12, pp. 10577–10592, 2022.
[22] A. Elnady, A. A. Adam, and M. Nasir, “Efficient adaptive con-
trollers and recursive filters for centralized control scheme of
islanded microgrid with linear and non-linear loads,”Interna-
tional Journal of Electrical Power & Energy Systems, vol. 141,
no. 2, p. 108104, 2022.
[23] M. Gao, G. Yu, and C. N. Li, “Incipient gear fault detection
using adaptive impulsive wavelet filter based on spectral
negentropy,”Chinese Journal of Mechanical Engineering,
vol. 35, no. 1, p. 22, 2022.
[24] G. Hu, B. Du, H. A. Li, and X. P. Wang, “Quadratic interpola-
tion boosted black widow spider-inspired optimization algo-
rithm with wavelet mutation,”Mathematics and Computers
in Simulation, vol. 200, no. 3, pp. 428–467, 2022.
[25] S. Karasu and Z. Sarac, “The effects on classifier performance
of 2D discrete wavelet transform analysis and whale optimiza-
tion algorithm for recognition of power quality disturbances,”
Cognitive Systems Research, vol. 75, no. 11, pp. 1–15, 2022.
12 Journal of Environmental and Public Health
Available via license: CC BY
Content may be subject to copyright.