cover
Contact Name
Mesran
Contact Email
mesran.skom.mkom@gmail.com
Phone
-
Journal Mail Official
jurnal.bits@gmail.com
Editorial Address
-
Location
Kota medan,
Sumatera utara
INDONESIA
Building of Informatics, Technology and Science
ISSN : 26848910     EISSN : 26853310     DOI : -
Core Subject : Science,
Building of Informatics, Technology and Science (BITS) is an open access media in publishing scientific articles that contain the results of research in information technology and computers. Paper that enters this journal will be checked for plagiarism and peer-rewiew first to maintain its quality. This journal is managed by Forum Kerjasama Pendidikan Tinggi (FKPT) published 2 times a year in Juni and Desember. The existence of this journal is expected to develop research and make a real contribution in improving research resources in the field of information technology and computers.
Arjuna Subject : -
Articles 926 Documents
Perbandingan Algoritma LSTM, Bi-LSTM, GRU, dan Bi-GRU untuk Prediksi Harga Saham Berbasis Deep Learning Tshamaroh, Muthia; Permana, Inggih; Salisah, Febi Nur; Muttakin, Fitriani; Afdal, M
Building of Informatics, Technology and Science (BITS) Vol 7 No 1 (2025): June (2025)
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i1.7252

Abstract

Stock price prediction is an important component in making investment decisions. This study aims to compare the performance of four deep learning models, namely LSTM, Bi-LSTM, GRU, and Bi-GRU, in predicting stock prices, in order to find the most optimal model for the implementation of an accurate stock price prediction system. Five years of historical data undergoes normalization, windowing, and is separated into training data, validation data, and test data. Model training is conducted with different settings of batch size, timestep, and three kinds of optimizers (Adam, SGD, RMSprop). Performance assessment employs MSE, RMSE, MAE, and R² measurements. The findings indicate that the Bi-GRU model utilizing Adam optimizer settings, a batch size of 8, and a timestep of 21 yields the highest performance, achieving an MSE of 0.0003, an RMSE of 0.0169, an MAE of 0.0129, and an R² of 0.9438. This model demonstrates a strong capability to identify intricate patterns and long-term temporal relationships, outperforming other models in accuracy. The results advocate for the establishment of a predictive system that aids investors and firms in making strategic decisions based on data.
Segmentasi Produk Minuman Tidak Termasuk Produk Susu Berdasarkan Informasi Nilai Gizi Menggunakan Metode DBSCAN Rachmatia, Baiq Wita; Primandari, Arum Handini
Building of Informatics, Technology and Science (BITS) Vol 7 No 1 (2025): June (2025)
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i1.7255

Abstract

Approximately 28.7% of Indonesians consume sugar, salt, and fat (SSF) in amounts that exceed the Ministry of Health's recommended limits. Over the past two decades, sweetened drink (MBDK: minuman berpemanis dalam kemasan) consumption has surged, making Indonesia the third highest in Southeast Asia for MBDK consumption. To mitigate this, consumers need clear information about GGL content, but nutritional labels are often complex and underutilized. Product segmentation can help consumers make healthier drink choices and support health interventions aimed at reducing risky consumption. Data on GGL values were collected from MBDK sold in three store types and analyzed using the DBSCAN method, which handles diversity and outliers without predefining cluster numbers. Descriptive statistics showed most products had low fat but higher sugar content, nearing 15 grams. After standardizing the data using z-scores, the DBSCAN clustering revealed two clusters and some noise. The evaluation indicated a silhouette coefficient of 0.396 and a Dunn index of 0.137, with t-tests showing significant differences between the clusters.
Perbandingan Algoritma NBC, SVM, Logistic Regression untuk Analisis Sentimen Terhadap Wacana KaburAjaDulu di Media Sosial X Rohman, Adib Annur; Trisnapradika, Gustina Alfa
Building of Informatics, Technology and Science (BITS) Vol 7 No 1 (2025): June (2025)
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i1.7261

Abstract

This research aims to analyze sentiment towards KaburAjaDulu discourse on X social media by utilizing Logistic Regression, Support Vector Machine (SVM), and Naive Bayes algorithms. Data was collected through a crawling process and resulted in 3,011 tweet data. Pre-processing stages include data cleaning, conversion of letters to lowercase, normalization, tokenization, stopword removal, and stemming. After preprocessing, the data was divided into two sentiment categories, namely positive and negative using a lexicon approach. The dataset is divided using an 80:20 scheme for training and test data, with feature representation utilizing the TF-IDF method. The modeling process is performed utilizing the three algorithms to be evaluated using accuracy, precision, recall, and f1-score metrics. As a solution to class inequality, the oversampling technique SMOTE (Synthetic Minority Over-sampling Technique) is applied. Based on the evaluation, it shows that before the application of SMOTE, Naive Bayes algorithm obtained 78.18% accuracy, 81.80% precision, 77.06% recall, and 77.35% f1-score; SVM obtained 85.63% accuracy, 86.49% precision, 85.68% recall, and 85.94% f1-score; while Logistic Regression obtained 83.05% accuracy, 85.31% precision, 82.47% recall, and 82.95% f1-score. After applying SMOTE, Naive Bayes improved to 81.90% accuracy, 82.27% precision, 81.67% recall, and 81.87% f1-score; SVM obtained 85.63% accuracy, 87.59% precision, 86.89% recall, and 87.13% f1-score; and Logistic Regression obtained 83.33% accuracy, 84.46% precision, 83.62% recall, and 83.88% f1-score. These findings prove that SVM has the most consistent and superior sentiment classification performance on this dataset, making an important contribution to the development of methods for analyzing people's views on social media platforms.
Optimalisasi Model SciBERT dengan Attention-BiLSTM-CRF untuk Pengenalan Entitas Penyakit dalam Teks Biomedis Pamungkas, Tahta Arya; Salam, Abu
Building of Informatics, Technology and Science (BITS) Vol 7 No 1 (2025): June (2025)
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i1.7263

Abstract

This research aims to improve the performance of medical entity recognition in biomedical text by modifying the SciBERT model with Attention-BiLSTM-CRF. Although SciBERT, based on the BERT architecture and trained on biomedical text data, has proven effective in entity recognition, it still has limitations in handling complex medical entities, especially nested entities. As a solution, this research integrates Attention, BiLSTM, and CRF components into the SciBERT model to enhance entity recognition accuracy. Experimental results show that the SciBERT + Attention-BiLSTM-CRF model outperforms the SciBERT model across all key evaluation metrics. Precision improved by 1.7% (from 0.8221 to 0.8364), Recall increased by 2.9% (from 0.8537 to 0.8768), and F1-Score increased by 2.1% (from 0.8372 to 0.8554). These improvements demonstrate that this modification significantly enhances the model's ability to recognize more complex medical entities in biomedical text. The addition of Attention and BiLSTM enriches contextual understanding, while CRF ensures consistency across entity labels. These results indicate that this approach could significantly contribute to automated systems in processing medical data.
Analisis Sentimen Masyarakat Terhadap Kebijakan IKN Pada Periode Jokowi dan Prabowo Menggunakan Algoritma NBC, SVM, dan K-NN Nasution, Nur Shabrina; Permana, Inggih; Salisah, Febi Nur; Afdal, M; Megawati, Megawati
Building of Informatics, Technology and Science (BITS) Vol 7 No 1 (2025): June (2025)
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i1.7276

Abstract

The relocation of the National Capital City (IKN) from Jakarta to East Kalimantan has generated a variety of responses from the Indonesian people recorded through social media, especially platform X. This study aims to analyze and compare public sentiment towards the IKN policy in two periods of government, namely President Joko Widodo and President Prabowo Subianto. This study aims to analyze and compare public sentiment towards the policy of the National Capital City during two periods of government, namely President Joko Widodo and President Prabowo Subianto, using a machine learning approach. The three algorithms used in sentiment classification are Naive Bayes Classifier (NBC), Support Vector Machine (SVM), and K-Nearest Neighbor (K-NN). The research process includes data crawling (600 data each per period), text preprocessing (cleaning, tokenizing, filtering, stemming), data labeling using Lexicon-Based approach with InSet dictionary, and weighting using TF-IDF method. The results of the analysis show that in the Jokowi period, public sentiment tends to be more balanced, with the dominance of negative sentiment (35.9%), followed by positive sentiment (33.4%) and neutral (30.7%). Whereas in the Prabowo period, negative sentiment increased to 40.3%, while positive decreased to 26.3%. Based on the model accuracy evaluation, in the Jokowi period, the NBC algorithm showed the best performance with an accuracy of 73%, while in the Prabowo period, the SVM algorithm excelled with the highest accuracy reaching 81%. These findings provide a dynamic picture of public perception of IKN policies under two different governments.
Studi Perbandingan Metode MABAC dan WASPAS dengan Pembobotan ROC dalam Sistem Pendukung Keputusan Pemilihan Supplier Terbaik Pratiwi, Heny; Sa’ad, Muhammad Ibnu; Hasiholan, Jundro Daud
Building of Informatics, Technology and Science (BITS) Vol 7 No 1 (2025): June (2025)
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i1.7278

Abstract

The selection of the right supplier is a crucial factor in the supply chain to ensure product quality, cost efficiency, and timely delivery. This study aims to determine the best supplier by comparing two multi-criteria decision-making methods: Multi-Attributive Border Approximation Area Comparison (MABAC) and Weighted Aggregated Sum Product Assessment (WASPAS). Five key criteria were used in the evaluation: product quality, price, delivery punctuality, service and responsiveness, and reputation and trust. The analysis results show that PT. Indo Makmur (A1) consistently ranked first in both methods, with the highest scores of 0.456 (MABAC) and 0.982 (WASPAS), making it the recommended supplier. PT. Sukses Bersama (A7) and PT. Cahaya Abadi (A3) ranked second and third in both methods, indicating good performance. Meanwhile, UD. Sentosa Jaya (A4) ranked the lowest in both methods, suggesting that this supplier is less competitive than the other alternatives. The comparison of results between MABAC and WASPAS methods demonstrates ranking consistency, confirming that both methods can be reliably used in decision-making. This study provides data-driven recommendations for companies in selecting the best supplier, thereby enhancing supply chain efficiency and supporting long-term business strategies.
Pengembangan Chatbot Kesehatan Mental Berbasis Web Menggunakan Model Long Short-Term Memory (LSTM) Ardin, Akbar Ilham; Salam, Abu
Building of Informatics, Technology and Science (BITS) Vol 7 No 1 (2025): June (2025)
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i1.7282

Abstract

Mental health issues such as stress, anxiety, and academic burnout are increasingly prevalent among university students. However, many students remain reluctant or unable to access counseling services due to time limitations, social stigma, and a lack of available professionals. This study aims to develop CuraBot, a web-based chatbot designed to provide preliminary emotional support and mental health education in an instant, anonymous, and easily accessible manner for students. The system was developed using the Long Short-Term Memory (LSTM) algorithm, which is proven to be effective in understanding contextual text-based conversations. The dataset used consists of 1,624 conversational entries across 77 intent classes, adapted and localized from an open-source corpus to reflect the linguistic style and needs of Indonesian students. The development process involved several stages, including data preprocessing (lemmatization, tokenization, stopword removal, and padding), model training using TensorFlow, and deployment into a Flask-based web application. The model was evaluated using a separate test set of 244 entries, resulting in an accuracy of 89.9%, precision of 90.4%, recall of 89.1%, and an F1-score of 89.8%. These results indicate that the model can classify user intent with high accuracy. This research contributes to the development of a contextual, practical, and AI-based digital solution that supports early access to psychological services within university environments.
Prediksi Potensi Kinerja Calon Karyawan Customer Service Call Center Menggunakan Model Machine Learning Berbasis Data Rekrutmen Pratama, Andriyan Yoga; Ghozi, Wildanil
Building of Informatics, Technology and Science (BITS) Vol 7 No 1 (2025): June (2025)
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i1.7285

Abstract

Employee selection process is a critical stage for companies in acquiring high-quality human resources (HR), particularly for customer service call center positions that demand excellent communication skills and strong work endurance. Data-driven recruitment methods have demonstrated improved accuracy compared to traditional, often subjective, approaches. This study aims to develop a predictive model to assess the potential performance of candidates during the HR interview stage, based on educational background, work experience, and other relevant factors, using machine learning algorithms. The dataset utilized includes demographic information, education levels, previous work experience, and other factors that may influence candidate performance in customer service roles. The models tested in this study include Decision Tree, Random Forest, and Artificial Neural Network algorithms. The analysis shows that GPA, prior work experience, and organizational involvement significantly correlate with the potential performance of candidates. The application of machine learning in the recruitment process can enhance selection effectiveness and improve HR efficiency. Through this approach, companies are expected to make more accurate hiring decisions and select the best candidates with greater precision.
Analisis Sentimen Masyarakat Terhadap Liga Indonesia Menggunakan Algoritma Naïve Bayes Classifier dan Support Vertor Machine Pada Platform X dan YouTube Irwanda, Mahyuda; Afdal, M; Novita, Rice; Zarnelly, Zarnelly
Building of Informatics, Technology and Science (BITS) Vol 7 No 1 (2025): June (2025)
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i1.7294

Abstract

The Indonesian League is a national football competition that attracts a lot of public attention. However, various problems such as controversial referee decisions, fan riots, and match-fixing issues are often in the spotlight. This study aims to analyze public sentiment towards the Indonesian League using the Naïve Bayes Classifier (NBC) and Support Vector Machine (SVM) algorithms. Data were collected from social media platform X (Twitter) as many as 2000 tweets and YouTube as many as 2000 comments in the period from January 2023 to December 2024. After going through preprocessing stages such as cleaning, case folding, tokenizing, stopword removal, and stemming, the data was classified into positive, negative, and neutral sentiments. The results showed that SVM had a higher accuracy (99%) than NBC (85%) in sentiment analysis.
Inventory Optimization through FP-Growth-Based Association Rule Mining of Material Stock Usage Patterns Ubaidillah, Ubaidillah; Sumiati, Sumiati
Building of Informatics, Technology and Science (BITS) Vol 7 No 1 (2025): June (2025)
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i1.7306

Abstract

This research focuses on analyzing the usage patterns of material stock at PT Baruna Energi Solusindo Teknik using Association Rule Mining (ARM) with the FP-Growth algorithm. The company often experiences material shortages or excess inventory due to manual and inaccurate stock planning. To address this issue, the study aims to discover frequent itemsets and association rules among materials used in water purification installations, enabling more data-driven procurement decisions. The research employs secondary data on material usage transactions from April 2024 to March 2025, which is processed using RapidMiner software. The FP-Growth algorithm identifies material combinations with high support and confidence values. For instance, Membrane RO has a support value of 75.9%, indicating that it is used in over three-quarters of all projects. Additionally, the combination {Membrane RO, PVC Pipe 1"} → SDL PVC 1" shows a confidence value of 83.3%, signifying a strong association among these items. The results suggest that stock optimization can be achieved by prioritizing frequently used items and associated combinations in procurement planning. This method not only improves inventory efficiency but also helps prevent stockouts and overstocking. The FP-Growth algorithm proves to be suitable and effective in identifying meaningful patterns in stock usage data.