Virtual game vs Physical game

1st Robin Das 2nd Sourov Biswas

dept. Computer Science and Engeneering dept. Computer Science and Engineering
Daffodil International University Daffodil International University
Dhaka, Bangladesh Dhaka, Bangladesh
Email: robin15-4153@diu.edu.bd Email: sourov15-4161@diu.edu.bd

Abstract—Today boys and girls are addicted to virtual sports those who participate primarily in physical sports.A more in-
leaving physical sports, the main purpose of this paper is the depth grasp of the socioeconomic elements that influence these
difference between virtual games and physical games and the outcomes.
reasons for virtual games addiction, what percentage of boys and
girls are now addicted to virtual games and what percentage are ”eSports and Physical Sports: Exploring the Motivations
addicted to physical games and the benefits of virtual and physical and Experiences of Youth in Online Competitions”Dr. Michael
games And disadvantages will be highlighted. So we collected Roberts Qualitative interviews and surveys.To explore the
some data and we predicted some results using 6 algorithms motives and experiences of young people participating in
using machine learning, LogisticRegression and SVC got 100 eSports and traditional physical sports, with a focus on social
percent accuracy, DecisionTreeClassifier, RandomForestClassifier
and GaussianNB got 75 percent accuracy and KNeighborsClas- and psychological factors. A larger cultural and geographical
sifier got 87 percent accuracy. context for a more comprehensive examination.
”Gamification of Physical Activity: A Comparative Study of
I. I NTRODUCTION Virtual and Real-World Incentive Systems” Dr. Sarah Johnson
Experiment with gamified interventions and surveys. To eval-
Nowadays boys and girls are turning towards virtual games. uate the effectiveness of gamification in promoting physical
It can be seen that from small children to adult boys and girls activity in both virtual and physical sports situations. An in-
are now leaning towards virtual games. The reason for this depth evaluation of the long-term sustainability of gamified
is that they are getting mobile phones in their hands from a therapies.
young age. Because of virtual games, no one wants to play any ”Virtual Gaming Addiction and Its Influence on Physical
physical game in the field anymore. Phone or PC is available Sports Engagement: A Mixed-Methods Approach” Dr. John
with great games like PUBG for mobile, Free Fire, COC, Call Williams Questionnaires, diagnostic instruments, and qualita-
of Duty and more addictive games like Grand Theft Auto tive interviews. To investigate the relationship between virtual
V, The Witcher 3: Wild Hunt, Apex Legend, Dishonored 2 gaming addiction and the fall in physical sports participation,
for PC. All addictive games like this. They spend their time with an emphasis on the underlying issues. Comprehensive
playing these games. Due to which many boys and girls are research on prevention and intervention measures.
becoming less addicted to physical games. These games are ”Virtual Sports Simulators in Education: Enhancing Skills
also called A-Support. And the joker reason behind virtual and Knowledge in Virtual and Physical Sports” Dr. Maria
games is that they can play at home or in bed without going Martinez Educational simulations and assessments using vir-
anywhere and not wasting body energy. So we researched tual sports simulators. To assess the educational potential of
virtual games and physical games using machine learning virtual sports simulators in developing virtual gaming abilities
research. Through this we can learn many things. Predictions as well as practical sports knowledge. Assessments of real-
or decisions are taken through machine learning. This paper world applicability in educational settings. ”Parental Guidance
will help us to increase our knowledge. And through this in Balancing Virtual Gaming and Physical Sports for Children:
research, we can know how much the children of today are An Exploratory Study” Dr. James Turner Parental surveys
at risk with this virtual game. We can always learn and know and interviews. To learn how parents view and maintain
two things here and we can warn how much boys and girls the balance between virtual gaming and physical sports for
are inclined towards virtual games and pros and cons and pros their children. In-depth insights into cultural and generational
and cons of physical games. disparities in parenting practices.
”Exergaming and Physical Activity Promotion in Adoles-
II. LITERATURE REVIEW cents: A Comparative Analysis” Dr. Laura Clark Controlled
experiments comparing physical activity during exergaming
”Virtual Sports vs. Physical Sports: A Comparative Analysis and traditional gaming. To evaluate the effectiveness of ex-
of Adolescent Health Outcomes”Dr. Emily Anderson Longi- ergaming in promoting physical activity among adolescents.
tudinal study using health assessments and questionnaires.To Comprehensive long-term health effect assessments.
evaluate the long-term physical and mental health results ”Peer Influence and the Choice Between Virtual Gaming
of adolescents who participate primarily in virtual sports vs and Physical Sports Among Teens: A Social Network Anal-
ysis” Dr. Robert Harris Surveys, social network analysis, and
interviews. To investigate the effect of peer groups and social
dynamics in developing adolescent preferences for virtual
gaming or physical sports. A global perspective that takes into
account cross-cultural differences in peer effects.
”Emerging Trends: Virtual Sports Leagues and Their Soci-
etal Implications” Dr. Jennifer Lee Analysis of content from
virtual sports leagues, questionnaires, and participant inter-
views. To investigate the rise and influence of virtual sports
• The logistic function starts near 0 for very negative values
leagues (eSports) on traditional sports participation among
of the linear combination, then gradually increases and
young adults. A thorough examination of the economic and
approaches 1 as the linear combination becomes positive.
cultural implications.
”The Technology and Screen Time Paradox: Analyzing Its It has an S-shaped curve.
Effects on Physical Activity Patterns Among the New Gen- The logistic function is characterized by its S-shaped curve,
eration” Dr. William Brown Longitudinal research, question- which makes it suitable for modeling probabilities in binary
naires, and screen time monitoring. To investigate the broader classification problems. The threshold for classification (e.g.,
effects of technology and screen time, including virtual gam- deciding whether an observation belongs to the positive class
ing, on juvenile physical activity and sedentary behavior. or not) is typically set at 0.5, but it can be adjusted based
Strategies for reducing negative impacts and fostering balanced on the specific problem and trade-offs between precision and
screen time. recall.

III. DATA GATHERING Decision Trees: A Decision Tree is a supervised machine

In data gathering, we opened an online Google form and learning algorithm used for both classification and regression
collected data from the dataset of research papers related to tasks. It is a tree-like structure that recursively splits the dataset
short videos by filling this form by the students of college and into subsets based on the most significant attribute or feature,
university with related questions. We have created 9 questions with the goal of making predictions or decisions. There is
and the question (e.g. ’Sports’, ’Favorite game’, ’Favorite no specific mathematical formula for a Decision Tree like
player’, ’Dislike game’, ’Play Time’, ’Spend time’, ’Reasons there is for some other machine learning algorithms. Instead,
Inactive in Sports’, ’Donot participate in sports reasons of Decision Trees are constructed using a recursive algorithm
mobile games?’, ’Reasons of sports reduced’, and the number that repeatedly selects the best attribute to split the data and
of 52 people who responded to this online survey. The identity create branches in the tree. The general algorithm follows this
of the participants was kept secret during the study and all data recursive pattern:
were handled with privacy. • Select the best attribute (feature) to split the data based
on a certain criterion (e.g., Gini impurity for classification
IV. METHODOLOGY
or mean squared error for regression).
Logistic Regression: Logistic Regression is a statistical • Create a branch for each possible outcome of the selected
model used for binary classification tasks, where the output attribute.
variable is a binary (two-class) categorical variable. It’s a • Recursively repeat steps 1 and 2 for each branch until a
popular and widely used machine learning algorithm that stopping condition is met, such as a maximum depth or
models the probability of a binary outcome as a function of one a minimum number of samples in a node.
or more predictor variables. The Logistic Regression model • Assign a prediction or decision label to each leaf node
uses the logistic function (also called the sigmoid function) to based on the majority class (for classification) or the mean
model the probability of the positive class (usually denoted as value (for regression) of the samples in that leaf.
”1”). The logistic function is defined as follows: Where:
A graphical representation of a simple Decision Tree might
• P(Y=1) is the probability of the positive class.
look like this:
• b0,b1,b2,. . . ,bn are the coefficients of the model.
• X1,X2,. . . ,Xn are the predictor variables.
• e is the base of the natural logarithm, approximately equal
to 2.71828.
The coefficients b0,b1,b2,. . . ,bn are estimated from the train-
ing data using methods like maximum likelihood estimation.
Where,
• The x-axis represents the linear combination
2b0+b1X1+b2X2, which can take any real value.
• The y-axis represents the probability P(Y=1), which is
constrained to be between 0 and 1. SVM Algorithm: A Support Vector Machine( SVM) is a
supervised machine learning algorithm used for bracket and
retrogression tasks. SVM is particularly effective for double
bracket problems and is known for chancing the optimal
hyperplane that stylish separates data points into different
classes while maximizing the periphery between the classes.
It can also be extended to handle multiclass bracket.
Equation (for Linear SVM): The equation for a linear SVM
model can be described as follows: For a binary classification
problem with a hyperplane represented by wx+b=0
where:
• w is the weight vector.
• x is the feature vector.
• b is the bias or intercept term. • The predictions of all the trees are combined (e.g., by ma-
jority vote for classification or averaging for regression)
The decision boundary is defined by: wx+b=0 The classifica- to make the final prediction.
tion decision is made based on the sign of wx+b: • The ”Random” aspect comes from the fact that each
• If wx+b¿0, the data point is classified as one class. decision tree is trained on a different subset of the
• If wx+b¡0, the data point is classified as the other class. data and may consider only a subset of the features,
Here’s a simplified graphical representation of a linear SVM introducing diversity into the ensemble.
decision boundary: KNN Algorithm: The k- Nearest Neighbors( k- NN) al-
In this graph: gorithm is a supervised machine learning algorithm used
• The blue and red points represent data points belonging for bracket and retrogression tasks. It’s a simple and in-
to two different classes. tuitive algorithm that makes prognostications grounded on
• The solid line represents the optimal hyperplane that the similarity between a new data point and its k- nearest
maximizes the margin between the classes. neighbors in a point space. In bracket, it assigns the class
• The dashed lines parallel to the hyperplane are support marker that’s most common among the k- nearest neighbors,
vectors. while in retrogression, it predicts a numerical value grounded
on the normal( or ladened average) of the target values of
The thing of SVM is to find the hyperplane( in this case, the
those neighbors. The Equation is: From the above equation
solid line) that maximizes the periphery between the classes
while minimizing bracket crimes. Support vectors are data
points closest to the hyperplane and play a pivotal part

x=Input features or independent variables, y=Target variable

or dependent variable

It’s important to note that the example above illustrates

a linear SVM, but SVMs can also be used with non-linear
kernels to handle complex data distributions where a linear
boundary is insufficient. In such cases, the decision boundary
can take on more complex shapes, such as curves or ellipses,
depending on the chosen kernel function. • The blue and red points represent data points belonging
Random Forest Algorithm: Random Forest is an ensemble to two different classes (e.g., ”A” and ”B” for a binary
machine learning algorithm used for both bracket and ret- classification problem).
rogression tasks. It’s grounded on the idea of constructing • The green point is the new data point that needs to be
multiple decision trees during training and combining their classified.
prognostications to ameliorate overall delicacy and reduce • The number ’k’ is the hyperparameter representing the
overfitting. Random Forest is known for its robustness, ver- number of nearest neighbors to consider (e.g., k=3).
satility, and capability to handle complex datasets. GaussianNB Algorithm: One probabilistic classification al-
In this conceptual representation: gorithm in the Naive Bayes family is Gaussian Naive Bayes.
• Each tree within the Random Forest is an individual This approach, which is specifically designed for situations
decision tree, depicted as a node in the forest. in which features are continuous and presume a Gaussian
(normal) distribution, makes use of the ”Naive” assumption,
regarding features as conditionally independent in light of the
class label. Because of this assumption, GaussianNB performs
particularly well in classification applications by simplifying
the computation of likelihood.

Fig. 1. shows the classification report of Decision Trees

precision recall f1-score support

1 1.00 0.86 8
accuracy 0.75 8
macroavg 0.50 0.38 0.43 8
weightedavg 1.00 0.75 0.86 8

where,
100 percent. The confusion matrix and class-specific metrics
• x is the value of the variable.
provide a thorough evaluation of its prediction ability. These
• The distribution mean is represented by .
results are essential for successfully identifying and handling
• The variance of the distribution is represented by 2 2.
the problem Figure-3 Shows the confusion matrix and The
V. EXPERIMENTAL ANALYSIS table shows the classification report.
In identifying Virtual game and Physical game addiction,
our data shows that the Logistic Regression approach had
an accuracy of 100 percent. The confusion matrix and class-
specific metrics provide a thorough evaluation of its prediction
ability. These results are essential for successfully identifying
and handling the problem Figure-1 Shows the confusion
matrix and The table shows the classification report.

Fig. 2. shows the classification report of SVC

precision recall f1-score support

1 1.00 1.00 1.00 8
accuracy 1.00 8
macroavg 1.00 1.00 1.00 8
weightedavg 1.00 1.00 1.00 8

In identifying Virtual game and Physical game addiction,

precision recall f1-score support our data shows that the Random Forest Algorithm approach
1 1.00 1.00 1.00 8 had an accuracy of 87 percent. The confusion matrix and
accuracy 1.00 8
macroavg 1.00 1.00 1.00 8 class-specific metrics provide a thorough evaluation of its
weightedavg 1.00 1.00 1.00 8 prediction ability. These results are essential for successfully
identifying and handling the problem Figure-4 Shows the
In identifying Virtual game and Physical game addiction, confusion matrix and The table shows the classification report.
our data shows that the Decision Trees approach had an ac-
curacy of 75 percent. The confusion matrix and class-specific In identifying Virtual game and Physical game addiction,
metrics provide a thorough evaluation of its prediction ability. our data shows that the KNN Algorithm approach had an ac-
These results are essential for successfully identifying and curacy of 87 percent. The confusion matrix and class-specific
handling the problem Figure-2 Shows the confusion matrix metrics provide a thorough evaluation of its prediction ability.
and The table shows the classification report. These results are essential for successfully identifying and
In identifying Virtual game and Physical game addiction, handling the problem Figure-5 Shows the confusion matrix
our data shows that the SVC approach had an accuracy of and The table shows the classification report.
 precision recall f1-score support
0 0.00 0.00 0.00 0.00
1 1.00 0.88 0.93 8
accuracy 0.88 8
macroavg 0.50 0.44 0.47 8
weightedavg 1.00 0.88 0.93 8

precision recall f1-score support

0 0.00 0.00 0.00 0.00
1 1.00 0.75 0.86 8
accuracy 0.75 8
macroavg 0.50 0.38 0.43 8
weightedavg 1.00 0.75 0.86 8

Fig. 3. shows the classification report of Random Forest Algorithm

precision recall f1-score support

of its own. While playing virtual games offers unparalleled
0 0.00 0.00 0.00 0.00 technological advancements and an immersive retreat, playing
1 1.00 0.88 0.93 8 physical games promotes social interaction, physical activity,
accuracy 0.88 8
macroavg 0.50 0.44 0.47 8
and cognitive development. These two worlds will likely col-
weightedavg 1.00 0.88 0.93 8 lide in video games of the future, creating hybrid experiences
that will push the boundaries of play as technology advances.

Fig. 4. shows the classification report of KNN Algorithm

In identifying Virtual game and Physical game addiction,

our data shows that the GaussianNB Algorithm approach
had an accuracy of 75 percent. The confusion matrix and
class-specific metrics provide a thorough evaluation of its
prediction ability. These results are essential for successfully
identifying and handling the problem Figure-6 Shows the
confusion matrix and The table shows the classification report.

Fig. 5. shows the classification report of GaussianNB Algorithm

VI. C ONCLUSION
Games, both real and virtual, are distinct but complementary
forms of entertainment, each having benefits and pleasures