Claim Missing Document
Check
Articles

Found 21 Documents
Search

Implementasi Model BiLSTM-Attention untuk Prediksi Nilai IHSG Berdasarkan Data Historis OHLCV Ramadhanti, Amirah Rizky; Putri, Safira Rahmalia; Trimono; Mohammad Idhom
Jurnal Ilmiah Media Sisfo Vol 19 No 2 (2025): Jurnal Ilmiah Media Sisfo
Publisher : LPPM Universitas Dinamika Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33998/mediasisfo.2025.19.2.2392

Abstract

The Composite Stock Price Index (IHSG) reflects the performance of the Indonesian capital market, but predicting it is challenging due to high volatility and the influence of various external factors. This study aims to develop and evaluate a deep learning-based predictive model using a Bidirectional Long Short-Term Memory (BiLSTM) architecture combined with an Attention Mechanism to predict the IHSG value based on historical numerical data (OHLCV). This method was chosen for its ability to recognize bidirectional sequential patterns and highlight the most relevant historical information in the prediction process. The research was conducted quantitatively using an experimental approach, and model evaluation was performed using regression metrics such as R², RMSE, MAE, and MAPE. The results obtained showed excellent predictive performance with an R² of 0.9485, MAPE of 0.63%, RMSE of 59.47, and MAE of 45.12. Additionally, attention weight analysis revealed that the model focuses more on the last two days within the prediction time window, indicating that recent information significantly influences IHSG movements. These findings suggest that the BiLSTM-Attention approach is effective in capturing stock market dynamics and has the potential to serve as a strategic tool for data-driven investment decision-making.
Analisis sentimen program makan bergizi gratis menggunakan bidirectional gated recurrent unit Krisnawan; Zufar Abdullah Rabbani; Trimono; Mohammad Idhom
IT Explore: Jurnal Penerapan Teknologi Informasi dan Komunikasi Vol 4 No 3 (2025): IT-Explore Oktober 2025
Publisher : Fakultas Teknologi Informasi, Universitas Kristen Satya Wacana

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24246/itexplore.v4i3.2025.pp282-294

Abstract

The Free Nutritious Meals (MBG) program launched by the Indonesian government aims to address the problem of malnutrition in children and students. However, the acceptance of this program in the community still requires in-depth evaluation because there are many negative sentiments that dominate on social media. This study aims to analyze the sentiment of the Indonesian community regarding the Free Nutritious Meals program on social media X (Twitter) using the Bidirectional Gated Recurrent Unit (BiGRU) model. Of the 1,405 tweet data obtained, 57% were negative opinions and 43% were positive opinions. The evaluation results show that the BiGRU model with FastText support to handle potential overfitting, is able to classify sentiment effectively, with an accuracy of 80%. Sentiment analysis shows that the majority of public responses to the Free Nutritious Meals (MBG) program tend to be negative, with 798 negative tweets and 607 positive. This reflects public dissatisfaction with the implementation of the program and highlights the need for evaluation and improvements so that the benefits can be more widely felt by the community.
Implementasi Manajemen Bandwidth Menggunakan Metode Queue Tree dengan PCQ di SMK Negeri 1 Surabaya Fikri Dwilaksono; Henni Endah Wahanani; Mohammad Idhom
Jurnal Ilmiah Teknik Informatika dan Komunikasi Vol. 5 No. 2 (2025): Juli: Jurnal Ilmiah Teknik Informatika dan Komunikasi
Publisher : Lembaga Pengembangan Kinerja Dosen

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55606/juitik.v5i2.1029

Abstract

With the rapid advancement of information technology, maintaining a stable and efficient computer network has become essential, particularly in educational settings. SMK Negeri 1 Surabaya encounters difficulties in managing its limited bandwidth because of the large number of users accessing the network simultaneously. This research focuses on applying bandwidth management through the Queue Tree technique combined with Per Connection Queuing (PCQ) on the school's computer network using Mikrotik hardware. The approach aims to enhance network performance by improving throughput, minimizing delays, and stabilizing jitter. Evaluation of Quality of Service (QoS) indicators—including throughput, packet loss, delay, and jitter—demonstrated favorable outcomes, with all metrics categorized as good. Findings confirm that using Queue Tree and PCQ methods contributes to fairer bandwidth distribution, mitigates network congestion, and improves overall network quality in the educational environment. This study offers a practical solution to support teaching and administrative processes by providing a more reliable and efficient network infrastructure.
Implementation of Bayesian Structural Time Series (BSTS) Method for Predicting Traditional Market Revenue Achievement in Surabaya Muizzadin, Muizzadin; Mohammad Idhom; Damaliana, Aviolla Terza
Indonesian Journal of Electronics, Electromedical Engineering, and Medical Informatics Vol. 7 No. 2 (2025): May
Publisher : Jurusan Teknik Elektromedik, Politeknik Kesehatan Kemenkes Surabaya, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35882/ijeeemi.v7i2.82

Abstract

Traditional markets play an important role in the regional economy, including in the city of Surabaya. However, the number of traditional markets in Surabaya has continued to decline in recent years due to competition with modern markets. In addition, the contribution of traditional markets to Regional Original Income (PAD) has fluctuated, for example 1.67% in 2013, 1.66% in 2014, and increased to 1.76% in 2015. This condition poses a challenge for the management of regional economic policies, so an accurate prediction method is needed to support strategic decision making. This study aims to predict the achievement of traditional market revenue in Surabaya using the Bayesian Structural Time Series (BSTS) method. The data used is the percentage of traditional market revenue achievement over the past fifteen years. The BSTS model is applied with various components, including Local Level, Local Linear Trend, and Seasonal, which allows flexibility in capturing trends, seasonal patterns, and structural changes in the data. Model evaluation is carried out using Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE) to assess prediction accuracy. The results of the study showed that the BSTS model with Local Level and Seasonal components and 1,000 MCMC iterations provided the best performance, with a MAPE value of 4.036% and an RMSE of 5.198. This model is able to capture trend and seasonal patterns well, making it effective in predicting traditional market revenue achievements. Based on these findings, the BSTS method has proven to be a reliable approach in predicting traditional market revenue achievements. The results of this study are expected to help market managers and policy makers in designing more adaptive strategies to maintain the competitiveness of traditional markets and increase their contribution to the regional economy.
Implementasi Spatial Durbin Model Berbasis Data Science Untuk Analisis Kemiskinan Jawa Timur Arif, Farah Yusnaida; Mohammad Idhom; Trimono, Trimono
Seminar Nasional Teknologi dan Multidisiplin Ilmu (SEMNASTEKMU) Vol. 5 No. 1 (2025): SEMNASTEKMU
Publisher : Universitas Sains dan Teknologi Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.51903/9w9pye50

Abstract

Poverty remains a major development challenge that requires data-driven analysis to understand its variation across regions. This study focuses on East Java, where spatial interdependence is suspected to influence poverty distribution, making spatial analysis relevant for supporting regional policy design. The study examines determinants of poverty using the Spatial Durbin Model, which captures both direct effects and indirect spatial spillovers through lagged independent variables. The analytical workflow is implemented using a Python-based data science pipeline to ensure a systematic, transparent, and reproducible process, in line with current trends in technology-supported research. The dataset consists of 2024 secondary data from the Indonesian Central Bureau of Statistics. The analysis includes data preprocessing, construction of a Queen Contiguity spatial weight matrix, Moran’s I test to detect spatial autocorrelation, and SDM estimation. Results indicate significant positive spatial autocorrelation (I = 0.4099; p = 0.0008), showing that poverty is not randomly distributed. While the spatial lag of the dependent variable is not significant, an indirect spatial effect appears through the Gini Ratio (θ₄ = –39.42168; p = 0.03855). Moreover, the Human Development Index significantly reduces poverty. These findings highlight the roles of regional inequality and human development in shaping poverty dynamics and provide insights for more targeted policy interventions.
Prediksi Harga Saham Menggunakan Model Mixture Autoregressive (MAR) (Studi Kasus : Saham Perusahaan Rokok) Ristiyani, Sintiya; Mohammad Idhom; Dwi Arman Prasetya
JURNAL PETISI (Pendidikan Teknologi Informasi) Vol. 7 No. 1 (2026): JURNAL PETISI (Pendidikan Teknologi Informasi)
Publisher : Universitas Pendidikan Muhammadiyah Sorong

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36232/jurnalpetisi.v7i1.2721

Abstract

Memprediksi harga saham merupakan tantangan besar dalam dunia penelitian karena tingginya risiko yang terlibat, meskipun potensi keuntungannya juga sangat besar. Hal ini mencerminkan bahwa pergerakan saham sangat dipengaruhi oleh perubahan faktor eksternal maupun internal. Oleh karena itu, memperoleh prediksi harga saham yang tepat menjadi sangat krusial guna meminimalkan kemungkinan kerugian bagi para investor. Penelitian ini dilakukan dengan tujuan untuk memprediksi harga saham dari tiga perusahaan rokok besar di Indonesia, yaitu PT. Hanjaya Mandala Sampoerna Tbk (HMSP), PT. Gudang Garam Tbk (GGRM), dan PT Wismilak Inti Makmur Tbk (WIIM). Penelitian ini menggunakan suatu model statistika bernama Mixture Autoregressive (MAR), dipilih karena kemampuannya dalam menangkap pola tidak linier serta perubahan kondisi yang sering muncul pada pergerakan harga saham. Menggunakan model MAR diperoleh model MAR dengan nilai Mean Absolute Percentage Error (MAPE) untuk masing-masing perusahaan yaitu Gudang Garam (5.46%), Sampoerna (4.52%), dan Wismilak (3.58%).
Implementasi Komunikasi Real-Time Berbasis WebSocket pada Sistem Quickshifter Dual-Mikrokontroler Erlangga Putra Ramadhan; Mohammad Idhom; Firza Prima Aditiawan
Jurnal Teknologi dan Manajemen Industri Terapan Vol. 5 No. 1 (2026): Jurnal Teknologi dan Manajemen Industri Terapan
Publisher : Yayasan Inovasi Kemajuan Intelektual

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55826/jtmit.v5i1.1631

Abstract

Penelitian ini mengimplementasikan komunikasi real-time berbasis WebSocket pada sistem quickshifter sepeda motor dengan arsitektur dual-mikrokontroler. Arduino Nano digunakan untuk menangani proses inti quickshifter, meliputi pembacaan sensor CKP, perhitungan RPM, serta pengendalian pemutusan pengapian, sementara ESP32-C3 berfungsi sebagai gateway komunikasi nirkabel melalui Wi-Fi dan WebSocket. Kedua mikrokontroler berkomunikasi menggunakan UART dengan baudrate 115200. Evaluasi kinerja difokuskan pada parameter latensi, jitter, packet loss, serta kepraktisan penggunaan, dengan USB Serial sebagai pembanding. Hasil pengujian menunjukkan bahwa WebSocket memiliki latensi rata-rata 19,81 ms dan jitter 1,79 ms tanpa packet loss, sedikit lebih tinggi dibandingkan USB Serial namun masih dalam batas toleransi untuk aplikasi real-time. Selain itu, sistem mampu melakukan penyetelan dan pemantauan RPM secara responsif serta stabil. Implementasi ini membuktikan bahwa protokol WebSocket pada arsitektur dual-mikrokontroler mampu menyeimbangkan kebutuhan performa real-time dan fleksibilitas koneksi nirkabel pada sistem quickshifter.
Prediksi Harga Saham Menggunakan ARIMA Outlier sebagai Pendekatan Awal Menuju Analisis AI Keuangan Cindi Adam; Mohammad Idhom; Trimono Trimono
Seminar Nasional Teknologi dan Multidisiplin Ilmu (SEMNASTEKMU) Vol. 5 No. 1 (2025): SEMNASTEKMU
Publisher : Universitas Sains dan Teknologi Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.51903/zwvk1v20

Abstract

The development of artificial intelligence (AI) has driven innovation in financial analysis, including the prediction of volatile stock prices. This study aims to predict the stock price of PT Garudafood Putra Putri Jaya Tbk using an ARIMA model with Outlier handling as an initial approach towards a more adaptive prediction system. Daily closing price data from Yahoo Finance was analyzed through stationarity testing, ARIMA model identification, log-return-based Outlier detection, and performance evaluation using RMSE, MAE, and MAPE. The results show that ARIMA Outlier performs better than the basic ARIMA. The standard ARIMA produces a MAPE of 1.32% and an AIC of –899.46, while ARIMA with three dummy Outliers achieves a MAPE of 1.16% and an AIC of –900.37. The 14-day forecast shows a stable pattern in the range of Rp 370–371. In the test data, the basic ARIMA provided the best accuracy in mid-August, while ARIMA Outlier achieved the highest accuracy at the end of August with a prediction of Rp 370.2, which was very close to the actual price of Rp 370.4. These results show that handling Outliers improves the accuracy of the model, so that ARIMA Outlier can be used as a starting point for the development of an AI-based financial prediction system.
KLASTERING WILAYAH DI JAWA TIMUR BERDASARKAN FAKTOR UNMET NEED MENGGUNAKAN FUZZY GUSTAFSON-KESSEL Chrysilla Citra Windyadari; Aviolla Terza Damaliana; Mohammad Idhom
Jurnal Sistem Informasi dan Informatika (Simika) Vol. 8 No. 2 (2025): Jurnal Sistem Informasi dan Informatika (Simika)
Publisher : Program Studi Sistem Informasi, Universitas Banten Jaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47080/simika.v8i2.3976

Abstract

The Family Planning Program is an effort to control the rate of population growth by regulating desired pregnancies. In its realization, the family planning program faces challenges in the form of unmet need (couples of childbearing age who do not use contraception). East Java Province in 2023 was recorded as the province with the third highest number of unmet need cases in Java. One method that can be used to analyze the phenomenon of unmet need is clustering analysis. Clustering analysis will help identify areas in East Java based on the priority level of the family planning program. Fuzzy Gustafson-Kessel (FGK) is one of the clustering methods developed as a refinement of the Fuzzy C-Means method. This study implements the Fuzzy Gustafson-Kessel (FGK) method with and without Principal Component Analysis (PCA) to cluster regions in East Java based on unmet need and determinant factors such as the availability of family planning facilities and resources. The results showed that the best model was obtained when using FGK with PCA, with the highest FSI value of 0.668 and XB of 0.235 at configuration c = 4 and m = 3.5. The clusters formed consist of 5 medium priority areas, 12 low priority areas, 9 high priority areas, and 12 developing priority areas. The results of this clustering can be used as a basis for policy makers in designing more effective intervention strategies to address unmet need in East Java.
IMPLEMENTASI SISTEM PENGENALAN WAJAH UNTUK PENDATAAN KEHADIRAN BERBASIS PYTHON Risky Pratama, Daffa; Mohammad Idhom
Journal of Data Science Theory and Application Vol. 5 No. 1 (2026): JASTA
Publisher : LP3M Universitas Putra Bangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32639/be3krg76

Abstract

Maraknya pertemuan daring membuka peluang besar dalam penerapan sistem otomatis untuk pendataan kehadiran. Penelitian ini bertujuan mengimplementasikan sistem pengenalan wajah berbasis Python untuk mendukung proses absensi peserta video meeting secara otomatis dan efisien. Sistem dikembangkan menggunakan algoritma Convolutional Neural Network (CNN) dengan arsitektur FaceNet sebagai ekstraktor fitur wajah dan Haar Cascade sebagai metode deteksi wajah awal. Aplikasi dilengkapi antarmuka grafis berbasis pustaka customtkinter sehingga pengguna dapat mengoperasikan seluruh fungsi sistem secara interaktif tanpa perintah berbasis teks. Metode pengembangan yang digunakan adalah Software Development Life Cycle (SDLC) dengan model Incremental yang terdiri dari empat modul utama, yaitu deteksi wajah, pengenalan wajah, antarmuka pengguna, dan penyimpanan data. Pengujian dilakukan menggunakan dua tangkapan layar video meeting yang memuat total 30 wajah peserta. Hasil evaluasi menunjukkan nilai precision sebesar 100%, recall 80%, accuracy 80%, dan F1-score 88,8%. Hasil ini menunjukkan sistem mampu melakukan pendataan kehadiran secara akurat meskipun dipengaruhi oleh pose wajah dan kondisi pencahayaan.