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 777 Documents
Mental Health Sentiment Analysis on Twitter using Ensemble Learning Algorithm Aziz, Kemal; Wahyudi, Bambang Ari; Palupi, Irma
Building of Informatics, Technology and Science (BITS) Vol 7 No 2 (2025): September 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i2.7763

Abstract

Mental health problems have become an important health issue around the world. Poor understanding as well as low mental health awareness contribute to mental health healing efforts. In particular, Social media is becoming a platform for people to convey feelings and emotions. A dataset of 20,000 English tweets, equally divided into 10,000 depressed and 10,000 non-depressed tweets, which were cleaned and processed using Term Frequency-Inverse Document Frequency (TF-IDF) for feature extraction. The method used in this sentiment analysis introduces an ensemble learning framework that combines Naïve Bayes, Support Vector Machine, and Random Forest classifiers, using majority voting for prediction. Each classifier was optimized using the best parameters, and the models were validated through 5-fold cross-validation. The experimental results show that Naïve Bayes with α = 1 achieved an accuracy of 76.23% while Random Forest with 5000 trees at 76.77%, and Support Vector Machine with a linear kernel at 75.32%. By combining these classifiers, the ensemble model reached the highest accuracy of 77.88%, demonstrating the effectiveness of combining multiple models to improve performance.
Sentiment Analysis of SiKasep Application Reviews on the Play Store Using the Naïve Bayes Approach Afrahtama, Ariiq; Gunawan, Putu Harry
Building of Informatics, Technology and Science (BITS) Vol 7 No 2 (2025): September 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i2.7777

Abstract

The Ministry of Public Works and Public Housing (PUPR) launched the SiKasep application (Subsidized Housing Mortgage Information System) to streamline subsidized housing loan applications. This research analyzes user sentiment toward SiKasep using 3,416 Google Play Store reviews through Naïve Bayes classification to provide actionable insights for government digital service improvement. The methodology encompasses data scraping, comprehensive preprocessing addressing Indonesian language challenges (slang normalization and morphological complexity), TF-IDF feature extraction, and Complement Naïve Bayes classification with hyperparameter optimization. The preprocessing pipeline reduced vocabulary sparsity by 47%, while RandomOverSampler addressed significant class imbalance. The Complement Naïve Bayes classifier achieved 75.98% accuracy with balanced performance across sentiment classes (precision: 79%, recall: 76%, F1-score: 76%). Analysis revealed predominantly negative sentiment (52.4%), primarily related to registration and authentication difficulties, including document verification, login functionality, and KTP integration issues. Positive sentiment highlighted user appreciation for core housing services when technical barriers were absent. The findings emphasize the importance of streamlined registration processes and robust technical infrastructure for government digital services. This research contributes to understanding Indonesian e-government user experiences and provides a replicable sentiment analysis framework supporting evidence-based policy development for enhanced digital service delivery.
Pengelompokkan Pola Perubahan Cuaca Menggunakan Metode K-Medoids dan Gap Statistic Julianthy, Denissya; Hadiana, Asep Id; Ramadhan, Edvin
Building of Informatics, Technology and Science (BITS) Vol 7 No 2 (2025): September 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i2.7824

Abstract

Clustering daily weather patterns is an important process for understanding complex weather variations. However, commonly used methods such as K-Means have weaknesses due to their sensitivity to outliers and the need for manual clustering. This study proposes a combination of the K-Medoids and Gap Statistics methods to produce more stable and accurate clusters. Semarang's daily weather data from 2017 to 2023 was processed through cleaning, standardization with Standard Scaler, and dimensionality reduction using PCA. The Gap Statistics results indicate the optimal number of clusters is three: rainy, sunny, and cloudy. The clustering evaluation yielded a Silhouette Score of 0.3793, a Calinski-Harabasz Index of 1490.5604, and a Davies-Bouldin Index of 0.9031. These results indicate a fairly good cluster structure, although there is still room for improvement, especially in the separation between clusters.
Comparative Analysis of Random Forest and Convolutional Neural Network (CNN) Algorithms for Pneumonia Detection in Chest X-ray Images: Accuracy, Interpretability, and Computational Efficiency Zaena, Siffa; Lhaksmana, Kemas Muslim
Building of Informatics, Technology and Science (BITS) Vol 7 No 2 (2025): September 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i2.7840

Abstract

Pneumonia is a lung infection that can be detected through chest X-ray images. Manual diagnosis requires radiological expertise and time, thus an accurate automated method is needed. This study aims to compare the performance of two image classification algorithms, Convolutional Neural Network (CNN) and Random Forest (RF), in detecting pneumonia. The dataset used was obtained from Kaggle, consisting of 5,863 X-ray images categorized into three classes: bacterial pneumonia, viral pneumonia, and normal. Preprocessing steps include image resizing, normalization, and data augmentation. The CNN model was built using multiple convolutional and pooling layers, while RF utilized numerical features derived from histograms and texture. The CNN model demonstrated superior performance, achieving 92.4% accuracy, 93.1% precision, 91.6% recall, and 92.3% F1-score, compared to 82.7%, 80.3%, 85.1%, and 82.6% for Random Forest, respectively. Although CNN offers better accuracy, RF excels in interpretability. In conclusion, CNN is more effective for image-based pneumonia classification, yet RF remains relevant in applications requiring transparent decision-making. Potential biases, such as class imbalance and limited demographic representation in the dataset, could influence model performance and generalizability across different patient populations.
Comparison of Convolutional Neural Network and Support Vector Machine for Student Question Classification in ChatGPT-based Learning Tools Jordan, Brilliant; Lhaksmana, Kemas Muslim
Building of Informatics, Technology and Science (BITS) Vol 7 No 2 (2025): September 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i2.7841

Abstract

Artificial Intelligence (AI) has revolutionized educational tools by enabling systems that proactively understand and respond to student needs. ChatGPT, a widely used generative model for education in Indonesia. However, it struggles to classify student questions accurately due to ambiguous phrasing, overlapping sentence structures, and difficulty recognizing intent, which limits its effectiveness as a learning assistant. This study compares the performance of Convolutional Neural Networks (CNN), which extract locally important features from word sequences with Support Vector Machines (SVM) in classifying student questions known for handling high-dimensional data and efficiently finding the optimal hyperplane for text classification. A dataset of 2,797 Indonesian ChatGPT interactions (71% clear vs. 29% unclear) was preprocessed through case folding, stop-word removal, stemming, and tokenisation, followed by data augmentation based on synonyms, which was applied to the minority class to balance the dataset. The models were tuned through grid or random search with prediction testing of the best model using 5-fold cross-validation comparisons across three data splits (70:30, 80:20, and 90:10). Results showed that CNN achieved balanced accuracy, precision, recall, and F1-score of 0.90 on the 90:10 split, outperforming SVM, which plateaued at 0.85 accuracy and dropped to 0.76 in F1-score. The embedded filters of the CNN found generality from lexical variation through the process of augmentation, while the TF-IDF sparse vectors in the SVM failed to maintain this level of semantics. These findings underscore that CNN is more adaptive to diverse data and better suited for integration into ChatGPT-based educational tools, particularly in supporting reliable classification and personalised AI feedback in student learning contexts.
Penerapan Penyeimbangan Data Pada Analisis Sentimen Ulasan Game Magic Chess Go Go di Play Store dengan Naive Bayes Mustaqim, Muhammad Hafizd; Santoso, Angga Bayu
Building of Informatics, Technology and Science (BITS) Vol 7 No 2 (2025): September 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i2.7845

Abstract

This study aims to perform sentiment analysis on reviews of the Magic Chess Go Go game from the Google Play Store, which exhibits data imbalance with 2,949 negative sentiment entries and 1,537 positive ones. To address this issue, a sentiment classification model was developed using the Naïve Bayes algorithm, along with a comparison of four data balancing methods: SMOTE, ADASYN, Random Oversampling (ROS), and Random Undersampling (RUS). Evaluation was conducted using a confusion matrix under two data splitting schemes, with the 80:20 split yielding the best performance. In this scheme, SMOTE achieved the highest accuracy at 84.2%, followed by ADASYN (83.8%), ROS (82.9%), and RUS (77.9%). These results indicate that SMOTE is the most effective method for handling data imbalance in this context. It can be concluded that applying SMOTE to the Naïve Bayes model in the 80:20 split scenario provides the best performance, demonstrating that synthetic data generation through SMOTE helps balance the dataset without significant information loss. Future work may explore alternative algorithms and parameter tuning to enhance sentiment classification performance.
Analisa Sentimen Pengguna Aplikasi DANA Pada Ulasan Google Play Store Menggunakan Algoritma Naive Bayes Classifier dan K-Nearest Neighbors Sabillah, Dian Ayu; Afdal, M; Permana, Inggih; Muttakin, Fitriani
Building of Informatics, Technology and Science (BITS) Vol 7 No 2 (2025): September 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i2.7861

Abstract

The use of digital wallets such as DANA in Indonesia continues to increase along with the need for fast and practical non-cash transactions. User reviews on the Google Play Store are an important source of information to assess satisfaction and service problems. This study aims to classify user sentiment towards the DANA application using the Naïve Bayes Classifier (NBC) and K-Nearest Neighbor (KNN) algorithms. A total of 1,000 reviews were collected and processed through text cleaning, tokenization, stopword removal, and stemming. Sentiments were classified into positive, neutral, and negative using the lexicon method and expert validation. The results showed that NBC was superior to KNN, with the highest accuracy of 71.83%, while KNN only reached 56.44%. NBC was also more effective in detecting negative sentiment, although both were less than optimal for neutral sentiment. Word cloud visualization displays the dominant words in each sentiment category. The conclusion of this study states that Naïve Bayes is more effective in analyzing sentiment reviews of digital wallet applications such as DANA.
Fake News Detection with Hybrid CNN-SVM on Data AI and Technology Lesmana, Aditya; Sibaroni, Yuliant; Prasetyowati, Sri Suryani
Building of Informatics, Technology and Science (BITS) Vol 7 No 2 (2025): September 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i2.7871

Abstract

The spread of fake news or hoaxes in this digital era, especially related to the issue of intelligence (AI) and Technology, is increasingly unsettling because it can trigger public misunderstanding and reduce trust in technological developments. News such as the claim that AI will lead to mass unemployment is a clear example of the spread of misleading information. Therefore, a system that can accurately detect fake news is needed. The purpose of this research is to develop a fake news detection system that is able to accurately identify hoaxes on topics related to AI and Technology. This study proposes a hybrid deep learning method that combines Convolutional Neural Network (CNN) and Support Vector Machine to improve the accuracy of hoax news detection. CNN is used to extract complex news text features, whereas SVM is used as a classifier because of its advantage of being able to separate classes within optimal margins. The selection of this method is based on the results of previous research which shows that each method has good performance, but has certain limitations. By combining the two, it is hoped that more optimal results can be obtained in detecting fake news, especially the topic of AI and Technology. The evaluation was carried out using news datasets related to AI and Technology that have gone through a process of preprocessing, feature extraction with TF – IDF, and feature expansion using Glove Embedding. The results obtained showed that the CNN-SVM hybrid model provided increased accuracy compared to using a single method.
Pengembangan Algoritma Convolutional Neural Network dalam Menganalisis Emosi Suara Menggunakan Mel-Spektogram Zakka, Iqlima Sabila; Rakhman, Abdul; Lindawati, Lindawati
Building of Informatics, Technology and Science (BITS) Vol 7 No 2 (2025): September 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i2.7875

Abstract

Speech Emotion Recognition (SER) still faces challenges in accuracy, especially in distinguishing acoustically similar emotions. Conventional approaches such as MFCC (Mel Frequency Cepstral Coefficients) are often ineffective in capturing the emotional nuances of voice. To address this, this study aims to develop a Convolution Neural Network (CNN) model based on the Spec-ResNet architecture that uses Mel-Spectrogram as input to improve the system's ability to extract and recognize emotional signatures from speech signals. Another objective is to evaluate the performance of primary emotion classification in the RAVDESS dataset and measure model consistency through 5-fold cross-validation. The model used, Spec-ResNet, is an adaptation of the ResNet architecture equipped with residual learning to maximize the multi-stage feature extraction process. Experiments were conducted with the RAVDESS dataset containing 1,440 voice samples from six primary emotions: neutral, happy, sad, angry, afraid, and surprised. The test results showed a significant increase in accuracy, with a macro score reaching 92%, up from the MLP/SVM baseline of 83%. Neutral and happy emotions were classified very well (F1-scores of 93% and 90%), but emotions such as fear and surprise remained difficult to distinguish due to the similarity of their vocal patterns. Validation through 5-fold cross-validation yielded an average accuracy of 91.5% ± 0.8%. This study demonstrates the great potential of Mel-spectrograms in SER, while also underscoring the need for advanced approaches such as attention mechanisms to handle ambiguous emotions.
Sistem Identifikasi Cerdas: Integrasi IOT dengan YOLOv8 Untuk Identifikasi Visual Kerusakan Dinding Bangunan Kamdan, Kamdan; Somantri, Somantri; Rohmat, Satria Rizki; Gumelar, Agung; Kharisma, Ivana Lucia
Building of Informatics, Technology and Science (BITS) Vol 7 No 2 (2025): September 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i2.7882

Abstract

Damage to non-structural building elements, particularly walls, can serve as an early indicator of more serious structural issues. Manual crack identification is often time-consuming, subjective, and lacks consistency. This study develops an automated identification system based on computer vision using the YOLOv8 architecture, integrated with Internet of Things (IoT) technology through the ESP32-CAM device. The system is designed to visually detect and classify wall damage into light, moderate, or severe categories based on field-captured images. The model was trained and evaluated using the confusion matrix metric to assess its classification performance. The test results show that the system achieved a solid performance with an mAP@50 score of 0.822 and a stricter mAP@50-95 score of 0.522, indicating the system’s strong capability in detecting damage objects with a good level of precision. The implementation of this system is expected to support building inspection processes in a more standardized, objective, and sustainable manner, and assist in decision-making regarding building maintenance and repair.