Article Per Year (5 Year)
p-Index From 2021 - 2026
0.817
P-Index
This Author published in this journals
All Journal Jurnal Teknologi Informasi dan Ilmu Komputer JIEET (Journal of Information Engineering and Educational Technology) Journal of Informatics and Computer Science (JINACS) Indonesian Journal of Electronics, Electromedical Engineering, and Medical Informatics
Nafisah, Nurun
Unknown Affiliation
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Electrical & Electronics Engineering Engineering Health Professions Materials Science & Nanotechnology

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

Found 4 Documents
Search

 1 
Pengembangan Website Pergudangan berbasis FIFO untuk Optimalisasi Persediaan Barang di LMI Pusat Surabaya Nafisah, Nurun; Yamasari, Yuni
Journal of Informatics and Computer Science (JINACS) Vol. 5 No. 04 (2024)
Publisher : Universitas Negeri Surabaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26740/jinacs.v5n04.p635-645

Abstract

Hybrid Autoencoder Architectures with LSTM and GRU Layers for Bitcoin Price Prediction Nafisah, Nurun; Yamasari, Yuni; Yohannes, Ervin
Indonesian Journal of Electronics, Electromedical Engineering, and Medical Informatics Vol. 7 No. 4 (2025): November
Publisher : Jurusan Teknik Elektromedik, Politeknik Kesehatan Kemenkes Surabaya, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35882/ijeeemi.v7i4.132

Abstract

The high volatility of cryptocurrency markets, particularly Bitcoin, poses significant challenges for accurate price forecasting. To address this issue, this study evaluates the performance of four autoencoder-based deep learning architectures: AE-LSTM, AE-GRU, AE-LSTM-GRU, and AE-GRU-LSTM. The models were developed and tested using a univariate approach, where only the closing price was used as input, and two different window sizes (30 and 60) were applied to analyse the effect of historical sequence length on prediction accuracy. Several parameter configurations, including the number of epochs, dropout rate, and learning rate, were explored to determine the optimal model performance. The dataset comprises Bitcoin’s daily closing prices from 2018 to 2025, encompassing diverse market phases, including both bullish and bearish trends. Model performance was assessed using four evaluation metrics: Root Mean Square Error (RMSE), Mean Absolute Error (MAE), the coefficient of determination (R²), and Mean Absolute Percentage Error (MAPE). The results indicate that the AE-LSTM-GRU consistently achieved the best overall performance across all configurations. For a window size of 30, it achieved an RMSE of 1.53067 and a MAPE of 1.98%, while for a window size of 60, the best performance recorded was an RMSE of 1.55217 and a MAPE of 2.09%. The hybrid structure combining LSTM’s capability to capture long-term dependencies with GRU’s efficiency in information decoding demonstrated strong robustness in modelling highly volatile time series. This study contributes to financial time series forecasting by presenting hybrid autoencoder architectures that strike a balance between predictive accuracy and computational efficiency, providing practical insights for researchers and practitioners in financial technology and cryptocurrency analytics
A Performance Comparison of LSTM and GRU Architectures for Forecasting Daily Bitcoin Price Volatility Nafisah, Nurun; Yamasari, Yuni; Yohannes, Ervin
JIEET (Journal of Information Engineering and Educational Technology) Vol. 9 No. 2 (2025)
Publisher : Universitas Negeri Surabaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26740/jieet.v9n2.p156-167

Abstract

The highly volatile movement of Bitcoin prices necessitates the use of prediction methods capable of accurately capturing complex and rapidly changing patterns. This study aims to compare the performance of two recurrent neural network architectures, namely Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), in forecasting Bitcoin prices based on historical time series data. The analysis was conducted using daily closing price data, with several parameter configurations applied, including dropout value, learning rate, and number of epochs at a window size of 30. The training process was carried out using a univariate approach to assess the fundamental ability of each model to learn temporal patterns without the influence of external variables. The results indicate that the GRU model consistently outperforms LSTM across most experimental settings. The best performance was achieved with 30 epochs, dropout 0.1, and a learning rate of 0.001, producing RMSE 1478.333, MAE 1000.900, R² 0.996081, and MAPE 1.973072. These metrics demonstrate a lower error level and a stronger fit to actual Bitcoin price movements. Moreover, a paired t-test confirmed that the performance gap between the two models is statistically significant. Overall, the findings suggest that the Gated Recurrent Unit architecture is more efficient in capturing nonlinear patterns and responding to the volatile dynamics of cryptocurrency price fluctuations, making it a promising approach for future predictive modeling in financial time series.
Tinjauan Literatur Tentang Event Log dan Process Mining: Implementasi, Tools, dan Algoritma Nafisah, Nurun; Yamasari, Yuni
Jurnal Teknologi Informasi dan Ilmu Komputer Vol 13 No 2: April 2026
Publisher : Fakultas Ilmu Komputer, Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.25126/jtiik.132

Abstract

Analisis event log melalui process mining telah digunakan secara luas untuk mengeksplorasi pola perilaku dalam berbagai domain, terutama pendidikan dan game. Pendekatan ini memanfaatkan alat bantu seperti Disco, ProM, dan metode berbasis machine learning untuk menggali pola perilaku, memahami proses, dan mengevaluasi hasil. Penelitian ini menggunakan metode Systematic Literature Review (SLR) dengan teknik Critical Appraisal Skills Programme (CASP) untuk memastikan kualitas dan validitas studi yang dianalisis. Dari total 48 artikel yang ditemukan, sebanyak 30 artikel lolos tahap seleksi, dan 18 artikel dinyatakan layak untuk dianalisis lebih lanjut. Dalam domain pendidikan, analisis menunjukkan perbedaan signifikan dalam pola belajar siswa, seperti perilaku kelompok dengan nilai tinggi yang cenderung konsisten dibandingkan kelompok dengan nilai rendah. Pola akses ke Learning Management System (LMS) dihubungkan dengan capaian akademik, memberikan wawasan tentang hubungan antara aktivitas belajar dan hasil pembelajaran. Di sisi lain, dalam domain game, algoritma process mining digunakan untuk mengidentifikasi kesalahan pemain, mengevaluasi jalur normatif, dan menganalisis keterkaitan antara model proses dan performa pemain. Hasilnya menunjukkan bahwa algoritma seperti Inductive Miner, Heuristic Miner, dan Fuzzy Miner memberikan fitness model tinggi (0,9–1,0) dan mampu mengekstraksi pola perilaku yang signifikan. Penggunaan tools seperti ProM dan Disco memungkinkan visualisasi model proses yang intuitif, sementara algoritma berbasis machine learning seperti XGBoost memberikan akurasi tinggi dalam memprediksi hasil berdasarkan data event log. Secara keseluruhan, hasil studi menegaskan bahwa process mining berperan penting dalam mengungkap pola perilaku kompleks serta mendukung peningkatan efektivitas pembelajaran dan evaluasi performa pengguna di berbagai domain.   Abstract Event log analysis through process mining has been widely used to explore behavioural patterns in various domains, particularly education and gaming. This approach utilizes Disco, ProM, and machine learning to examine behavioural patterns, understand processes, and evaluate results. This study employed a Systematic Literature Review (SLR) with the Critical Appraisal Skills Programme (CASP) technique to ensure the quality and validity of the analyzed studies. Of the 48 articles found, 30 passed the selection stage, and 18 were deemed suitable for further analysis. In the education domain, the analysis revealed significant differences in student learning patterns, such as the behaviour of high-scoring groups tending to be more consistent than that of low-scoring groups. Access patterns to Learning Management Systems (LMS) were linked to academic achievement, providing insight into the relationship between learning activities and learning outcomes. Process mining algorithms were used in the gaming domain to identify player errors, evaluate normative pathways, and analyze the relationship between process models and player performance. The results show that algorithms such as Inductive Miner, Heuristic Miner, and Fuzzy Miner provide high model fitness (0.9–1.0) and can extract significant behavioural patterns. Tools such as ProM and Disco allow for intuitive visualization of process models. In contrast, machine learning-based algorithms such as XGBoost provide high accuracy in predicting outcomes based on event log data. The study results confirm that process mining is crucial in uncovering complex behavioural patterns and supporting improved learning effectiveness and user performance evaluation across various domains.  
Co-Authors Ervin Yohannes YUNI YAMASARI