ArticlePDF Available

Abstract and Figures

This study tries to find out the best model for prediction video game rate categories. A representation from four rating categories (everyone, everyone 10+, teen, mature) was used for the analysis. The paper follows CRISP-DM approach under Rapid Miner software to business and data understanding, Data preparation, model building and evaluation. The researchers compared prediction among six model and the results showed that the Generalized Linear Models (GLMs) achieved a best accuracy (0.9027), also results highlighted eight important content descriptions to have the highest influence on prediction.
Content may be subject to copyright.
Opción, Año 35, Especial No.19 (2019): 1368-1393
ISSN 1012-1587/ISSNe: 2477-9385
Recibido: 14-12-2019Aceptado: 19-03-2019
Prediction of the Digital Game Rating Systems
based on the ESRB
Khaled Mohammad Alomari1,
1Faculty of Computer Sciences, Abu Dhabi University, Abu Dhabi, UAE
khaled.alomari@adu.ac.aeb
Ahmad Qasim AlHamad1,
1Faculty of Computer Sciences, Abu Dhabi University, Abu Dhabi, UAE
aqd14@adu.ac.ae
Hisham o. Mbaidin2
1Faculty of Management Information Systems, Mutah University,
Karak, Jordan
Hisham.mbaidin@adu.ac.ae
Said Salloum3
3Fujairah University, Fujairah, UAE
ssalloum@uof.ac.ae
Abstract
This study tries to find out the best model for prediction video
game rate categories. A representation from four rating categories
(everyone, everyone 10+, teen, mature) was used for the analysis. The
paper follows CRISP-DM approach under Rapid Miner software to
business and data understanding, Data preparation, model building and
evaluation. The researchers compared prediction among six model and the
results showed that the Generalized Linear Models (GLMs) achieved a
best accuracy (0.9027), also results highlighted eight important content
descriptions to have the highest influence on prediction.
Keywords: Video game, Digital Game Rating Systems, Machine
Learning.
Predicción de los sistemas de clasificación de juegos
digitales basados en el sistema de calificación ESRB
Resumen
Este estudio trata de encontrar el mejor modelo para las categorías de
predicción de videojuegos. Para el análisis se utilizó una representación de
1369 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
cuatro categorías de calificación (todos, todos los mayores de 10 años,
adolescentes, adultos). El documento sigue el enfoque CRISP-DM del
software Rapid Miner para la comprensión de negocios y datos, preparación
de datos, construcción de modelos y evaluación. Los investigadores
compararon la predicción entre los seis modelos y los resultados mostraron
que los Modelos lineales generalizados (GLM) lograron una mejor precisión
(0.9027); los resultados también resaltaron ocho descripciones de contenido
importantes para tener la mayor influencia en la predicción.
Palabras clave: Videojuego, Sistemas de Clasificación de Juegos
Digitales, Aprendizaje Automático.
1. INTRODUCTION
In the wake emergence of the technological and communication
revolution in the last quarter of the twentieth century and after the
proliferation of “Globalization “, doubts and fears appeared among many
peoples especially the developing communities on the impact of the
modern technologies and media on cultures, traditions, heritage, and social
structure (S.A. Salloum, AlHamad, Al-Emran, & Shaalan, 2018; Said A
Salloum, Al-Emran, & Shaalan, 2017). Many studies have shown no
communities are fully immune on the impacts of modern new
technologies where it becomes an influential force on the global social
structure. Because of its fast-rapid spread in our community, it becomes a
source of information entertainment and cultural trait.
In the shadows of modern communities, information technology
plays a huge impact on the members of the community. People find
themselves automatically dealing with huge amount of information and
managing their daily complex and variedproblems through the usage of
this technology (Said A Salloum, Al-Emran, & Shaalan, 2018). This
technology provides them with the energy unprecedented thinking tool
Local wisdom based modelling strategy of SME business
development in Indonesia
1370
and the art of finding new modern and sophisticated solution (S. A.
Salloum, Al-Emran, Monem, & Shaalan, 2017). In addition, technology
gave the opportunity to the community to learn more and absorb new
concepts to stay up-to-date with the current world, while children being
sensitive and more adaptive to these changes where in many cases they
become in competition between parents and educators in the socialization
and education.
Perhaps the most important achievements of information
technology are the advent of computers, Internet, mobile devices, games
consoles, and tablet, which reshaped the child's life at home and school in
unexpected and profound ways. Children of the „e-community‟ are more
susceptible to the pros and cons of this community. Computers push
children to learn better, through the learning they found more efficient
environment and while trying new technologies that makes them more
prepared for the future. For children, learning information technology is
essential for giving them tools of success at the future because in the
modern world, technology plays a major role of knowledge provider
without retreating to other fields.
2. BACKGROUND
2.1. Digital Game Effect
According to (Turner et al., 2012) as researchers, it is important to
separate the hype from the reality. A fundamental issue for the study of
problems related to digital gaming is that there is no agreement upon
1371 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
definition of types of problems digital games produce and also, there is no
clear boundaries between normal play and excessive play. The digital
games today are spur to players to become part of it. It plays a big role in
physical and emotional effect on the player. It took them to a deeper level
in engagement and interaction (Norcia, 2014). According to ,McGonigal,
2011, American children spend long hours in playing video games
compared to the time spent on their studying. Where in many casesin other
places, the time spent on video games vs. studying were equal. This would
result in a fact that by the time this American child gets to the age of 21 he
or she would have spent 10000 hours on playing video games. Other
statistics (latest statistics) showed that 170 million American are video
gamers. According to , Greitemeyer & Osswald, 2009, pro-social video
game playing led to a short-term reduction in the tendency to see the
world as hostile and an immediate reduction in anti-social thoughts. The
pro-social game would lead to increased helpful and decreased hurtful
behavior, relative to violent games, with neutral games yielding
intermediate behaviors (Saleem, Anderson, & Gentile, 2012). Video
games can also include opportunities for self-assessment and are often
becoming important social learning environments that allow for additional
learning from different perspectives (McLean & Griffiths, 2013). Some
researchers have argued that video games are the “training wheels” for
computer literacy (Gentile & Anderson, 2006); (McLean & Griffiths,
2013). Video game helps to improve perceptual skills and visual attention
task regardless of previous video game experience. Other studies have
documented relations between video game play and visual selective
attention, mental rotation, spatial visualization, and reaction times (Gentile
& Anderson, 2006). Video games can also provide opportunities for
practice in different directions and in the use of fine motor skills (Gentile
Local wisdom based modelling strategy of SME business
development in Indonesia
1372
& Anderson, 2006). In spite of the benefits that may be in some video
games, however, drawbacks in the horizon arise, because most of the
games used by children and adolescents with negative implications
affecting them at all stages of growth. In addition, a large percentage of
video games depends on the entertainment and enjoy killing others,
destroying property and assaulting them unjustly. Consequently, children
and adolescents methods of committing the crime and their arts, tricks,
capabilities and skills of violence, aggression, and the outcome of crime
developed in their minds. These capabilities acquired through the exercise
of intimacy in those games (Weiss, 2010). Excess synapses during early
adolescence in emotionally laden situations may lead to the increased
aggression of early adolescence (more so for boys than girls) (Kirsh,
2002).
Researchers have disagreed in regard to whether digital games such
as Mortal Kombat, Splatterhouse, and Grand Theft Auto have the capacity
to „blur‟ the social boundaries between good and evil and to invoke
catastrophic effects through subsequent violent acts by gamers (Jaslow,
2013). According to ,Engelhardt, Bartholow, Kerr, & Bushman, 2011, it is
confirmed that violent digital games with apocalyptic themes desensitize
the gamer to the violent graphics over the short term; and that games such
as Grand Theft Auto, Killzone and Hitman caused the participants to act
more aggressively than the participants who played non-violent games.
The study in which 70 participants played either violent or non-violent
digital games for 25 minutes, and after which measurements were made
on the participants‟ brain activities which supported that the violent
content and interaction, enhanced the neural desensitization to violence.
According to (Engelhardt et al., 2011) the neural desensitization also
1373 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
served as a predictor to increased aggressive behaviors after repeated
exposure to the game violence. However, Massachusetts Institute of
Technology (MIT) Director, Henry Jenkins (Jenkins, 2006) argued that
games are “merely games” and that the youth are well aware of the
difference between game rules and rules that apply in the real world.
Further, (Jenkins, 2006) supported that the youth who transfers the tragedy
in digital game content to real-world tragedies are “severely emotionally
disturbed”.
3. LITERATURE REVIEW
3.1.Prediction Models
Reliable outputs from predictive analyses of big data are of
tremendous value to business analysts forecasting tools. Statistical models
are typically used to build relationships and to test them in order to defend
theories of causal relationships that are used as a basis for prediction.
(Hee, 1966) outlined the importance of both qualitative and quantitative
testing of the validity of forecasting model formulas and values. (Vucevic
& Yaddow, 2012) supported that game design approaches that are
unorganized and that are not clearly defined prevent the prediction and
repetition of the quality of the product.
The need to test the validity of theoretical scientific proclamations
has been well served by predictive modeling and testing methods which
identify existing causal mechanisms and discover differentiations in
construction operations (Shmueli, 2010). The predictive model may reveal
Local wisdom based modelling strategy of SME business
development in Indonesia
1374
potential improvements to existing models and bridge the gap between
theoretical assumptions and application. Classical models for probability
and prediction were developed prior to the information technology
revolution and pre-programmed algorithms. Furthermore, the
quantification of predictability within the model is commonly achieved by
a statistical evaluation of distinct features through both explanatory,
descriptive, and predictive approaches (Kumalasari, 2019).
According to (Shmueli, 2010) the predictive model is considered a
tool for applications of data mining in order to make new and future
predictions. The approach to making predictions may be Bayesian,
frequentists, parametric, or statistical models. Statistical predictions within
dynamic systems fit the Bayesian model; when time and relative system
state are not referenced, the distribution selection is generated by
equilibrium distribution. Model-to-data fit may be measured by a prior and
posterior predictive checks or from a hybrid of checks designed for the
hierarchical model (Gelman, Hwang, & Vehtari, 2014). The most common
prediction model is the multiple linear prediction models which utilizes
predictor and response variables.
The primary objective of logistic regression is the modeling of
some probability. More specifically, predictive modeling is used to
forecast unknown values based on other values or attributes that are
known. The learning algorithm is used to generate a dependable rule for
the prediction of probable outputs for future data (Friedman, 1997). In the
case of non-stochastic prediction, the objective is the prediction of output
(Y) for any new observations with input values (X), and observations to
1375 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
time (t) is the basis for forecasting future values at a time t + k, where
k > 0 (Geisser, 1993).
4. METHODOLOGY
The methodology for this study is based on the assumptions and
structures of the CRISP-DM to prediction models for digital game
rating. The initial data is collected from large datasets of ESRB rating
system; and is prepared, modeled and evaluated based upon the
CRISP-DM general tasks. The video game data is extracted and
segmented are created in order to create the game rating prediction
model.
The first step of the CRISP-DM model is to identify and clarify
the business objective. The business objective for this study is to gain
an understanding of video game rating and prediction models
performance through the extraction of data from a large dataset for
analysis. This section will describe sample data and outline the method
of data analysis for 2053 video game titles rating that ware extracted
from a listing ESRB ratings system in order to explore different
prediction models that will identify the most significant model to
predict a rating for video game titles.
Local wisdom based modelling strategy of SME business
development in Indonesia
1376
4.1 .Date
The research will use a total of 2053 video game titles which were
selected from the listing of ESRB ratings. The method of selection was to
extract all video game under the PlayStation four platform until June
2018. The final 2053 title sample is comprised of representations from
four rating categories (Everyone, Everyone 10+, Teen, Mature). A total of
2053 video game titles were used for the analysis along with the 38 titles
of Content Descriptors as shown Table 1 and Classification Rate
generated for each game. The data was stored in Excel Sheet where
content on (Game Title, Rate, 38 Content Descriptors). The Content
Descriptors convert to binary data through using 1 if exist content
descriptors and 0 if not exist according to the game data in the Content
Descriptors based on ESRB data.
Table 1 Content Descriptors
Alcohol
Reference
Animated
Blood
Blood
Blood and
Gore
Comic
Mischief
Crude
Humor
Drug
Referenc
e
Fantasy
Violence
Language
Lyrics
Mature
Humor
Mild Blood
Mild Fantasy
Violence
Mild
Language
Mild
Lyrics
Mild
Suggestive
Themes
Nudity
Partial
Nudity
Real
Gamblin
g
Sexual Content
Sexual
Violence
Simulated
Gambling
Strong
Languag
e
Strong Lyrics
Suggestive
Tobacco
Use of
Use of Alcohol
1377 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
Themes
Reference
Alcohol
and Tobacco
Use of Drugs
and Alcohol
Use of
Tobacco
Violence
4.2.Transformation
The data was scraped from the ESRB, cleaned and trained based
upon the steps of the CRISP-DM model, using the RapidMiner tool. The
data which was stored on excel files uploaded as sources which were used
to create a dataset that could be selected for the different types of
analytics. The dataset was split randomly to 80% train and 20% validation
set as shown in Table 2. Where validation set outputs are then used to
predict game rate classification for 411 game titles.
Table 2 Dataset split
Rate
Everyone
Everyone
10+
Teen
Mature
Total
Train (0.8)
409
366
543
324
1642
validation
(0.2)
102
92
136
81
411
Total
511
458
679
405
2053
i. Generalized Linear Models (GLMs)
Executes GLM algorithm using H2O 3.8.2.6.
Local wisdom based modelling strategy of SME business
development in Indonesia
1378
Generalized linear models (GLMs) are an extension of traditional
linear models that allows the specification of models whose response
variable follows different distributions. It covers widely used statistical
models, such as linear regression for normally distributed responses,
logistic models for binary data, log linear models for count data,
complementary log-log models for interval-censored survival data, plus
many other statistical models through its very general model formulation.
The operator starts a 1-node local H2O cluster and runs the
algorithm on it.
Table 3 Generalized Linear Model result
Generalized
Linear
true
E
true
T
true
M
true
E10
FN
pred. E
101
1
0
5
6
pred. T
0
124
14
8
22
pred. M
0
1
67
0
1
pred. E10
1
10
0
79
11
ground
truth
102
136
81
92
Predict
107
146
68
90
FP
1
12
14
13
sum
411
TP & TN
371
Table 4 illustrate the Accuracy, Precision, Recall, and F-score of
each class in the Generalized Linear model.
1379 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
Table 4 Generalized Linear Model categories result
Generalized Linear
class
E
T
M
E10
TP
101
124
67
79
TN
270
247
304
292
FP
1
12
14
13
FN
6
22
1
11
Accuracy
0.981
0.916
0.961
0.939
Precision
0.99
0.912
0.827
0.859
Recall
0.944
0.849
0.985
0.878
F-score
0.967
0.879
0.899
0.868
ii. Decision Tree
Decision tree employs the divide and conquer method and
recursively divides a training set until each division consists of examples
from one class. A general algorithm for decision tree building will include
creating a root node and assign all of the training data to it, selecting the
best splitting attribute, adding a branch to the root node for each value of
the split, splitting the data into mutually exclusive subsets along the lines
of the specific split and at the end a repetition is made for steps 2 and 3 for
each and every leaf node until the stopping criteria is reached. A
prediction for the class label Attribute is determined depending on the
majority of Examples which reached this leaf during generation, while an
estimation for a numerical value is obtained by averaging the values in a
leaf.
This Operator can process Example Sets containing both nominal
and numerical Attributes. The label Attribute must be nominal for
classification and numerical for regression.
Local wisdom based modelling strategy of SME business
development in Indonesia
1380
Table 5 illustrate the prediction using the Decision Tree Model and
collocation the False Negative (FN), False Positive (FP), True Negative
(TN) and True Positive (TP) to evaluation model.
Table 5 Decision Tree Model result
Decision
Tree
true
E
true
T
true
M
true
E10
FN
pred. E
101
1
0
6
7
pred. T
0
126
14
11
25
pred. M
0
2
67
0
2
pred. E10
1
7
0
75
8
Ground
Truth
102
136
81
92
Predict
106
149
67
89
FP
1
10
14
17
sum
411
TP & TN
369
Table 6 illustrate the Accuracy, Precision, Recall, and F-score of
each class in the Decision Tree model.
Table 6 Decision Tree Model categories result
Decision Tree
Class
E
T
M
E10
TP
101
126
67
75
TN
268
243
302
294
FP
1
10
14
17
FN
7
25
2
8
Accuracy
0.97878
0.913366
0.958442
0.936548
Precision
0.990196
0.926471
0.82716
0.815217
Recall
0.935185
0.834437
0.971014
0.903614
F-score
0.961905
0.878049
0.893333
0.857143
1381 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
iii. Deep Learning
Deep Learning is a machine learning technique that imitate the
human experience so that the computer will appear as normal human. In
deep learning, a computer model learns to perform classification tasks
directly from images, text, or sound. In deep learning, models are trained
with stochastic gradient descent using back-propagation. The network can
contain a large number of hidden labeled layers consisting of neurons with
tech, rectifier, and max out activation functions. It also requires very large
computing power.
Table 7 illustrate the prediction using the Deep Learning Model
and collocation the False Negative (FN), False Positive (FP), True
Negative (TN) and True Positive (TP) to evaluation model.
Table 7 Deep Learning Model result
Deep Learning
true E
true T
true M
true E10
FN
pred. E
99
1
1
5
7
pred. T
0
126
14
9
23
pred. M
0
1
66
0
1
pred. E10
3
8
0
78
11
ground truth
102
136
81
92
Predict
108
151
69
83
FP
3
10
15
14
sum
411
TP & TN
369
Local wisdom based modelling strategy of SME business
development in Indonesia
1382
Table 8 illustrate the Accuracy, Precision, Recall, and F-score of
each class in the Deep Learning model.
Table 8 Deep Learning Model categories result
Deep
Learning
Class
E
T
M
E10
TP
99
126
66
78
TN
270
243
303
291
FP
3
10
15
14
FN
7
23
1
11
Accurac
y
0.97361
5
0.91791
0.95844
2
0.93654
8
Precision
0.97058
8
0.92647
1
0.81481
5
0.84782
6
Recall
0.93396
2
0.84563
8
0.98507
5
0.87640
4
F-score
0.95192
3
0.88421
1
0.89189
2
0.86187
8
iv. Gradient Boosted
A gradient boosted model is an ensemble of either regression or
classification tree models. Both are forward-learning ensemble methods
that obtain predictive results through gradually improved estimations.
Boosting is a flexible nonlinear regression procedure that helps to improve
the accuracy of trees. By sequentially applying weak classification
algorithms to the incrementally changed data, a series of decision trees are
created that produce an ensemble of weak prediction models. While
boosting trees increases their accuracy, it also decreases speed and human
interpretability. The gradient boosting method generalizes tree boosting to
minimize these issues.
1383 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
The operator starts a 1-node local H2O cluster and runs the
algorithm on it. Although it uses one node, the execution in parallel.
Table 9 illustrate the prediction using the Gradient Boosted Model
and collocation the False Negative (FN), False Positive (FP), True
Negative (TN) and True Positive (TP) to evaluation model.
Table 9 Gradient Boosted Model result
Gradient
Boosted
true
E
true
T
true
M
true
E10
F
N
pred. E
100
2
1
5
8
pred. T
0
119
13
7
20
pred. M
0
0
67
0
0
pred. E10
2
15
0
80
17
ground
truth
102
136
81
92
Predict
108
139
67
97
FP
2
17
14
12
sum
411
TP & TN
366
Table 10 illustrate the Accuracy, Precision, Recall, and F-score of
each class in the Gradient Boosted model.
Local wisdom based modelling strategy of SME business
development in Indonesia
1384
Table 10 Gradient Boosted Model categories result
Gradient
Boosted
Class
E
T
M
E10
TP
100
119
67
80
TN
266
247
299
286
FP
2
17
14
12
FN
8
20
0
17
Accurac
y
0.97340
4
0.90818
9
0.96315
8
0.92658
2
Precisio
n
0.98039
2
0.875
0.82716
0.86956
5
Recall
0.92592
6
0.85611
5
1
0.82474
2
F-score
0.95238
1
0.86545
5
0.90540
5
0.84656
1
v. Random Forest
A collection of several random trees that are defined by the amount
of trees parameter form up a random forest. The creation/training of the
trees takes place at the Example Set‟s bootstrapped sub-sets available at
the Input Port. The splitting rule for a particular Attribute is signified for
every node present in the tree. The splitting rule is chosen on the basis of a
sub-set of Attributes, showed along with subset ratio. The values are
separated in best possible way through this rule for the chosen parameter
criterion. The values are separated according to the rule for various classes
in classification, whereas, when regression is concerned, the errors due to
the estimation are minimized by separating. Till the ending criteria are
reached, new nodes are built continuously.
1385 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
The model‟s complexity can be minimized by leverage of the
pruning technique after replacing sub-trees, as its predictive power with
leaves is quite limited. The details of parameter shall indicate the various
kinds of pruning.
A somewhat identical technique to random forest is extremely
randomized trees; the method for acquiring this includes assessing the
split random parameter and disabling pruning. Tuning for this technique
include the parameters that are minimal leaf size and split ratio, while
disabling guess split ratio will undo it. The best picks for the minimal leaf
size classification and regression problems include two and five
respectively.
Table 11 illustrate the prediction using the Random Forest Model
and collocation the False Negative (FN), False Positive (FP), True
Negative (TN) and True Positive (TP) to evaluation model.
Table 11 Random Forest Model result
Random Forest
true E
true T
true M
true E10
FN
pred. E
101
7
2
26
35
pred. T
0
115
16
1
17
pred. M
0
0
63
0
0
pred. E10
1
14
0
65
15
ground truth
102
136
81
92
Predict
136
132
63
80
FP
1
21
18
27
sum
411
TP & TN
344
Local wisdom based modelling strategy of SME business
development in Indonesia
1386
Table 12 illustrate the Accuracy, Precision, Recall, and F-score of
each class in the Random Forest model.
Table 12 Random Forest Model categories result
Random
Forest
Class
E
T
M
E10
TP
101
115
63
65
TN
243
229
281
279
FP
1
21
18
27
FN
35
17
0
15
Accuracy
0.905263
0.900524
0.950276
0.891192
Precision
0.990196
0.845588
0.777778
0.706522
Recall
0.742647
0.871212
1
0.8125
F-score
0.848739
0.858209
0.875
0.755814
vi. Naive Bayes
A small data set can be the basis for creating a reasonable model
through Naive Bayes which is a high-bias, low-variance classifier. It‟s
quite user-friendly and computationally inexpensive. The major
utilizations are for text categorization, including spam detection, sentiment
analysis, and recommender systems. The Naive Bayes has a fundamental
assumption on the theory that if the label (the class) is provided, the result
from an Attribute is independent from the result of the other Attributes.
The Naive Bayes classifier is quite useful, despite of the fact that these
assumptions are rarely true (they are "naive"!). The mathematical work
required for creating the Naive Bayes probability model is minimized with
the independence assumption. Certain assumption regarding distributions
of conditional probability should be carried out to give the final touch to
the probability model, for specific Attributes according to the class. The
1387 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
Gaussian probability densities are used by Operator for modeling the
Attribute data.
Table 13 illustrate the prediction using the Naive Bayes Model and
collocation the False Negative (FN), False Positive (FP), True Negative
(TN) and True Positive (TP) to evaluation model.
Table 13 Naïve Bayes Model result
Naive Bayes
true E
true T
true M
true E10
FN
pred. E
100
3
1
9
13
pred. T
0
58
0
0
0
pred. M
0
58
79
0
58
pred. E10
2
17
1
83
20
ground truth
102
136
81
92
Predict
113
58
137
103
FP
2
78
2
9
sum
411
TP & TN
320
Table 14 illustrate the Accuracy, Precision, Recall, and F-score of
each class in the Naive Bayes model.
Table 14 Naie Bayes Model categories result
Naive Bayes
Class
E
T
M
E10
TP
100
58
79
83
TN
220
262
241
237
FP
2
78
2
9
FN
13
0
58
20
Local wisdom based modelling strategy of SME business
development in Indonesia
1388
Accuracy
0.955224
0.80402
0.842105
0.916905
Precision
0.980392
0.426471
0.975309
0.902174
Recall
0.884956
1
0.576642
0.805825
F-score
0.930233
0.597938
0.724771
0.851282
5. DISCUSSION
For model‟s analysis, the authors used two methods. First, through
evaluating each class in the model used accuracy and F-Score. Second,
evaluating the models used accuracy.
Table 15 compares each class in all models
Class E
Class T
Class M
Class E10
Accura
cy
F-
scor
e
Accura
cy
F-
scor
e
Accura
cy
F-
scor
e
Accura
cy
F-
scor
e
Generalized
Linear
0.981
0.96
7
0.916
0.87
9
0.961
0.89
9
0.939
0.86
8
Decision Tree
0.979
0.97
1
0.913
0.88
4
0.958
0.90
5
0.937
0.82
9
Deep Learning
0.974
0.94
3
0.918
0.87
8
0.958
0.88
0.937
0.89
1
Gradient
Boosted
0.973
0.95
2
0.908
0.86
5
0.963
0.90
5
0.927
0.84
7
Random
Forest
0.905
0.84
9
0.901
0.85
8
0.95
0.87
5
0.891
0.75
6
Naive Bayes
0.955
0.93
0.804
0.59
8
0.842
0.72
5
0.917
0.85
1
From table 15 above that illustrate the comparing Accuracy and F-
score of class in each model and if we exclude the random forest model, it
is obvious that the accuracy results in class E was very high. It is
important to note that any game in principal is for everyone (means, class
1389 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
E), the content description will change the class to be some other classes
please see table 17 for content descriptions. In general, the accuracy
results were in somehow high in all categories, but not in the class T
category, which implies that the prediction for this class was not achieved
perfectly. The researchers believe that this is due to the conflict between
class T content description with (class E10 and class M) content
descriptions.
Table 16 comparing the models evaluation
Model
Overall Accuracy
Generalized Linear
0.9027
Decision Tree
0.8978
Deep Learning
0.8978
Gradient Boosted
0.8905
Random Forest
0.837
Naive Bayes
0.7786
Table 16 shows that the generalized linear model achieved the best
results in terms of accuracy (0.9027) compared to other models used in
this research. Although, the decision tree and deep learning models
achieved very well. The most important content description that has high
influence on prediction were those who has weight more than 0.5 shown
in table 17 and proved by the receiver operating characteristic curve
(ROC) shown in figure 2 below.
Local wisdom based modelling strategy of SME business
development in Indonesia
1390
Figure1 content description weight
Table 17 content description weight result
Attribute
weights
Attribute
weights
Violence
1
Cartoon Violence
0.188934216
Strong Language
0.838057824
Use of Alcohol and
Tobacco
0.168348965
Fantasy Violence
0.794854985
Mild Cartoon Violence
0.159409298
Blood
0.756097962
Strong Sexual Content
0.153716103
Sexual Themes
0.696178674
Use of Drugs and
Alcohol
0.153716103
Blood and Gore
0.681380689
Mild Violence
0.133808617
Intense Violence
0.662858896
Comic Mischief
0.132589223
Mild Fantasy
Violence
0.598541607
Alcohol Reference
0.124247704
Language
0.388589151
Use of Tobacco
0.10678485
Suggestive Themes
0.384235154
Animated Blood
0.082399214
Mild Language
0.33754685
Mature Humor
0.063136969
Drug Reference
0.284278128
Lyrics
0.048609269
Crude Humor
0.24096595
Simulated Gambling
0.017131237
Sexual Content
0.236182457
Tobacco Reference
0.013750678
Use of Drugs
0.230909405
Sexual Violence
0
1391 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
Use of Alcohol
0.220620196
Strong Lyrics
0
Mild Suggestive
Themes
0.204935285
6. CONCLUSIONS
The researchers compared prediction among six model and the
results showed that the Generalized Linear Models (GLMs) achieved a
best accuracy (0.9027), also results highlighted eight important content
descriptions to have the highest influence on prediction. The most
important content description that has high influence on prediction were
those who has weight more than 0.5 shown in table 17 which are:
Violence, Strong Language, Fantasy Violence, Blood, Sexual Themes,
Blood and Gore, Intense Violence, and Mild Fantasy Violence. These
content description results are proved by the receiver operating
characteristic curve (ROC).
The result is useful for researchers and digital rating system
developer so that it formulates a base for parents to be advised before their
children can use these games. Which could also be very beneficial in
predicting their behaviors if these games are to be used, so the best is to
have restriction in these games.
REFERENCES
Bartlmae, K. ,2000,. An experience factory for knowledge discovery in
databases: representing KDD-experience. In Proceedings of the
2000 information resources management association international
Local wisdom based modelling strategy of SME business
development in Indonesia
1392
conference on Challenges of information technology management
in the 21st century (pp. 10641065). IGI Global.
Diebold, F. X., & Mariano, R. S. ,1995,. Comparing Predictive
Accuracy. Journal of Business & Economic Statistics,
Vol.13,No.3,pp. 253263.
Diebold, F. X., & Kilian, L. ,2001,. Measuring predictability: theory
and macroeconomic applications. Journal of Applied
Econometrics, Vol.6,No.6, pp. 657669.
Engelhardt, C. R., Bartholow, B. D., Kerr, G. T., & Bushman, B. J. ,
2011,. This is your brain on violent video games: Neural
desensitization to violence predicts increased aggression
following violent video game exposure. Journal of Experimental
Social Psychology, Vol.47,No.5, pp. 10331036.
Felini, D. , 2015, Beyond today’s video game rating systems: A critical
approach to PEGI and ESRB, and proposed improvements.
Games and Culture, Vol.10. No.1,pp. 106122.
Geisser, S. ,1993, Predictive inference, vol. 55. CRC Press, Boca
Raton, FL. Gullett, W.(1997). Environmental Protection and the
Precautionary Principle: A Response to Scientific Uncertainty in
Environmental Management. Environmental and Planning Law
Journal, Vol.14,No.1, pp. 5269.
Gelman, A., Hwang, J., & Vehtari, A. ,2014,. Understanding predictive
information criteria for Bayesian models. Statistics and
Computing, Vol.24,No.6, pp. 9971016.
Gentile, D. A., & Anderson, C. A. ,2006,. Video games. Encyclopedia of
Human Development, Vol.3, No.8, pp.13031307.
Greitemeyer, T., & Osswald, S, 2009, Prosocial video games reduce
aggressive cognitions. Journal of Experimental Social Psychology,
Vol.45,No.4,pp.896900.
Gruber, M. , 2017,. Decision Modeling and CRISP-DM for Modern
Data Science Projects. Decision Management Solutions. Retrieved
from Http://Www.Decisionmanagementsolutions.Com/Crisp-Dm-
and-Decision-Modeling-for-Modern-Data-Science-Projects/.
Hee, O. ,1966, Tests for Predictability of Statistical Models. Journal of
Farm Economics, Vol.48, No.5, pp.14791484.
Jaslow, R. ,2013,. Violent video games and mass violence: a complex
1393 Iha Haryani Hatta et al.
Opción, Año 35, Especial No.19 (2019): 1368-1393
link. CBS News, 18.
Jenkins, H. , 2006, Reality bytes: Eight myths about video games
debunked. Impact of Gaming Essays. Retrieved Feb, 1, 2007.
Kirsh, S. J. ,2002, The effects of violent video games on adolescents.
Aggression and Violent Behavior, Vol.7, pp.113.
Kumalasari, G. 2019. Local government policy model of indonesia
rattan handicraft creative industry in trangsan village.
Humanities & Social Sciences Reviews. Vol. 7, No 3: 87-91. India.
Laczniak, R. N., Carlson, L., Walker, D., & Brocato, E. D., 2017,
Parental Restrictive Mediation and Children’s Violent Video
Game Play: The Effectiveness of the Entertainment Software
Rating Board (ESRB) Rating System. Journal of Public Policy &
Marketing, Vol.36,No.1, pp.7078.
McGonigal, J, 2011,. Reality is broken: Why games make us better
and how they can change the world. Penguin.
McLean, L., & Griffiths, M. D. ,2013, The psychological effects of video
games on young people. Aloma: Revista de Psicologia, Ciències de
l‟Educació i de l‟Esport, Vol.31,No.1.
Norcia, A. M. A., 2014,. The Impact of Video Games. Retrieved from
Http://Www.Pamf.Org/Parenting-Teens/General/Media-
Web/Videogames.Html.
Saleem, M., Anderson, C. A., & Gentile, D. A., 2012,. Effects of
prosocial, neutral, and violent video games on children’s helpful
and hurtful behaviors. Aggressive Behavior, Vol.38,No.4,pp.
281287.
Salloum, S. A., Al-Emran, M., Monem, A., & Shaalan, K,2017, A Survey
of Text Mining in Social Media: Facebook and Twitter
Perspectives. Advances in Science, Technology and Engineering
Systems Journal, Vol.2,No.1, pp. 127133.
UNIVERSIDAD
DEL ZULIA
Revista de Ciencias Humanas y Sociales
Año 35, Especial N° 19, 2019
Esta revista fue editada en formato digital por el personal de la Oficina de
Pubñlicaciones Científicas de la Facultad Experimental de Ciencias,
Universidad del Zulia.
Maracaibo - Venezuela
www.luz.edu.ve
www.serbi.luz.edu.ve
produccioncientifica.luz.edu.ve
... To measure the discriminant validity, two criteria require measurement which is the Fornell-Larker criterion, and the Heterotrait-Monotrait ratio (HTMT) [45]. The Fornell-Larker condition, affirmed the prerequisites as shown in the results of Table 2 as all the AVEs alongside with their square roots are larger than their relationship with other structures from the source [54]. Table 3 given below has the HRMT ratio showing the threshold value to be higher than the values of individual constructs [55]. ...
... was applied to check the predictive model. Different classifiers such as J48, Logistic, OneR, BayesNet, LWL, and AdaBoostM1 were introduced for testing the predictive model [53,54]. The outcomes seen in Table 6 conclude that the J48 had much better performance compared to many other classifiers in terms of forecasting speed & vividness (SV), perceived concentration (PC), perceived ease of use (PE), and perceived usefulness (PU) on the acceptance of audio-video material (AA). ...
... Also, it is attracting and influencing the advancement of humans and their society. Deep Learning, especially LSTM, is used on most of the text generation tasks in recent years [1][2][3][4][5]. We have found that poem generation uses different techniques and consists of many steps of word embedding. ...
... Since we have a vocabulary group, each word has a word_id that identifies it and is used for one-hot encoding. The words' sequence presents a two-dimensional array of length fixed to the experiment sequence required (1,2,5,10) and width of vocabulary count (which was 10000). On one hot encoding, one field is only set to 1, and the rest are zeros, so the Dataset is divided into several sequences with the same length, then converted to a one-hot encoding format then passed to the model for training. ...
Chapter
Full-text available
Text Generation, especially poetry synthesis, is a promising and challenging AI task. We have used LSTM and word2vec methods to explore this area. We do forward and backward word training with different word sequences lengths. Two Datasets of Arabic poems were used. Preprocessing for unification and cleaning was done too, but the Data size was big and required very high memory and processing, so we used a sub-Datasets for training; this affected our experiments since the model is trained on fewer data. A user-supplied keyword was implemented. We have found the shorter training sequence models were better in generating more meaningful text, and longer models prefer most frequent words, repeat text, and use small words. Best predicted sentences were selected by measuring each of its words conditional probability and multiply them; this avoids local maxima if we used a greedy method that chooses the best next-word only. Moreover, the AraVecword2vec module was not very helpful since it was provided synonyms much more that related words. Many enhancements can be done in the future, such as Arabic prosody constraints, and overcome the hardware issue.
... The IT industry has gone through tremendous amount of change, growth, restructuring and advancements in recent years and it doesn't stop here [16][17][18][19][20][21]. Every year new software machineries, tools like cross-platform application development [16,[22][23][24][25][26][27], Blockchain development [28], machine learning [29][30][31][32], cloud computing etc. is getting improved and enhanced, along with their progression demands of their utilizations and consumptions are increasing worldwide constantly. Before we go into detail discussion on all the latest trends evolving around today's health care industry, we should highlight a critical note that for any organization or industry to adopt [33,34], consume, harness innovative steps and the latest trends, it should maintain operational integrity by modernizing legacy core systems, value for money and transforming the business technology while closing complying with or I should say zero compromise on the evolving cyber security and privacy risks [35]. ...
... was utilized to test the predictive model. Different classifiers, such as J48, Logistic, OneR, BayesNet, LWL, and AdaBoostM1, were used to test the predictive model [102][103][104][105][106]. ...
Article
Full-text available
This study makes use of a cohesive yet innovative research model to identify the determinants of the adoption of smart watches using constructs from the Technology Acceptance Model (TAM) and constructs of smartwatches, including effectiveness, content richness, and personal innovativeness. The chief objective of the study was to encourage the use of smartwatches for medical purposes so that the role of doctors can be made more effective and to facilitate access to patient records. Our conceptual framework highlights the association of TAM constructs (i.e., perceived usefulness and perceived ease of use) with the content richness, the construct of user satisfaction, and innovativeness. To measure the effectiveness of the smartwatch, an external factor based on the flow theory was added, which emphasizes the control over the smartwatch and the degree of involvement. The study employs data from 385 respondents involved in the field of medicine, such as doctors, patients, and nurses. The data were gathered through a survey and used for evaluation of the research model using partial least squares structural equation modeling (PLS-SEM) and machine learning (ML) models. The significance and performance of factors impacting THE adoption of smartwatches were also identified using Importance-Performance Map Analysis (IPMA). User satisfaction is the most important predictor of intention to adopt a medical smartwatch according to the ML and IPMA analyses. The fitting of the structural equation model to the sample showed a high dependence of user satisfaction on perceived usefulness and perceived ease of use. Furthermore, two critical factors, in-novativeness and content richness, are demonstrated to enhance perceived usefulness. However, one should consider that perceived usefulness or behavioral intention could not be determined based on perceived ease of use. In general, the findings suggest that smartwatch usage could become critically important in the medical field as a mediator that allows doctors, patients, and other users to access essential information.
... The use of DM for educational activities has been increased based on staff decisions, student performance, and administration decisions [23]. Data-driven knowledge discovery is a paradigm that can be applied to DM [24][25][26]. DM is a multifaceted area that encompasses a variety of topics such as statistics, AI, information technology, learning, data visualization, and retrieval [27]. ...
Article
Nowadays, the modern management is promoted to resolve the issue of unreliable information transmission and to provide work efficiency. The basic aim of the modern management is to be more effective in the role of the school to train talents and serve the society. This article focuses on the application of data mining (DM) in the development of information management system (IMS) in universities and colleges. DM provides powerful approaches for a variety of educational areas. Due to the large amount of student information that can be used to design valuable patterns relevant to student learning behavior, research in the field of education is continuously expanding. Educational data mining can be used by educational institutions to assess student performance, assisting the institution in recognizing the student’s accomplishments. In DM, classification is a well-known technique that has been regularly used to determine student achievement. In this study, the process of DM and the application research of association rules is introduced in the development of IMS in universities and colleges. The results show that the curriculum covers the whole field and the minimum transaction support count be 2, minconf = 70%. The results also suggested that students who choose one course also tend to choose the other course. The application of DM theory in university information will greatly upsurge the data analysis capability of administrators and improve the management level.
... Although previous studies have demonstrated the promise of the XAI technique regarding interpretability analysis, no study has used it to examine video games. Current research studies attempt to obtain metrics with the highest prediction accuracy (Alomari et al. 2019) but lack a thorough analysis of the harmful content in video games. The fact that no game studies focus on explainability has also been raised in previous review studies (Barredo Arrieta et al. 2020;Tjoa and Guan 2020). ...
Article
Full-text available
Studies reported that playing video games with harmful content can lead to adverse effects on players. Therefore, understanding the harmful content can help reduce these adverse effects. This study is the first to examine the potential of interpretable machine learning (ML) models for explaining the harmful content in video games that may potentially cause adverse effects on players based on game rating predictions. First, the study presents a performance analysis of the supervised ML models for game rating predictions. Secondly, using an interpretability analysis, this study explains the potentially harmful content. The results show that the ensemble Random Forest model robustly predicted game ratings. Then, the interpretable ML model successfully exposed and explained several harmful contents, including Blood, Fantasy Violence, Strong Language, and Blood and Gore. This revealed that the depiction of blood, the depiction of the mutilation of body parts, violent actions of human or non-human characters, and the frequent use of profanity might potentially be associated with adverse effects on players. The findings suggest the strength of interpretable ML models in explaining harmful content. The knowledge gained can be used to develop effective regulations for controlling identified video game content and potential adverse effects.
... In 2012 NN has been discussed thoroughly in many other deep learning capabilities, e.g., Long Short Term Memory (LSTM), Deep-forward Neural Networks (DNN), and Convolutional Neural Networks (CNN) [1][2][3][4][5][6]. Many solutions exist to produce music based on using computers, e.g., neural networks, Markov chains Recombinancy, sampling, morphing, and rule-based methodologies [2,3,7,8]. ...
Chapter
Full-text available
This paper discusses applying different types of neural networks to classify a dataset of type audio. We used a GTZAN dataset that includes various audio music records representing different conventional categories of music gen-res. Each shares a set of common traditions; these traditions we call features. We build our proposed Python models using the Anaconda toolkit with TensorFlow (TF) an open-source deep-learning library. In our previous research, we build a multilayer sequential model to classify the dataset and then solve the overfitting issue in that model. In this paper, we build a Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) with Long Short Time Memory (LSTM). Finally, we compared the results to know the capabilities and limitations of Deep Learning (DL). CNN outperformed the other models in terms of training and test accuracy, having 83.74% and 74%, respectively.
... Neural Networks (NN) have achieved a wide consideration as algorithms in artificial intelligence, where it first been addressed in 1957 as a perception algorithm. After that, in 1986, it was marked in the algorithm of backpropagation [1][2][3][4][5]. Finally, in recent years it successfully progressed in Deep Learning (DL) in many topics such as image classification and speech recognition. ...
Chapter
Full-text available
In this paper, we discuss applying a neural network model to classify a dataset of type audio. We used a GTZAN dataset that includes different audio music records representing different conventional categories of music genres. Each shares a set of common traditions; these traditions we call it features. We build our proposed Python models using the Anaconda toolkit that has TensorFlow (TF), an open-source deep-learning library. We build a multilayer sequential model to classify the dataset and then try to solve the model's overfitting issue. Results show that our proposed overfitting solution tends to increase Testing Accuracy equals 60.55% and reduce the Testing Error by 1.1726.
... Exploring this data plays a vital role in consumer retention by improving the quality of services [6,7]. Social media sites such as Instagram, Facebook, and Twitter offer valuable data that can be used by business owners not only to track and analyse customers' opinions about their businesses but also that of their competitors [8][9][10][11]. Moreover, these valuable data attracted decision-makers who seek to improve the services provided [8,9,12,13]. ...
Article
Full-text available
The growing popularity of social media sites has generated a massive amount of data that attracted researchers, decision-makers, and companies to investigate people's opinions and thoughts in various fields. Sentiment analysis is considered an emerging topic recently. Decision-makers, companies, and service providers as well-considered sentiment analysis as a valuable tool for improvement. This research paper aims to obtain a dataset of tweets and apply different machine learning algorithms to analyze and classify texts. This research paper explored text classification accuracy while using different classifiers for classifying balanced and unbalanced datasets. It was found that the performance of different classifiers varied depending on the size of the dataset. The results also revealed that the Naive Byes and ID3 gave a better accuracy level than other classifiers, and the performance was better with the balanced datasets. The different classifiers (K-NN, Decision Tree, Random Forest, and Random Tree) gave a better performance with the unbalanced datasets.
... In the twenty-first century, Artificial Intelligence has become one of the most advanced technologies employed in various sectors [1][2][3][4][5][6][7][8][9][10][11][12][13]. The United Arab Emirates was the first country to launch AI Strategy in the region and world; that shows the adoption of AI in the Federal government's strategic plans is inevitable [14][15][16]. ...
Chapter
The purpose of this chapter is to highlight the main technologies of Artificial Intelligence used in power system where the traditional methods will not be able to catch up all condition of operating and dispatching. Moreover, for each technology mentioned in the chapter there is a brief description where is used exactly power system. Moreover, these methods improve the operation and productivity of the power system by controlling voltage, stability, power-flow, and load frequency. It also permits to control the network such as location, size, and control of equipment and devices. The automation of the power system ensures to support the restoration, fault diagnosis, management, and network security. It is necessary to identify the appropriate AI technique to use it in planning, monitoring, and controlling the power system. Finally the chapter will highlight briefly sustainable side of using AI in power system.
Article
Full-text available
Text mining has become one of the trendy fields that has been incorporated in several research fields such as computational linguistics, Information Retrieval (IR) and data mining. Natural Language Processing (NLP) techniques were used to extract knowledge from the textual text that is written by human beings. Text mining reads an unstructured form of data to provide meaningful information patterns in a shortest time period. Social networking sites are a great source of communication as most of the people in today's world use these sites in their daily lives to keep connected to each other. It becomes a common practice to not write a sentence with correct grammar and spelling. This practice may lead to different kinds of ambiguities like lexical, syntactic, and semantic and due to this type of unclear data, it is hard to find out the actual data order. Accordingly, we are conducting an investigation with the aim of looking for different text mining methods to get various textual orders on social media websites. This survey aims to describe how studies in social media have used text analytics and text mining techniques for the purpose of identifying the key themes in the data. This survey focused on analyzing the text mining studies related to Facebook and Twitter; the two dominant social media in the world. Results of this survey can serve as the baselines for future text mining research.
Article
Full-text available
This study seeks to determine whether Entertainment Software Rating Board (ESRB) ratings are an effective tool for parents seeking to enhance their mediation efforts and ultimately reduce their children’s play of violent video games (and their children’s corresponding engagement in negative behaviors, such as getting in fights at school). Results demonstrate that children of parents who employ restrictive mediation efforts tend to play less violent video games, and the effect of these efforts is enhanced when parents use the ESRB system. This moderated effect also extends to reduce children’s engagement in negative behaviors in school. These findings suggest that both the video game industry and the government could play a role in establishing and maintaining educational programs whose purpose is to inform parents/caretakers about the potential usefulness of the ESRB guidelines.
Article
Full-text available
Recent research reveals that playing prosocial video games increases prosocial cognitions, positive affect, and helpful behaviors [Gentile et al., 2009; Greitemeyer and Osswald, 2009, 2010, 2011]. These results are consistent with the social-cognitive models of social behavior such as the general learning model [Buckley and Anderson, 2006]. However, no experimental studies have examined such effects on children. Previous research on violent video games suggests that short-term effects of video games are largely based on priming of existing behavioral scripts. Thus, it is unclear whether younger children will show similar effects. This research had 9–14 years olds play a prosocial, neutral, or violent video game, and assessed helpful and hurtful behaviors simultaneously through a new tangram measure. Prosocial games increased helpful and decreased hurtful behavior, whereas violent games had the opposite effects. Aggr. Behav. 38:281-287, 2012. © 2012 Wiley Periodicals, Inc.
Article
Full-text available
We review the Akaike, deviance, and Watanabe-Akaike information criteria from a Bayesian perspective, where the goal is to estimate expected out-of-sample-prediction error using a biascorrected adjustment of within-sample error. We focus on the choices involved in setting up these measures, and we compare them in three simple examples, one theoretical and two applied. The contribution of this review is to put all these information criteria into a Bayesian predictive context and to better understand, through small examples, how these methods can apply in practice.
Article
Full-text available
Previous research has shown that media violence exposure can cause desensitization to violence, which in theory can increase aggression. However, no study to date has demonstrated this association. In the present experiment, participants played a violent or nonviolent video game, viewed violent and nonviolent photos while their brain activity was measured, and then gave an ostensible opponent unpleasant noise blasts. Participants low in previous exposure to video game violence who played a violent (relative to a nonviolent) game showed a reduction in the P3 component of the event-related brain potential (ERP) to violent images (indicating physiological desensitization), and this brain response mediated the effect of video game content on subsequent aggressive behavior. These data provide the first experimental evidence linking violence desensitization with increased aggression, and show that a neural marker of this process can at least partially account for the causal link between violent game exposure and aggression.
Article
Purpose of Study: Creative industry is one of the economic sectors that important to get attention from the government. Absorption workers in the creative industry are very large considering the main capital in this industry is idea and creativity. An employee who works on large industry, often replaced by mechanical power by efficiency reason. This is different from the creative industry which had to take creativity owned by a human being to maintain its existence. The rattan industry in this area has been able to penetrate the export market in the Netherlands, Germany, The United States, and Australia. Although has penetrated the export market, a number of problems experienced by actors of the creative industry, such as difficulty in competition with other countries related to the design of products and the difficulties to participate in the event of the international exhibition. Methodology: The methodology used in this research is nondoctrinal, its describe clearly on local government policy in the development of the rattan creative industry in Trangsan Village, Gatak, Sukoharjo, Central Java. Results: This study found that one of an industry that relatively has been shown to be able to survive in the center of turbulence and national and global economic crisis is the rattan industry in Trangsan Village. Implications/Applications: Looking at these problems, the local government should take an active role to provide support for rattan industries of Trangsan Village by facilitating ease licensing and integrated technical assistance in connection to promoting both through social media and facilitation to be fit for as well as in international events.
Article
The current classification systems for video games are first attempts at protecting children from the real or imaginary influence of potentially harmful contents. These systems, however, are based on questionable principles, for two reasons. First, analyzing the Pan European Game Information (PEGI) and the Entertainment Software Rating Board (ESRB) from a pedagogical point of view, one cannot but notice that they are inherently flawed by contradictions and confusion of different perspectives. Second, these contradictions increase the difficulty for parents who buy video games to understand the rating. This is a considerable drawback, as parents and child caregivers should be the primary targets of such rating systems. This article offers a critical examination of the European PEGI and the North American ESRB rating systems, and, starting from this analysis, suggests improvements that could make video game rating systems more appropriate in terms of their function as parental guidance.
Article
Previous research has shown that playing violent video games increased aggressive tendencies. However, as pointed out by the General Learning Model (GLM) [Buckley, K. E., & Anderson, C. A. (2006). A theoretical model of the effects and consequences of playing video games. In: P. Vorderer & J. Bryant (Eds.), Playing video games motives responses and consequences (pp. 363–378). Mahwah, NJ: Lawrence Erlbaum], depending on their content, video games do not inevitably increase but may also decrease aggressive responses. Accordingly, the present research tested the hypothesis that playing prosocial video games decreases aggressive cognitions. In fact, playing a prosocial (relative to a neutral) video game reduced the hostile expectation bias (Experiment 1) and decreased the accessibility of antisocial thoughts (Experiment 2). Thus, these results lend credence to GLMs assumption that the effects of video game exposure depend to a great extent on the content of the game played.