Claim Missing Document
Check
Articles

Found 3 Documents
Search
Journal : Science and Technology Indonesia

LSTM-CNN Hybrid Model Performance Improvement with BioWordVec for Biomedical Report Big Data Classification Kurniasari, Dian; Warsono; Usman, Mustofa; Lumbanraja, Favorisen Rosyking; Wamiliana
Science and Technology Indonesia Vol. 9 No. 2 (2024): April
Publisher : Research Center of Inorganic Materials and Coordination Complexes, FMIPA Universitas Sriwijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26554/sti.2024.9.2.273-283

Abstract

The rise in mortality rates due to leukemia has fueled the swift expansion of publications concerning the disease. The increase in publications has dramatically affected the enhancement of biomedical literature, further complicating the manual extraction of pertinent material on leukemia. Text classification is an approach used to retrieve pertinent and top-notch information from the biomedical literature. This research suggests employing an LSTM-CNN hybrid model to tackle imbalanced data classification in a dataset of PubMed abstracts centred on leukemia. Random Undersampling and Random Oversampling techniques are merged to tackle the data imbalance problem. The classification model’s performance is improved by utilizing a pre trained word embedding created explicitly for the biomedical domain, BioWordVec. Model evaluation indicates that hybrid resampling techniques with domain-specific pre-trained word embeddings can enhance model performance in classification tasks, achieving accuracy, precision, recall, and f1-score of 99.55%, 99%, 100%, and 99%, respectively. The results suggest that this research could be an alternative technique to help obtain information about leukemia.
On the Characteristic Function of the Four-Parameter Generalized Beta of the Second Kind (GB2) Distribution and Its Approximation to the Singh-Maddala, Dagum, and Fisk Distributions Warsono; Kurniasari, Dian
Science and Technology Indonesia Vol. 10 No. 1 (2025): January
Publisher : Research Center of Inorganic Materials and Coordination Complexes, FMIPA Universitas Sriwijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26554/sti.2025.10.1.201-211

Abstract

Researchers have thoroughly investigated generalized distributions due to their inherent flexibility, which allows them to include several well-known distributions as special cases. Among these, the four parameter Generalized Beta of the Second Kind (GB2) distribution stands out as one of the most versatile frameworks in probability theory. Despite its broad applications, the GB2 distribution’s characteristic function, a critical tool in probability and statistical analysis, lacks a closed-form solution in the existing literature. This study pursues two primary objectives: first, to derive the characteristic function and the kth moment of the GB2 distribution, and second, to demonstrate how the GB2 distribution can serve as a close approximation to the Singh-Maddala, Dagum, and Fisk distributions using its characteristic function and kth moment. These derivations and approximations rely on gamma and beta functions, supplemented by the Maclaurin series expansion.
Modeling Vector Error Correction with Exogeneous (VECMX) Variable for Analyzing Nonstationary Variable Energy Used and Gross Domestic Product (GDP) Usman, Mustofa; Wamiliana; Russel, Edwin; Kurniasari, Dian; Widiarti; Elfaki, Faiz A.M
Science and Technology Indonesia Vol. 10 No. 1 (2025): January
Publisher : Research Center of Inorganic Materials and Coordination Complexes, FMIPA Universitas Sriwijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26554/sti.2025.10.1.283-293

Abstract

Analysis of energy used, GDP and population has been carried out in many countries and has become a topic of interest for many researchers and governments. This is because energy used is an important factor for society and industry in a country. In this study, the modeling of the relationship between energy used, GDP and population as an exogenous variable for the cases of Indonesia from 1967-2023 will be discussed. The energy used and GDP data are nonstationary with order one, I(1), and there is cointegration between energy used and GDP. Therefore, the model which will be used is the Vector Error Correction Model with Exogenous variable (VECMX) with population as the exogenous variable. From the results of analysis, the best model is VECMX(3,1) with cointegration rank R=1. Based on this model, the pattern of the relationship among the three variables, Granger-causality between energy used and GDP, exogenous impact on energy used and GDP, and forecasting for the next 10 years will be discussed.