Jurnal Varian

Finding the Factors Influencing the Severity of Traffic Accident Victims in Sleman Regency Using Ordinal Logistic Regression Analysis

Amalia Rizqi Cahyani, Mujiati Dwi Kartikasari — Mon, 25 Nov 2024 05:41:39 +0800

Special Region of Yogyakarta (Daerah Istimewa Yogyakarta, DIY) is well-known for its tourist, cultural, and educational attractions, but it also has a high accident rate. Sleman Regency is among the DIY regions with the greatest number of traffic accidents. According to Yogyakarta Police records, Sleman Regency had 1,825 traffic incidents in 2022, while 637 accidents occurred there in a short period of time in 2023, specifically from January to April. To stop the rising number of people injured in road accidents, this issue needs to be taken into account. The objective of this study was to examine the profile of traffic accidents that happened in Sleman Regency between January and April of 2023 and use the ordinal logistic regression method to find characteristics that influence the severity of traffic accidents. Sleman Regency traffic accident data was used in this study. The opponent's vehicle factor, with the category of four or more wheeled vehicles and non-motorized vehicles, is one of the elements that influences the severity of traffic accident victims in Sleman Regency, according to the study's findings.

Clustering of Study Program Using of Block-Based K-Medoids

Asa Nugrahaini Itsal Muna, Kariyam Kariyam — Mon, 25 Nov 2024 05:57:16 +0800

The purpose of this research is to classify Study Programs based on eleven mixed data from Internal
Quality Management System (QMS) indicators. This grouping can provide a clearer picture of how
QMS affects the performance and quality of study programs. By understanding these clusters, universities can identify and design more effective strategies to improve the quality of education. The data
used comes from the National Accreditation Board for Higher Education (BAN-PT) and the website
database, which consists of seven numerical variables: number of lecturers, percentage of doctors, percentage of professors and associate professors, student enumeration, percentage of graduates, program
experience, and availability of laboratories. Meanwhile, the categorical variable consists of four variables: National Accreditation Board of Higher Education (BAN-PT) research ranking, accreditation,
international recognition, and level of community service. The clustering method used is the blockbased k-medoids (block-based KM), and multivariate analysis of variance (MANOVA). We applied the
Deviation Ratio Index based on K-Medoids (DRIM) to determine the number of clusters. This research
results that the optimal number of groups that must be formed is three. Based on MANOVA the results
showed that the group consisting of 12 study programs had better QMS outcomes than the other two
groups.

Analyzing Lightning Strike Susceptibility Using the Elliptical Fitting Method with a Principal Component Analysis Approach

Mon, 25 Nov 2024 06:19:41 +0800

Lightning is a high-current discharge that occurs in Cumulonimbus clouds, with CG (Cloud to Ground)
lightning strikes posing significant dangers, especially to human life. Pasuruan, located in the highlands
between mountains and the ocean in Indonesia, is particularly vulnerable to such strikes. This study
aims to mitigate the impact of lightning strikes, particularly in industrial areas like Pasuruan, by delineating lightning-prone areas using a sophisticated methodological approach. Our research employs a
robust Ellipse Fitting Method, parameterized with Principal Component Analysis (PCA), to accurately
define the boundaries of these high-risk zones. The Ellipse Fitting Method, which involves forming
an ellipse from the intersection of a plane and a cone, uses five key parameters: a center point, two
vertex points, and two focus points. PCA is then applied to these parameters to determine the ellipse’s
configuration, with the center point derived from the mean of all data points. The major and minor
axes are defined by the first and second eigenvalues of the principal components, respectively. The size
of the ellipse correlates with the confidence level, with higher confidence resulting in a larger ellipse.
The result of integrating these advanced techniques is the generation of two PCA models from data
collected across 28 sub-districts in Pasuruan, with findings indicating a high level of vulnerability in
Lumbang District and a moderate level of risk in Gempol District. This methodological framework not
only enhances the precision in identifying lightning-prone areas but also provides a scalable approach
for similar studies in other regions. Suggestion for the further research are to overcome extreme points
or extreme points in the PCA confidence ellipse such as MVEE.

Mathematical Modelling and Simulation Strategies for Controlling Damage to Forest Resources Due to Illegal Logging

Irene R Naben, Elinora N Bano, Fried M. A. Blegur — Mon, 25 Nov 2024 06:31:34 +0800

Forests are one of the natural resources that provide many benefits for the welfare of living things.
The dense population causes people to depend more on forest resources. One of them is illegal logging.
Various strategies to control forest damage due to illegal logging have been carried out, namely by direct
handling to improve damaged conditions while preventing the recurrence of forest damage. The purpose
of this research is to build a mathematical model of forest resource damage control strategies due to
illegal logging, determine assumptions, formulate the model, and conduct analysis and problem solving
including: determining the equilibrium point, determining the stability analysis of the equilibrium point,
and conducting numerical simulations of the equilibrium point. The last step is to interpret the results of
the analysis obtained and make conclusions. Based on the research and simulation results of the model,
it can be concluded that taking into account the variable of forest resource damage control strategy
due to illegal logging, the result shows that if the density of forest resources has been affected by the
disturbance of population density around the forest, it is necessary to have a forest resource damage
control strategy in order to compensate for the people around the forest who do a lot of illegal logging.
In order to maintain the forest so that the forest does not quickly become extinct and can overcome
drought, prevent flooding, maintain groundwater quality, protect animals, reduce air pollution, climate
control, reduce dust particles, prevent the greenhouse effect, supply natural fertilizers, prevent erosion,
and maintain springs.

Analysis of Underdeveloped Regency Using Logistic Threshold Regression Model

Annisa Nur Salsabila, Siskarossa Ika Oktora — Mon, 25 Nov 2024 06:38:11 +0800

Regional development inequality causes some regions to lag behind other regions. An underdeveloped
regency is a regency where territories and people are less developed than other regions nationally. The
government has set a Human Development Index (HDI) target of 62.2 to 62.7 to accelerate the development of underdeveloped regency and prevent the regions from lagging. This study aims to evaluate
the HDI target and obtain the HDI value that reduces the risk of underdeveloped regency and acquires
variables that affect underdeveloped regency’s status. The logistic threshold regression model is used
in this study with HDI as the threshold variable, 22 indicators for determining underdeveloped regency
as explanatory variables, and the underdeveloped regency’s status as the response variable. Threshold
regression can handle non-linear relationships between response and explanatory variables, including
various types of threshold models such as step, segmented, hinge, stegmented, and upper hinge. By applying a hinge threshold regression model using the R package ’chngpt,’ this study addresses non-linear
relationships and categorical responses. The results showed a threshold effect with a threshold value of
62.9, indicating that the HDI target can reduce the region’s risk of being left behind.

The Weibull Regression Model Analysis of Mahakam River Water Pollution Potential

Zalva Pradipa, Suyitno Suyitno, Meiliyani Siringoringo — Mon, 25 Nov 2024 06:50:16 +0800

Mahakam River has a vital role in the lives of the people of the East Kalimantan province, including
providing a raw source of clean water. The multi-activity of the Mahakam River watershed, as a water
traffic lane, mining, fisheries, hotels, restaurants, and resident houses, has the potential to produce waste
into the water. Increasing waste in the water flow can increase the pollution potential of river water,
threatening people’s health. Therefore, precaution is necessary. In this research, statistical prevention
was proposed, providing information to the East Kalimantan people regarding the factors affecting the
pollution potential of the Mahakam River through Weibull regression (WR) modeling on dissolved oxygen (DO) data 2022. Research data was secondary data provided by the Life Environmental Department
of East Kalimantan province. The WR model is a Weibull distribution that is directly influenced by covariates. WR model consists of Weibull survival regression, cumulative distribution regression, hazard
regression, and Weibull mean regression. This research aims to obtain the factors affecting the pollution potential and to provide the pollution potential information of Mahakam River 2022. The research
concluded that factors influencing the pollution potential of the Mahakam River were watercolor degree
and nitrate concentration. Applying the WR model to DO data 2022 was able to provide the pollution
potential information of Mahakam River, namely the probability of river water isn’t polluted is 0.6555,
or the probability of the polluted river water is 0.3445, the pollution rate is 6 locations are polluted for
every 10 mg/L DO, and the DO average of river water is 5.7450 mg/L. Increasing water color degree
and nitrate concentration will decrease the probability of the Mahakam River being polluted, increase
the probability of the Mahakam River being polluted, increase the pollution rate, and reduce the DO of
Mahakam River water.

A Kernel Logistic Regression Approach to Understanding the “Banyak Anak Banyak Rezeki” Stigma

Mon, 25 Nov 2024 06:51:48 +0800

Indonesia, the world’s 4th largest country with a population of 270 million in 2020, faces many challenges due to rapid population growth, including biodiversity loss and increased consumption of natural
resources. One of the cultural factors underlying the high rate of population growth is the perception of
“Banyak Anak Banyak Rezeki“ that develops in the community. This study aims to identify and model
the factors that influence the “Banyak Anak Banyak Rezeki” stigma and find solutions to overcome this
problem. The research method used was quantitative, with a sample of 384 people in South Sulawesi,
consisting of Bugis, Makassar, Toraja, and Mandar tribes. The variables studied include religiosity,
tradition, number of children, and cognitive dissonance. The analysis techniques used were logistic
regression (LR) and kernel logistic regression (KLR). The results showed that religiosity, number of
children, and cognitive dissonance had a significant effect on the “Banyak Anak Banyak Rezeki” stigma.
The accuracy of the LR model reached 87.01% and increased to 93.51% after using KLR.

Application of VAR-GARCH for Modeling the Causal Relationship of Stock Prices in the Mining Sub-sector

Muhammad Nasrudin, Endah Setyowati, Shindi Shella May Wara — Mon, 25 Nov 2024 06:58:48 +0800

Accurate modeling is expected to minimize risk and maximize profit in investment portfolios, one of
which is in stock price modeling. This research aims to model the causal relationship between stock
prices using the Vector Autoregressive - Generalized Autoregressive Conditional Heteroskedasticity
(VAR-GARCH) model. The VAR-GARCH model is used to overcome heteroscedasticity and model
dynamic volatility. The data used for the modeling consists of daily stock prices from July 2023 to
May 2024 for mining sub-sector companies listed on the Jakarta Islamic Index (JII), including ADMR,
ADRO, and ANTM. The results showed that the VAR(1) model is stable, but this model indicates the
presence of heteroskedasticity or ARCH effects. Therefore, the VAR(1) model was combined with the
GARCH model, and the results showed that the best model is VAR(1)-GARCH(1,1). The VAR(1)-
GARCH(1,1) model is appropriate and meets the homoskedasticity assumptions for modeling the stock
prices of the mining sub-sector in the Jakarta Islamic Index (JII). This indicates that the VAR-GARCH
model could successfully handle the volatility of stock price data. In general, this research is in line
with previous research, i.e., the VAR-GARCH model showed a better model for capturing the volatility
patterns in the data.

Modeling the Number of High School Dropouts Using GWGPR

Mon, 25 Nov 2024 07:04:20 +0800

In 2022, the high school dropout rate is the highest compared to other levels of education in Indonesia.
Seeing the urgency of the 12-year Compulsory Education program, completing education up to the high
school level is an important thing that needs to be considered. Thus, it is necessary to know the factors
that influence the dropout rate in the hope that this problem can be reduced. This study aims to model
the high school dropout rate using geographically weighted generalized poisson regression (GWGPR)
based on the factors that influence it. GWGPR is used if the response variable is overdispersed and
depends on the location observed. The results of this study indicate that each province has a different regression model. The GWGPR model with the adaptive tricube kernel weighting function is the
best model because it has the smallest AIC value compared to other weighting functions. In Central
Sulawesi Province, the GWGPR model with the adaptive tricube kernel weighting function formed is
µˆ26 = exp (8, 1267 − 0, 1267X4 + 0, 0344X5 + 0, 0957X6 + 0, 1173X7). With the significant variables are the average length of schooling, the percentage of the population aged 7-17 years who receive
PIP, the open unemployment rate, and the percentage of children who do not live with parents.

Stock Price Index Prediction Using Random Forest Algorithm for Optimal Portfolio

Putri Humairah, Dina Agustina — Mon, 25 Nov 2024 00:00:00 +0800

With a majority Muslim population in Indonesia, Islamic capital markets such as the Jakarta Islamic
Index (JII) are a relevant choice because the JII is an investment index that complies with Sharia principles. This research aims to predict stock prices in the JII using the Random Forest (RF) algorithm and
form an optimal portfolio with the Mean-Variance Efficient Portfolio (MVEP) model. The data used is
the daily closing price of JII stocks from April 2023 to March 2024, obtained from the Indonesia Stock
Exchange and Yahoo Finance. The RF method is used to predict stock prices, with model performance
evaluation using Mean Absolute Percentage Error (MAPE). The results showed that the application of
ML with the RF algorithm in predicting stock prices produced very good predictions because the evaluation results using MAPE were in the 0%-10% range, namely a value of 2.522% for ACES shares;
1.222% for ICBP shares, and 0.760% for INDF shares. The optimal portfolio formed using MVEP
produces a stock composition with a weight of 7.64% for ACES, 22.46% for ICBP, and 69.90% for
INDF. The optimal portfolio’s estimated expected return and risk are 0.0546% and 0.0103%.