Article Per Year (5 Year)
p-Index From 2021 - 2026
5.511
P-Index
This Author published in this journals
All Journal Techno.Com: Jurnal Teknologi Informasi Scan : Jurnal Teknologi Informasi dan Komunikasi Proceeding of the Electrical Engineering Computer Science and Informatics Journal of Information System JIPI (Jurnal Ilmiah Penelitian dan Pembelajaran Informatika) Jurnal Mantik bit-Tech JATI (Jurnal Mahasiswa Teknik Informatika) Jurnal Abdi Insani Indonesian Journal of Electrical Engineering and Computer Science International Journal of Advances in Data and Information Systems Jurnal Teknologi dan Manajemen Reputasi: Jurnal Rekayasa Perangkat Lunak International Journal Of Computer, Network Security and Information System (IJCONSIST) COMSERVA: Jurnal Penelitian dan Pengabdian Masyarakat Jurnal Informatika Dan Tekonologi Komputer (JITEK) Jurnal Informatika Teknologi dan Sains (Jinteks) ILTEK : Jurnal Teknologi Jurnal Informatika Dan Tekonologi Komputer
Nurlaili, Afina Lina
Program Studi Informatika, UPN Veteran Jawa Timur
Agriculture, Biological Sciences & Forestry Humanities Civil Engineering, Building, Construction & Architecture Computer Science & IT Control & Systems Engineering Economics, Econometrics & Finance Education Electrical & Electronics Engineering Engineering Environmental Science Health Professions Industrial & Manufacturing Engineering Languange, Linguistic, Communication & Media Law, Crime, Criminology & Criminal Justice Mathematics Mechanical Engineering Social Sciences Other

Published : 44 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 44 Documents
Search

 1   2   3   4   5 
Random Forest – Deep Convolutional Neural Network Ensemble Model for Skin Disease Classification Ananda Rheza Kurniawan; Yisti Vita Via; Afina Lina Nurlaili
bit-Tech Vol. 8 No. 1 (2025): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i1.2528

Abstract

Skin diseases such as psoriasis, atopic dermatitis, and tinea are chronic conditions that significantly affect quality of life and require rapid and accurate classification to support early treatment. However, limited medical personnel and inadequate classification tools in various regions remain major challenges in handling these cases. This study proposes an automatic skin disease classification system based on digital images using an ensemble method that combines Deep Convolutional Neural Network (DCNN) and Random Forest (RF). The dataset used comprises 4,246 images categorized into four classes (psoriasis, atopic dermatitis, tinea, and normal skin), sourced from Kaggle and DermNet. Preprocessing steps include image resizing, normalization, and data augmentation, while hyperparameter tuning is conducted using Bayesian Optimization. The ensemble model applies a soft voting mechanism to integrate predictions from both DCNN and RF. Experimental results show that the RF-DCNN model achieves an accuracy of up to 84.35% in the 80:10:10 data split scenario, surpassing the performance of the conventional CNN model. These results suggest that the hybrid DCNN-RF approach enhances accuracy, stability, and generalization in skin disease classification. The proposed model holds strong potential for implementation in artificial intelligence-based clinical decision support systems, especially in regions with limited access to dermatology specialists. Future work is encouraged to explore more advanced architectures such as EfficientNet and Swin Transformer for further performance improvements.
Design and Development of a Counseling Service System Using Extreme Programming Methodology Ikhsan Nobrian; Afina Lina Nurlaili; Firza Prima Aditiawan
bit-Tech Vol. 8 No. 1 (2025): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i1.2928

Abstract

This study addresses the inefficiency and error-prone nature of manual counseling and student violation point recording processes in schools, which often result in delays and inaccuracies. To overcome these challenges, we propose the development of a digital guidance and counseling service system designed to improve data management and enhance service accessibility for school administrators and counselors. The innovation lies in the creation of an integrated, browser-accessible application built using the MERN (MongoDB, Express.js, React, Node.js) stack, which ensures robust functionality and scalability. By applying modern development and testing methodologies, the system is designed to be both reliable and user-friendly. The core objective of this system is to streamline processes such as counseling appointment scheduling, alumni tracking, certificate submission, and student behavior reporting. It was developed using the Extreme Programming (XP) methodology, which encourages flexibility and iterative planning through close collaboration with end users. White Box Testing techniques, including cyclomatic complexity analysis and independent path testing, were employed to validate the system's internal logic. The system’s usability was assessed using the System Usability Scale (SUS), achieving an excellent score of 93.25, indicating high user satisfaction. Furthermore, the Lighthouse performance test yielded a perfect score of 100, confirming the system's high responsiveness. These results demonstrate that the developed system significantly enhances the efficiency, accuracy, and accessibility of guidance services, reduces administrative burdens, and enables better monitoring of student development, making it ideal for deployment in real-world school environments.
Design and Implementation of Digital Medical Record System Using the Boyer–Moore Algorithm Fidela Carissa Aramintha; Afina Lina Nurlaili; Firza Prima Aditiawan
bit-Tech Vol. 8 No. 2 (2025): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i2.2959

Abstract

Praktik Mandiri Bidan (PMB) Siti Nur Aini, A.Md.Keb., a primary healthcare facility specializing in maternal and child health, continues to rely on manual, paper-based medical records. This practice causes slow data retrieval, high risk of data loss, and storage limitations. To address these issues, this study developed BIDIG (Bidan Medis Digital), a web-based medical record system integrating the Boyer–Moore string matching algorithm to accelerate and enhance the accuracy of patient record searches. The Boyer–Moore algorithm was selected for its proven efficiency in pattern matching, enabling larger shift steps during mismatches and outperforming other methods such as Rabin–Karp, Brute Force, and Knuth–Morris–Pratt in prior research. BIDIG was implemented using the Laravel framework and MySQL database, with features for patient registration, medical record management, and reporting. Usability testing with the System Usability Scale (SUS) involved 10 respondents and achieved an average score of 87.25 (category: Acceptable, grade B, adjective rating: Excellent). Comparative performance evaluation using a Confusion Matrix on 160 records yielded accuracy of 0.9977, precision of 0.9318, and recall of 1.00, confirming near-perfect retrieval capability with minimal false positives. These results demonstrate that integrating an efficient search algorithm with an intuitive interface can substantially improve service quality and operational efficiency in small-scale healthcare facilities. The findings underscore the potential of algorithm-driven digitalization to overcome resource constraints, reduce errors, and accelerate access to critical health information offering a scalable model for broader adoption in similar clinical settings.
Implementation of GRU with Attention Mechanism for Classifying Lung Diseases from Respiratory Sounds Kartika Sari; Anggraini Puspita Sari; Afina Lina Nurlaili
bit-Tech Vol. 8 No. 2 (2025): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i2.3210

Abstract

Early and accurate detection of lung diseases plays a crucial role in improving treatment outcomes and reducing mortality rates, particularly in low-resource healthcare settings. Conventional auscultation using a stethoscope is a fundamental, fast, and affordable method for initial lung examination. However, its effectiveness is limited by subjectivity, as it depends on the examiner’s expertise and can be influenced by environmental noise. To overcome these limitations, this study proposes a deep learning approach for lung diseases classification using a combination of Gated Recurrent Unit (GRU) and Attention Mechanism with log Mel spectrogram as an input based on respiratory sound. Unlike previous works that employed standalone methods such as GRU or CNN, the integration of Attention mechanism in this study allows the model to focus on prominent temporal patterns within respiratory sounds, thereby enhancing classification accuracy. Experiments were conducted on the ICBHI 2017 dataset, which underwent preprocessing stages consisting of minor class removal, recording location restriction, data augmentation, and log Mel spectrogram feature extraction. The test results show that the model produces high performances with an accuracy of 90.85%, precision of 93%, recall of 90.85%, and an F1-score of 91.14%, outperforming several works that reported in prior studies. These results demonstrate the effectiveness of combining GRU and Attention mechanism in capturing the temporal features of respiratory signals. Future research could focus on enhancing model robustness through improved data quality, other model architecture, and multimodal integration for broader clinical applicability.
Implementation of the Weighted Product Method for Determining Poor Households Alya Izzah Zalfa Rihadah Ramadhani Nirwana Putri; Afina Lina Nurlaili; Muhammad Muharrom Al Haromainy
bit-Tech Vol. 8 No. 2 (2025): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i2.3214

Abstract

Many village-level poverty programs still depend on manual deliberation, which is slow to audit and difficult to reproduce across localities. This study addresses that gap by delivering an end-to-end, transparent implementation of the Weighted Product (WP) method for ranking poor households in Prunggahan Kulon, Tuban Regency. We assess whether a clearly specified WP pipeline complete with documented polarity (benefit/cost), normalized weights, and run logs can convert heterogeneous village records into reproducible preferences suitable for operational targeting. Household data supplied by the village and the Social Office were coded on a 0–1 scale for eight agreed criteria; expenditure (C2) was treated as a cost while others were benefits. Equal weights were used in this initial deployment for clarity and explainability. The method was implemented in a Laravel-based system that records bases, signed exponents, the multiplicative score , and normalized preferences . A five-household subset (A1–A5) is reported for illustration, with the full system supporting larger lists. The computation yielded a clear ordering (A4 > A1 > A2 > A3 > A5). The multiplicative rule preserved penalties for critical shortfalls and prevented strong indicators from masking severe deprivations, while the software artifacts ensured traceability from inputs to final . The dataset comprised 491 households encoded across eight criteria, with one cost criterion and seven benefits. Compared with prior WP applications, our contribution is an end-to-end, district-ready pipeline with explicit polarity, documented weights, and preserved run logs enabling third-party replication. This design measurably improves transparency and reproducibility for local poverty targeting.
Indonesian Sign Language (SIBI) Recognition from Audio Mel-Spectrograms Using LSTM Architecture Enryco Hidayat; Mohammad Idhom; Afina Lina Nurlaili
bit-Tech Vol. 8 No. 2 (2025): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i2.3229

Abstract

Persistent communication barriers continue to challenge Deaf and Hard of Hearing (DHH) individuals in accessing spoken language, underscoring the need for effective and inclusive translation technologies. Existing audio-to-sign language systems typically employ multi-stage pipelines involving speech-to-text transcription, which may propagate recognition errors and fail to preserve acoustic nuances. Addressing these limitations, this study developed and evaluated a deep learning framework for translating spoken Indonesian audio directly into classifications of the Indonesian Sign Language System (SIBI), eliminating explicit text conversion. The dataset comprised 495 eight-second WAV recordings (22,050 Hz) representing five SIBI phrase classes, augmented through time stretching, pitch shifting, and noise addition to improve generalization. Mel-Spectrogram features were extracted and input to a stacked Long Short-Term Memory (LSTM) network implemented in TensorFlow/Keras, trained to learn temporal–spectral mappings between audio patterns and SIBI categories. Evaluation on a held-out test set demonstrated robust performance, achieving 98 % accuracy with consistently high precision, recall, and F1-scores. The trained model was further integrated into a prototype web application built with Flask and React, confirming its feasibility for real-time assistive communication. While results highlight the viability of direct Mel-Spectrogram-to-LSTM translation for SIBI recognition, current findings are constrained by the limited dataset size and restricted speaker diversity. Future research should therefore expand the dataset to include more speakers, varied acoustic environments, and continuous-speech inputs to ensure broader applicability and real-world robustness.
Comparative Analysis of IndoBERT, IndoBERTweet, and XLM-RoBERTa for Detecting Online Gambling Comments on YouTube Kevin Iansyah; Afina Lina Nurlaili; Muhammad Muharrom Al Haromainy
bit-Tech Vol. 8 No. 2 (2025): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i2.3257

Abstract

The proliferation of online gambling promotions in YouTube comment sections poses significant challenges for content moderation on Indonesian digital platforms. Although transformer models have proven effective for various Indonesian-language NLP tasks, no systematic comparative evaluation exists for detecting online gambling promotions on YouTube, nor has research explored model sensitivity to hyperparameters in this context. This research identifies the optimal transformer model and configuration for detecting Indonesian-language online gambling promotion comments on YouTube. A total of 26,455 YouTube comments were collected from February to July 2025 and stratified into balanced training (18,926 comments) and validation sets (3,340 comments), plus an imbalanced testing set (4,189 comments with 28.05% promotions and 71.95% non-promotions) reflecting realistic platform conditions. Nine fine-tuning experiments were conducted with three transformer models (IndoBERT, IndoBERTweet, XLM-RoBERTa) using three learning rates (1e-5, 2e-5, 3e-5). Evaluation employed accuracy, precision, recall, F1-score, and AUC-ROC metrics. Results show IndoBERT with learning rate 1e-5 achieved best performance (F1-score 99.57%, recall 99.49%), outperforming IndoBERTweet (F1-score 98.58%) and XLM-RoBERTa (F1-score 99.28%). Interestingly, the formal corpus-trained model (IndoBERT) proved more effective than the social media model (IndoBERTweet), indicating that gambling promotion language patterns tend to be structured despite appearing in informal contexts. IndoBERT demonstrated greatest stability to learning rate variations (standard deviation 0.0011), while XLM-RoBERTa offered fastest inference time (2.48 ms) with optimal performance-efficiency balance. These findings provide practical recommendations for automated content moderation systems on Indonesian social media platforms, with IndoBERT for maximum accuracy scenarios and XLM-RoBERTa for large-scale real-time deployment.
Optimization of Ride Routes in a Tourist Attraction Using Dijkstra’s and Genetic Algorithm Firyal Wishal Nabili; Eva Yulia Puspaningrum; Afina Lina Nurlaili
bit-Tech Vol. 8 No. 2 (2025): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i2.3264

Abstract

This research presents a hybrid optimization framework that integrates Dijkstra’s Algorithm, the Genetic Algorithm (GA), and a 2-Opt local search procedure to generate optimal and demographically tailored tourist routes at Wisata Bahari Lamongan (WBL). The methodological novelty lies in the layered design of the hybrid pipeline: Dijkstra is used as a pre-processing stage to reconstruct a complete shortest-path distance matrix from partially measured field data, ensuring that GA operates on accurate inter-attraction distances and avoids unrealistic transitions. The GA then performs route evolution using PMX crossover, swap mutation, and elitism, while 2-Opt refines local segments to prevent suboptimal edge structures. Experiments involved 12 parameter-testing scenarios (CR = 0.7–0.9, MR = 0.05–0.1, population sizes of 50 and 100) across three visitor categories children, adults, and seniors. Benchmark validation on ATSP datasets from TSPLIB (BR17, P43, RY48, FT53) resulted in a mean error rate of 6.189%, confirming the robustness and generalizability of the method. The optimal configuration (CR = 0.7, MR = 0.05, PopSize = 100) produced route distances of 184,750 cm (children), 197,340 cm (adults), and 180,190 cm (seniors), yielding efficiency improvements of 30–50% compared to a pure GA and 3–7% compared to the initial measured paths. These findings demonstrate that the proposed hybrid Dijkstra–GA–2Opt framework offers a conceptually distinct, scalable, and empirically validated approach for real-world tourism route optimization.
Comparative Analysis of Memory and Render Performance: BLoC vs Provider on Low-End Devices Muhammad Albert Nur Agathon; Afina Lina Nurlaili; Muhammad Muharrom Al Haromainy
bit-Tech Vol. 8 No. 2 (2025): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i2.3268

Abstract

Mobile application performance is a critical determinant of user retention, yet optimization strategies for low-end hardware remain underexplored in the Flutter ecosystem. The choice of state management, specifically between Provider and BLoC, is a pivotal architectural decision affecting resource efficiency, particularly when integrated with Clean Architecture. Addressing the scarcity of empirical studies on constrained hardware, this research quantitatively compares the performance of these two libraries on a Vivo 1719 device running Android 7. Two identical Al-Qur'an applications were developed to facilitate a controlled experiment, isolating state management as the single variable. Performance metrics, including Resident Set Size (RSS), Garbage Collection (GC) frequency, and Frame Janks, were measured using Flutter DevTools during intensive scrolling and data loading scenarios. The results demonstrate that BLoC significantly outperforms Provider on low-end specifications. In the heaviest scenario, BLoC recorded lower peak memory usage (201.88 MB) compared to Provider (221.64 MB) and triggered 33% fewer GC events. Furthermore, BLoC reduced frame janks by 40% (15 janks vs. 21 janks). From a software engineering perspective, these findings indicate that BLoC's stream-based, event-driven architecture offers superior resource isolation compared to Provider's listener propagation mechanism, which tends to induce higher garbage collection overhead. Consequently, this study recommends BLoC as the preferred strategy for deployments targeting emerging markets, offering a worthwhile trade-off between development complexity and runtime stability to ensure broader digital inclusivity.
Feature Augmentation with XGBoost to Improve 1D CNN Performance in Anemia Recognition Raissa Atha Febrianti; Anggraini Puspita Sari; Afina Lina Nurlaili
bit-Tech Vol. 8 No. 2 (2025): bit-Tech
Publisher : Komunitas Dosen Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32877/bt.v8i2.3282

Abstract

Anemia is one of the most prevalent nutritional and hematological disorders worldwide, characterized by low hemoglobin levels caused by iron deficiency, genetic factors, or chronic diseases. Diagnosis commonly relies on Complete Blood Count (CBC) interpretation, a manual process that is time-consuming and susceptible to human error. This study proposes a novel hybrid framework that integrates Extreme Gradient Boosting (XGBoost) and a One-Dimensional Convolutional Neural Network (1D-CNN) to enhance anemia classification. The methodological novelty lies in employing XGBoost as a feature-augmentation mechanism, where its class-probability outputs are fused with the original CBC features before being processed by the 1D-CNN, enabling richer representation learning compared to conventional single-model approaches. The model was trained and evaluated using a CBC dataset consisting of 364 samples covering four anemia classes (normocytic, microcytic, macrocytic, and normal), with performance assessed through an 80:20 stratified train–test split. Experimental results demonstrate that the proposed XGB–1DCNN model achieves a testing accuracy of 97.26%, precision of 98.68%, recall of 96.46%, and F1-score of 97.48%, outperforming the baseline 1D-CNN model (83.56%). These findings demonstrate that combining ensemble learning and deep learning significantly improves the model’s ability to capture complex nonlinear patterns in CBC data, offering a more reliable solution for AI-based early anemia diagnosis and clinical decision support.
Co-Authors Achmad Andrian Maulana Adinda Putri Budi Saraswati Agung Mustika Riski Ajeng Listya Devani Alya Izzah Zalfa Rihadah Ramadhani Nirwana Putri Ananda Rheza Kurniawan Anggraini Puspita Sari Anggraini Puspita Sari Ani Dijah Rahajoe Anya Ningrum Nur'afifah Aulia Saharani Basuki Rahmat Budi Mukhamad Mulyo Dhian Satria Yudha Kartika Egar Firmansyah Enryco Hidayat Erlin Widyastuti Erwin Erdiyanto Eva Yulia Puspaningrum Faisal Muttaqin Faishal Fernando Hutama Fetty Tri Anggraeny Fidela Carissa Aramintha Firlie Aurellia Az-zahra Firyal Wishal Nabili Firza Prima Aditiawan Firza Prima Aditiawan Henni Endah Wahanani I Dewa Gde Satria Pramana Erlangga Ikhsan Nobrian Indah Rahmawti Utami Isfa Fadil Muhammad Kartika Sari Kesya Nursyahada Kevin Iansyah Luthfiyana Mahrurin Abadi M. Muharrom Al Haromainy M. Ryan Nurdiansyah N.A Made Hanindia Prami Swari Maulana, Hendra Moh. Mario Subagio Mohammad Habim Hazidan Rifqi Mohammad Idhom Mohammad Idhom Muhamad Fihris Aldama Muhammad Albert Nur Agathon Muhammad Ariq Hawari Adiputra Muhammad Faizul Ulum Muhammad Muharrom Al Haromainy Muhammad Muharrom Al Haromainy Muhammad Muharrom Al Haromainy Muhammad Rohman Irsyadi Muhsin Mulyani Satya Bhakti Nadia Dita Salsabila Nanda Syarla Hariyanti Nurul Hidajati Raissa Atha Febrianti Retno Mumpuni Retno Mumpuni Riyanarto Sarno Rizki, Agung Mustika Rizky Amelia Ryan Eka Wiratna Ryan Purnomo Safitri, Eristya Maya Silvia Dwi Cahyani Sugiarto, Sugiarto Theressa Marry Christianty Volem Alvaro Azira Azira yisti vita via Yisti Vita Via