Article Per Year (5 Year)
p-Index From 2020 - 2025
0.61
P-Index
This Author published in this journals
All Journal BAREKENG: Jurnal Ilmu Matematika dan Terapan Zeta - Math Journal Jurnal Nasional Pendidikan Teknik Informatika (JANAPATI)
Salsabila, Fatiha Nadia
Unknown Affiliation
Computer Science & IT Control & Systems Engineering Economics, Econometrics & Finance Education Energy Engineering Mathematics Mechanical Engineering Physics Transportation

Published : 4 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 4 Documents
Search

 1 
Optimizing Brain Tumor MRI Classification with Transfer Learning: A Performance Comparison of Pre-Trained CNN Models Mardianto, M. Fariz Fadillah; Pusporani, Elly; Salsabila, Fatiha Nadia; Nitasari, Alfi Nur; Lu’lu’a, Na’imatul
Jurnal Nasional Pendidikan Teknik Informatika: JANAPATI Vol. 14 No. 1 (2025)
Publisher : Prodi Pendidikan Teknik Informatika Universitas Pendidikan Ganesha

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.23887/janapati.v14i1.87377

Abstract

This study aims to classify brain MRI images into several types of brain tumors using the Convolutional Neural Network (CNN) approach with transfer learning. This method has the advantage of processing complex images in a shorter time than conventional CNN approaches. In this study, the data used was a public database from Kaggle, which consisted of four categories: glioma, meningioma, no tumor, and pituitary. Before entering the transfer learning process, data augmentation is carried out on the training data. Four pre-trained CNN models were used: VGG19, ResNet50, InceptionV3, and DenseNet121. The four models compared their ability to classify MRI images with several evaluation metrics: accuracy, precision, recall, and F1 score. The results of the performance comparison of the four pre-trained models show that the ResNet50 is the best model, with an accuracy of 98%. Meanwhile, VGG19, DenseNet121, and InceptionV3 produce 97%, 96%, and 95% accuracy, respectively. The ResNet50 architecture demonstrated superior performance in brain tumor classification, achieving 98% accuracy. It can be attributed to its residual learning structure, which efficiently manages complex MRI features.  Further research should concentrate on larger, more diverse datasets and advanced preprocessing techniques to enhance model generalizability.
COMPARISON OF POISSON REGRESSION AND GENERALIZED POISSON REGRESSION IN MODELING THE NUMBER OF INFANT MORTALITY IN WEST JAVA 2022 Saifudin, Toha; Salsabila, Fatiha Nadia; Fitriani, Mubadi'ul; Kholidiyah, Azizatul; Auliyah, Nina; Ariani, Fildzah Tri Januar; Suliyanto, Suliyanto
BAREKENG: Jurnal Ilmu Matematika dan Terapan Vol 19 No 1 (2025): BAREKENG: Journal of Mathematics and Its Application
Publisher : PATTIMURA UNIVERSITY

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30598/barekengvol19iss1pp35-50

Abstract

In line with the Sustainable Development Goals (SDGs), the Infant Mortality Rate (AKB) is a very important health indicator, especially in neonatal and perinatal care. West Java Province consistently ranks third nationally in terms of infant mortality in 2020 and 2021. This study analyzes the factors influencing infant mortality in West Java in 2022 using secondary data from the 2022 West Java Provincial Health Profile. The response variable is the number of infant deaths, while the predictor variables include the percentage of K-4 coverage (X1), high-risk pregnancy (X2), family with PHBS (X3), exclusive breastfeeding (X4), and complete immunization coverage (X5). Given the count-based nature of the data, Poisson regression was used, which assumes equidispersion where the variance is equal to the mean. However, the analysis found overdispersion, where the variance significantly exceeds the mean, making Poisson regression unsuitable. To address this, Generalized Poisson Regression (GPR) was applied, as GPR introduces a dispersion parameter that accounts for overdispersion, thus better fitting the data. The initial Poisson regression results showed that X1, X2, X4, and X5 significantly influenced infant mortality, while the GPR model showed that only X2 and X3 were significant factors, with a dispersion parameter of -3.116. The GPR model shows that every additional one high-risk pregnancy increases the infant mortality rate by 1.00006, while an increase of one unit of clean and healthy living practices reduces the mortality rate by 2.66%. Model evaluation using AIC, BIC, and RMSE confirmed that the GPR model better described the relationship between predictor variables and infant mortality rates compared to Poisson regression. These findings emphasize the need to use GPR to model cases with overdispersion in count data, so as to provide more reliable information for policy and intervention strategies.
Reduksi Faktor-Faktor yang Mempengaruhi Kualitas Air Hujan dengan Metode Analisis Komponen Utama Nitasari, Alfi Nur; Salsabila, Fatiha Nadia; Ramadhanty, Devira Thania; Anggriawan, Muhammad Rizal; Amelia, Dita; Mardianto, M. Fariz Fadillah; Ana, Elly
Zeta - Math Journal Vol 8 No 1 (2023): Mei
Publisher : Universitas Islam Madura

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31102/zeta.2023.8.1.7-15

Abstract

The increase in air pollution that occurs due to industrial and economic activities will affect the rainwater content. The vital function of one of the sources of clean, namely rainwater, makes the urgency of research on indications of a decrease in rainwater quality by analyzing the substances contained. The parameters used for the determination of rainwater quality are the degree of acidity, conductivity, calcium, magnesium, sodium, potassium, ammonium, sulfate ions, nitrate ions, chloride ions, total hardness, and acidity. The data source used is data from the Central Statistics Agency (BPS) entitled "Indonesian Environmental Statistics 2022". By using the main component analysis method to reduce the variables of rainwater quality influence factors in 35 city/station samples in Indonesia, results were obtained, namely rainwater quality influence factors that can be formed into three main components with each containing 10 variables consisting of acidity, conductivity, magnesium, sodium, potassium, chloride ions, sulfate ions, nitrate ions, total hardness, and acidity capable of explaining 73.695% of the total variance.
UNILEVER STOCK PRICES FORECASTING WITH ENSEMBLE AVERAGING APPROACH ARIMA-GARCH AND SUPPORT VECTOR REGRESSION Pusporani, Elly; Nitasari, Alfi Nur; Salsabila, Fatiha Nadia; Indrasta, Irma Ayu; Mardianto, M. Fariz Fadillah
BAREKENG: Jurnal Ilmu Matematika dan Terapan Vol 20 No 1 (2026): BAREKENG: Journal of Mathematics and Its Application
Publisher : PATTIMURA UNIVERSITY

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30598/barekengvol20iss1pp0137-0154

Abstract

Investment, mainly in stock prices, plays a significant role in the Indonesian economy. Accurate stock price forecasting can help investors make informed decisions. Unilever Indonesia Tbk (UNVR) exhibits high volatility in its closing stock prices, making it crucial to develop a reliable forecasting model. This study applies an ensemble averaging method that integrates the ARIMA-GARCH model and Support Vector Regression (SVR) to predict UNVR's closing stock prices from January 6, 2019, to November 5, 2023. The results indicate that the data can be modeled using ARIMA (0,2,1). However, the squared residuals of the model show heteroscedasticity, necessitating variance modeling using the ARCH-GARCH approach. The best combination of mean and variance modeling is achieved with ARIMA (0,2,1) – GARCH (1,1), yielding a Mean Absolute Percentage Error (MAPE) of 2.865%. Additionally, a nonparametric SVR model with parameters C = 4 and ε = 0 is applied, resulting in a MAPE of 2.94%. An ensemble averaging approach is implemented to optimize forecasting accuracy further, combining ARIMA-GARCH and SVR models. This ensemble approach improves predictive performance, achieving a final MAPE of 1.682%. These findings demonstrate that ensemble averaging effectively enhances stock price forecasting accuracy by leveraging linear and nonlinear modeling techniques.
Co-Authors Anggriawan, Muhammad Rizal Ariani, Fildzah Tri Januar Auliyah, Nina Dita Amelia, Dita Elly Ana Elly Pusporani Fitriani, Mubadi'ul Indrasta, Irma Ayu Kholidiyah, Azizatul Lu’lu’a, Na’imatul M. Fariz Fadillah Mardianto Nitasari, Alfi Nur Ramadhanty, Devira Thania Suliyanto Suliyanto Toha Saifudin