Claim Missing Document
Check
Articles

Found 9 Documents
Search

Optimasi Deep Learning untuk Prediksi Saham di Masa Pandemi Covid-19 Hastomo, Widi; Karno, Adhitio Satyo Bayangkari; Kalbuana, Nawang; Nisfiani, Ervina; ETP, Lussiana
JEPIN (Jurnal Edukasi dan Penelitian Informatika) Vol 7, No 2 (2021): Volume 7 No 2
Publisher : Program Studi Informatika

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26418/jp.v7i2.47411

Abstract

Penelitian ini bertujuan untuk meningkatkan akurasi dengan menurunkan tingkat kesalahan prediksi dari 5 data saham blue chip di Indonesia. Dengan cara mengkombinasikan desain 4 hidden layer neural nework menggunakan Long Short Term Memory (LSTM) dan Gated Recurrent Unit (GRU). Dari tiap data saham akan dihasilkan grafik rmse-epoch yang dapat menunjukan kombinasi layer dengan akurasi terbaik, sebagai berikut; (a) BBCA dengan layer LSTM-GRU-LSTM-GRU (RMSE=1120,651, e=15), (b) BBRI dengan layer LSTM-GRU-LSTM-GRU (RMSE =110,331, e=25), (c) INDF dengan layer GRU-GRU-GRU-GRU (RMSE =156,297, e=35 ), (d) ASII dengan layer GRU-GRU-GRU-GRU (RMSE =134,551, e=20 ), (e) TLKM dengan layer GRU-LSTM-GRU-LSTM (RMSE =71,658, e=35 ). Tantangan dalam mengolah data Deep Learning (DL) adalah menentukan nilai parameter epoch untuk menghasilkan prediksi akurasi yang tinggi.
Optimizing Investment: Combining Deep Learning for Price Prediction and Moving Average for Return-Risk Analysis Hastomo, Widi; Karno, Adhitio Satyo Bayangkari; Masriyanda, Masriyanda; Sestri, Ellya; Kardian, Aqwam Rosadi; Azis, Nur; Dewanto, Ignatius Joko; Rasyiddin, Ahmad; Sundoro, Aries; Kamilia, Nada
Jurnal Teknik Elektro Vol 14, No 2 (2022): Jurnal Teknik Elektro
Publisher : Jurusan Teknik Elektro, Fakultas Teknik, Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/jte.v14i2.45002

Abstract

The ability to analyze predictions marks something going up or down, as well as the level of possible risk taken into account by much-needed stock investors. In a study, this analysis of risk and correlation between shares was calculated using the method of moving averages (MA). Besides that, a dataset of 4 stocks (Apple, Google, Microsoft, and Amazon) also performed prediction mark stock in period time next (future) with the use of the neural network method (deep learning) Long Short-Term Memory (LSTM) model. The result of programming in the Python language is several visualizations for easy graph-reading information. This article presents new research that aims to fill the gap in understanding investment analysis for beginners by visualizing risk and return analysis on shares. The results reveal that changes in stock sales volume did not occur significantly, although the short and long-term MA charts for the four stocks tended to fluctuate, offering new insights into investment analysis and providing a basis for future development. The best accuracy results were on MSFT shares, with an achievement of 0.9532 and a loss value of 0.0014. Thus, MSFT shares can be used as a priority for investment. Therefore, this research adds a new dimension to the literature and paves the way for further investigations in risk and return analysis and stock prediction using deep learning.
Classification of Brain Image Tumor using EfficientNet B1-B2 Deep Learning Hastomo, Widi; Karno, Adhitio Satyo Bayangkari; Sestri, Ellya; Terisia, Vany; Yusuf, Diana; Arman, Shevty Arbekti; Arif, Dodi
Semesta Teknika Vol 27, No 1 (2024): MEI
Publisher : Universitas Muhammadiyah Yogyakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.18196/st.v27i1.19691

Abstract

In this study, a new neural network model (EfficientNet B1-B2) was sought for the detection of brain tumors in magnetic resonance imaging (MRI) images. The primary objective was to achieve high accuracy rates so as to classify the images. The deep learning techniques meticulously processed and increased the data augmentation as much as possible for the EfficientNet B1-B2 models. Our experimental results show an accuracy of 98% in the B1 version in Table II. This provides a potentially optimistic view of the application of artificial intelligence technology to disease diagnosis based on medical image analysis. Nonetheless, we must remind ourselves that the dataset we used has limitations in terms of the challenges it can pose. Although the number of potential variations of actual medical images constitutes a major challenge, it is not the only one. Most medical datasets are unbalanced, contain highly variable noise, have a slow internal structure, and are often small in size. Hence, our end goal is to help stimulate not only the field of brain tumor detection and treatment but also the development of more sophisticated classification models in the health context.
Stacked LSTM-GRU Long-Term Forecasting Model for Indonesian Islamic Banks Sujatna, Yayat; Karno, Adhitio Satyo Bayangkari; Hastomo, Widi; Yuningsih, Nia; Arif, Dody; Handayani, Sri Setya; Kardian, Aqwam Rosadi; Wardhani, Ire Puspa; Rere, L.M Rasdi
Knowledge Engineering and Data Science Vol 6, No 2 (2023)
Publisher : Universitas Negeri Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.17977/um018v6i22023p215-250

Abstract

The development of the Islamic banking industry in Indonesia has become a significant concern in recent years, with rapid growth in the number of banks operating based on Sharia principles. To face emerging challenges and opportunities, a deep understanding of the long-term financial behavior of Islamic banks is becoming increasingly important. This study aims to predict the share price of PT Bank Syariah Indonesia Tbk, over 28 days using the LSTM-GRU stack. The observation stage includes importing the dataset, data separation, model variations, the training process, output, and evaluation. Observations were conducted using 10 model variations from 4 stacks of LSTM and GRU. Each model performs the training process in four epochs (200, 500, 750, and 1000). The results of observations in this study show that long-term predictions (28 days ahead) using four stacks of LSTM-GRU and daily training accumulation techniques produce better accuracy than the general method (using multiple outputs). From the observations we have made for predictions for the next 28 days, the model with the LGLG stack arrangement (LSTM-GRU-LSTM-GRU) produces the best accuracy at epoch 750 with an MSE LSTM-GRU 63.43762863. This study will undoubtedly continue in order to achieve even better precision, either by utilizing a new design or by further improving the technology we are now employing.
PREDICTING SOLAR POWER GENERATION: A MACHINE LEARNING APPROACH FOR GRID STABILITY AND EFFICIENCY Setiawati, Popong; Karno, Adhitio Satyo Bayangkari; Hastomo, Widi; Sestri, Ellya; Kasoni, Dian; Arif, Dodi; Razi, Fahrul
Jurnal Pilar Nusa Mandiri Vol. 21 No. 1 (2025): Pilar Nusa Mandiri : Journal of Computing and Information System Publishing Pe
Publisher : LPPM Universitas Nusa Mandiri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33480/pilar.v21i1.6126

Abstract

In countries with high levels of insolation, the demand for renewable energy sources has driven the rapid emergence and growth of solar power plants. Maintaining grid stability and efficient power management in response to weather variations that affect solar radiation intensity and battery consumption limits remains a major challenge. This study aims to develop a machine learning-based prediction model to estimate the electricity generated by solar power plants using weather data. Four algorithms are utilized: Linear Regression, Random Forest Regressor, Decision Tree Regressor, and Gradient Boosting Regressor. The results show that the Random Forest algorithm produces the best model, with MAE and RMSE values of 0.1114281 and 0.3187232, respectively. This research contributes to the literature, particularly on the relatively unexplored topic of using multiple machine learning models to predict energy output from photovoltaic systems. The findings have the potential to inform more efficient energy policies and improve energy integration technologies for grid-connected solar power systems.
Prediksi Data Time Series Saham Bank BRI Dengan Mesin Belajar LSTM (Long ShortTerm Memory) Karno, Adhitio Satyo Bayangkari
Journal of Informatic and Information Security Vol. 1 No. 1 (2020): Juni 2020
Publisher : Program Studi Informatika, Fakultas Ilmu Komputer Universitas Bhayangkara Jakarta Raya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31599/jiforty.v1i1.133

Abstract

Abstract This study aims to measure the accuracy in predicting time series data using the LSTM (Long Short-Term Memory) machine learning method, and determine the number of epochs needed to produce a small RMSE (Root Mean Square Error) value. The result of this research is a high level of variation in RMSE value to the number of epochs needed in the data processing. This variation is quite difficult to obtain the right epoch value. By doing an iteration of the LSTM process on the number of different epochs (visualized in the graph), then the number of epochs with a minimum RMSE value will be easier to obtain. From the research of BBRI's stock data prediction, a good RMSE value was obtained (RMSE = 227.470333244533). Keywords: long short-term memory, machine learning, epoch, root mean square error, mean square error. Abstrak Penelitian ini bertujuan untuk mengukur ketelitian dalam memprediksi data time series menggunakan metode mesin belajar LSTM (Long Short-Term Memory), serta menentukan banyaknya epoch yang diperlukan untuk menghasilkan nilai RMSE (Root Mean Square Error) yang kecil. Hasil dari penelitian ini adalah tingkat variasi yang tinggi nilai rmse terhdap jumlah epoch yang diperlukan dalam proses pengolahan data. Variasi ini cukup menyulitkan untuk memperoleh nilai epoch yang tepat. Dengan melakukan iterasi dari proses LSTM terhadap jumlah epoch yang berbeda (di visualisasikan dalam grafik), maka jumlah epoch dengan nilai RMSE minimal akan lebih mudah diperoleh. Dari penelitan prediksi data saham BBRI diperoleh nilai RMSE yang cukup baik yaitu 227,470333244533. Kata kunci: long short-term memory, machine learning, epoch, root mean square error, mean square error.
OPTIMASI CONVOLUTION NEURAL NETWORK UNTUK DETEKSI COVID-19 Hastomo, Widi; Karno, Adhitio Satyo Bayangkari; Bakti, Indra
RADIAL : Jurnal Peradaban Sains, Rekayasa dan Teknologi Vol. 10 No. 2 (2022): RADIAL: JuRnal PerADaban SaIns RekAyasan dan TeknoLogi
Publisher : Universitas Bina Taruna Gorontalo

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37971/radial.v10i2.299

Abstract

Abstrak: Optimasi Convolution Neural Network Untuk Deteksi Covid-19. Kondisi pandemi seperti sekarang ini diperlukan sebuah algoritma pembelajaran mesin untuk mendeteksi covid-19 secara otomatis berdasarkan pada gambar rontgen dada guna memudahkan dalam mambantu pengambil keputusan. Penelitian ini ingin membandingkan arsitektur CNN AlexNet dan MobileNetV2 untuk mendeteksi (a) covid-19, (b) lung opacity, (c) normal, (d) viral pneumonia. Data himpunan rontgen dada yang digunakan sejumlah 4000 yang berasal dari kaggle.com, 0.8 data dibagi untuk pelatihan sedangkan 0.2 nya digunakan untuk pengujian. Optimizer yang digunakan yaitu keras SGD momentum, dengan nilai learning rate 0.005 dan momentum 0.9, serta epoch 50. Ukuran gambar untuk input yaitu 224x224 serta ukuran batch 32. Hasil optimasi dari kedua algoritma tersebut yaitu, MobileNetV2 lebih baik untuk mendeteksi covid-19 dengan nilai akurasi presisi mencapai 99%. Penelitian selanjutnya dapat membandingkan algoritma CNN yang lainnya serta data himpunan yang lebih banyak. Kata kunci: CNN; AlexNet; MobileNetV2; Covid-19 Abstract: Convolution Neural Network Optimization for Covid-19 Detection. In the current pandemic conditions, a machine learning algorithm is needed to detect COVID-19 automatically based on chest X-ray images to make it easier to assist decision makers. Aim study be disposed for compare the architecture of CNN AlexNet and MobileNetV2 to detect (a) covid-19, (b) lung opacity, (c) normal, (d) viral pneumonia. The data set of chest X-rays used are 4000 from kaggle.com, 0.8 of the data is shared for training while 0.2 is used for testing. The optimizer used is hard SGD momentum, with a value of leaning rate 0.005 and momentum 0.9, and epoch 50. The image size for the input is 224x224 and the batch size is 32. The optimization results from the two algorithms are, MobileNetV2 is better for detecting covid-19 with an accuracy value The precision reaches 99%. Future research can compare other CNN algorithms and larger data sets. Keywords: CNN; AlexNet; MobileNetV2; Covid-19
Menggunakan Xception, Transfer Learning, dan Permutasi untuk Meningkatkan Klasifikasi Ketidaksempurnaan Permukaan Baja: Using Xception, Transfer Learning, and Permutation to Improve the Classification of Steel Surface Imperfections Setiawati, Popong; Karno, Adhitio Satyo Bayangkari; Hastomo, Widi; Setiawan, Iwan
MALCOM: Indonesian Journal of Machine Learning and Computer Science Vol. 4 No. 1 (2024): MALCOM January 2024
Publisher : Institut Riset dan Publikasi Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.57152/malcom.v4i1.1258

Abstract

Kualitas permukaan baja yang diproduksi sangat penting untuk meningkatkan daya saing dalam industri baja. Tingginya tingkat cacat pada permukaan baja merupakan masalah serius yang berdampak pada kualitas keluaran. Pengendalian yang masih dilakukan secara manual dan visual saat ini hanya dapat dilakukan oleh orang-orang dengan bakat dan keahlian tertentu. Pengamatan dengan metode konvensional ini memerlukan waktu yang lama, lamban, dan presisi yang rendah. Saat ini, perkembangan teknik pembelajaran mendalam memungkinkan deteksi cacat permukaan baja secara otomatis dengan tingkat akurasi yang tinggi. Arsitektur Xception digunakan dalam pekerjaan ini untuk menerapkan strategi pembelajaran mendalam. Teknik permutasi dan augmentasi digunakan untuk mengatasi ketidakseimbangan data. Model yang dikembangkan dapat membedakan empat jenis cacat pada permukaan baja. Koleksi 7.095 foto permukaan baja digunakan dalam prosedur pelatihan. Jika dibandingkan dengan tidak menggunakan transfer learning, hasil pengukuran kinerja proses pelatihan dengan menggunakan transfer learning (Imagenet) menunjukkan hasil yang lebih baik. Pelatihan pembelajaran transfer menghasilkan skor akurasi masing-masing sebesar 94,9% dan 97,7% untuk data pelatihan dan validasi. Sedangkan hasil penilaian nilai kerugian untuk data latih dan validasi masing-masing sebesar 19,4% dan 14,4%.
OPTIMASI CONVOLUTION NEURAL NETWORK UNTUK DETEKSI COVID-19 Hastomo, Widi; Karno, Adhitio Satyo Bayangkari; Bakti, Indra
RADIAL : Jurnal Peradaban Sains, Rekayasa dan Teknologi Vol. 10 No. 2 (2022): RADIAL: JuRnal PerADaban SaIns RekAyasan dan TeknoLogi
Publisher : Universitas Bina Taruna Gorontalo

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37971/radial.v10i2.299

Abstract

Abstrak: Optimasi Convolution Neural Network Untuk Deteksi Covid-19. Kondisi pandemi seperti sekarang ini diperlukan sebuah algoritma pembelajaran mesin untuk mendeteksi covid-19 secara otomatis berdasarkan pada gambar rontgen dada guna memudahkan dalam mambantu pengambil keputusan. Penelitian ini ingin membandingkan arsitektur CNN AlexNet dan MobileNetV2 untuk mendeteksi (a) covid-19, (b) lung opacity, (c) normal, (d) viral pneumonia. Data himpunan rontgen dada yang digunakan sejumlah 4000 yang berasal dari kaggle.com, 0.8 data dibagi untuk pelatihan sedangkan 0.2 nya digunakan untuk pengujian. Optimizer yang digunakan yaitu keras SGD momentum, dengan nilai learning rate 0.005 dan momentum 0.9, serta epoch 50. Ukuran gambar untuk input yaitu 224x224 serta ukuran batch 32. Hasil optimasi dari kedua algoritma tersebut yaitu, MobileNetV2 lebih baik untuk mendeteksi covid-19 dengan nilai akurasi presisi mencapai 99%. Penelitian selanjutnya dapat membandingkan algoritma CNN yang lainnya serta data himpunan yang lebih banyak. Kata kunci: CNN; AlexNet; MobileNetV2; Covid-19 Abstract: Convolution Neural Network Optimization for Covid-19 Detection. In the current pandemic conditions, a machine learning algorithm is needed to detect COVID-19 automatically based on chest X-ray images to make it easier to assist decision makers. Aim study be disposed for compare the architecture of CNN AlexNet and MobileNetV2 to detect (a) covid-19, (b) lung opacity, (c) normal, (d) viral pneumonia. The data set of chest X-rays used are 4000 from kaggle.com, 0.8 of the data is shared for training while 0.2 is used for testing. The optimizer used is hard SGD momentum, with a value of leaning rate 0.005 and momentum 0.9, and epoch 50. The image size for the input is 224x224 and the batch size is 32. The optimization results from the two algorithms are, MobileNetV2 is better for detecting covid-19 with an accuracy value The precision reaches 99%. Future research can compare other CNN algorithms and larger data sets. Keywords: CNN; AlexNet; MobileNetV2; Covid-19