Claim Missing Document
Check
Articles

Found 28 Documents
Search

Comparison of LSTM and GRU Models for Forex Prediction Pahlevi, Mohammad Rezza; Kusrini, Kusrini; Hidayat, Tonny
Sinkron : jurnal dan penelitian teknik informatika Vol. 7 No. 4 (2023): Article Research Volume 7 Issue 4, October 2023
Publisher : Politeknik Ganesha Medan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33395/sinkron.v8i4.12709

Abstract

Trading foreign currencies worth trillions of dollars takes place daily in the forex market, characterized by highly volatile movements. The forex market operates on bid and ask prices, with exchange rates determined by the principles of supply and demand. Trading involves currency pairs like EUR/USD, where the value of the Euro is compared to the US Dollar, serving as a basis for analyzing price fluctuations. Due to the volatile nature of forex, market participants must make informed decisions when buying and selling, as improper choices can result in financial losses. One approach to mitigating risk in forex trading decisions is through the use of forecasting techniques. This research study employs LSTM and GRU methods to predict forex trends, which are evaluated using various dataset divisions. The most accurate results are obtained using a dataset of 4979, split into three equal parts: 80% for training, 10% for validation, and 10% for testing. This approach yields an RMSE value of 0.054, MAPE of 0.037, and R-square of 97%
Image Augmentation for BreaKHis Medical Data using Convolutional Neural Networks Istighosah, Maie; Sunyoto, Andi; Hidayat, Tonny
Sinkron : jurnal dan penelitian teknik informatika Vol. 7 No. 4 (2023): Article Research Volume 7 Issue 4, October 2023
Publisher : Politeknik Ganesha Medan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33395/sinkron.v8i4.12878

Abstract

In applying Convolutional Neural Network (CNN) to computer vision tasks in the medical domain, it is necessary to have sufficient datasets to train models with high accuracy and good general ability in identifying important patterns in medical data. This overfitting is exacerbated by data imbalances, where some classes may have a smaller sample size than others, leading to biased predictive results. The purpose of this augmentation is to create variation in the training data, which in turn can help reduce overfitting and increase the ability of the model to generalize. Therefore, comparing augmentation techniques becomes essential to assess and understand the relative effectiveness of each method in addressing the challenges of overfitting and data imbalance in the medical domain. In the context of the research described, namely a comparative analysis of augmentation performance on CNN models using the ResNet101 architecture, a comparison of augmentation techniques such as Image Generator, SMOTE, and ADASYN provides insight into which technique is most suitable for improving model performance on limited medical data. By comparing these techniques' accuracy, recall, and overall performance results, research can identify the most effective and relevant techniques in addressing the challenges of complex medical datasets. This provides a valuable guide for developing better CNN models in the future and may encourage further research in developing more innovative augmentation methods suitable for the medical domain.
Breast Cancer Detection in Histopathology Images using ResNet101 Architecture Istighosah, Maie; Sunyoto, Andi; Hidayat, Tonny
Sinkron : jurnal dan penelitian teknik informatika Vol. 7 No. 4 (2023): Article Research Volume 7 Issue 4, October 2023
Publisher : Politeknik Ganesha Medan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33395/sinkron.v8i4.12948

Abstract

Cancer is a significant challenge in many fields, especially health and medicine. Breast cancer is among the most common and frequent cancers in women worldwide. Early detection of cancer is the main step for early treatment and increasing the chances of patient survival. As the convolutional neural network method has grown in popularity, breast cancer can be easily identified without the help of experts. Using BreaKHis histopathology data, this project will assess the efficacy of the CNN architecture ResNet101 for breast cancer image classification. The dataset is divided into two classes, namely 1146 malignant and 547 benign. The treatment of data preprocessing is considered. The implementation of data augmentation in the benign class to obtain data balance between the two classes and prevent overfitting. The BreaKHis dataset has noise and uneven color distribution. Approaches such as bilateral filtering, image enhancement, and color normalization were chosen to enhance image quality. Adding flatten, dense, and dropout layers to the ResNet101 architecture is applied to improve the model performance. Parameters were modified during the training stage to achieve optimal model performance. The Adam optimizer was used with a learning rate 0.0001 and a batch size of 32. Furthermore, the model was trained for 100 epochs. The accuracy, precision, recall, and f1-score results are 98.7%, 98.73%, 98.7%, and 98.7%, respectively. According to the results, the proposed ResNet101 model outperforms the standard technique as well as other architectures.
Optimizing Facial Expression Recognition with Image Augmentation Techniques: VGG19 Approach on FERC Dataset Ilmawati, Fahma Inti; Kusrini, Kusrini; Hidayat, Tonny
Sinkron : jurnal dan penelitian teknik informatika Vol. 8 No. 2 (2024): Article Research Volume 8 Issue 2, April 2024
Publisher : Politeknik Ganesha Medan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33395/sinkron.v8i2.13507

Abstract

In the field of facial expression recognition (FER), the availability of balanced and representative datasets is key to success in training accurate models. However, Facial Expression Recognition Challenge (FERC) datasets often face the challenge of class imbalance, where some facial expressions have a much smaller number of samples compared to others. This issue can result in biased and unsatisfactory model performance, especially in recognizing less common facial expressions. Data augmentation techniques are becoming an important strategy as they can expand the dataset by creating new variations of existing samples, thus increasing the variety and diversity of the data. Data augmentation can be used to increase the number of samples for less common facial expression classes, thus improving the model's ability to recognize and understand diverse facial expressions. The augmentation results are then combined with balancing techniques such as SMOTE coupled with undersampling to improve model performance. In this study, VGG19 is used to support better model performance. This will provide valuable guidelines for optimizing more advanced CNN models in the future and may encourage further research in creating more innovative augmentation techniques.
Optimasi Algoritma C4.5 Menggunakan Metode Forward Selection Dan Stratified Sampling Untuk Prediksi Kelayakan Mahasiswa Penerima Beasiswa Bentar Candra P; Kusrini, Kusrini; Tonny Hidayat
KLIK: Kajian Ilmiah Informatika dan Komputer Vol. 4 No. 6 (2024): Juni 2024
Publisher : STMIK Budi Darma

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30865/klik.v4i6.1933

Abstract

Every prospective student has the opportunity to get a scholarship within an educational institution, but it is often not on target so a more accurate data mining approach is needed. However, the C4.5 algorithm has a weakness in its level of accuracy when managing large amounts of data so it needs to be optimized. This research aims to optimize the C4.5 algorithm using stratified sampling and forward selection methods in determining the eligibility of scholarship recipients. The data came from prospective students at Anwar Medika University with a sample size of 263 records which were then processed using the RapidMinner application for the C4.5 algorithm without optimization and the C4.5 algorithm with optimization of the stratified sampling + forward selection method. The research results show a higher level of accuracy in the C4.5 algorithm with optimization using the stratified sampling + forward selection method, namely 81.75% compared to the accuracy level in the C4.5 algorithm without optimization, namely 80.23%. Thus, the conclusion of this research is that the C4.5 algorithm with optimization using stratified sampling and forward selection methods is more effective and can overcome the shortcomings of the C4.5 algorithm without optimization
Analisis pencapaian kinerja menggunakan regresi linier dan ARIMA (studi kasus: KSP Kredit Union Pancur Solidaritas) Safril, Martinus; Hidayat, Tonny
Jurnal Ilmiah Teknologi Informasi Asia Vol 19 No 1 (2025): Volume 19 nomor 1 2025 (8)
Publisher : LP2M Institut Teknologi dan Bisnis ASIA Malang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32815/jitika.v19i1.1078

Abstract

Continuous performance measurement is important for organizations, especially cooperatives such as KSP (Koperasi Simpan Pinjam) Credit Union Pancur Solidaritas. This ensures work programs are planned and completed effectively and allows for ongoing success evaluation. Linear Regression and ARIMA are methods applied to set targets for organizational work programs and measure the cooperative's performance over time. This study aimed to examine performance achievement using Linear Regression and ARIMA (Auto Regressive Integrated Moving Average). The research used a quantitative descriptive approach. Study data included documentation on assets, member numbers, outreach activities, loan disbursements, overdue loans, and staff count for KSP Credit Union Pancur Solidaritas from 2021 to 2024. Data analysis employed Linear Regression and ARIMA tests performed using Python software. The study results showed that combining Linear Regression and ARIMA can produce three different performance possibilities: the highest anticipated performance (upper performance), the predicted performance (predicted performance), and the lowest possible performance (lower performance). Based on this analysis, the prediction for KSP CUPS member growth indicates an increase each month, with growth predicted to be 1,426 members by June 2025.
A Systematic Literature Review Of Mental Health Diagnostic Using K-Nearest Neighbour - Whale Optimization Algorithm Septian, Firza; Kusrini , Kusrini; Hidayat, Tonny
SISKOMTI: Jurnal Sistem Informasi Komputer dan Teknologi Informasi Vol. 5 No. 1 (2023): Februari 2023
Publisher : Universitas Lembah Dempo

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54342/ptc0pb11

Abstract

People including unborn infants are negatively impacted by a number of things, such as noise. Noise and the others aspect could affect somebody mental health. Mental health as natural problem might be easier detected using metaheuristic algorithm, K-Nearest Neighbour - Whale Optimization Algorithm (KNN-WOA) is one of them. A variety of trustworthy sources, including IEEE and Scopus, are used to collect the data. In the action research technique, practical applications work as a "Laboratory" for testing hypotheses on synthesized products. There are three fundamental ideas in regard to using WOA for medical purposes. KNN will be used according to the plan for medical diagnostics. WOA, a population-based approach, uses a randomized collectivist humpback whale sample to enhance potential solutions as feature selection while KNN as the main algorithm. Only three of the 94 journals collected met the set standards.
DiG-MFV: Dual-integrated Graph for Multilingual Fact Verification Agustina, Nova; Kusrini; Utami, Ema; Hidayat, Tonny
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 9 No 4 (2025): August 2025
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v9i4.6695

Abstract

The proliferation of misinformation in political domains, especially across multilingual platforms, presents a major challenge to maintaining public information integrity. Existing models often fail to effectively verify claims when the evidence spans multiple languages and lacks a structured format. To address this issue, this study proposes a novel architecture called Dual-integrated Graph for Multilingual Fact Verification (DiG-MFV), which combines semantic representations from multilingual language models (i.e., mBERT, XLM-R, and LaBSE) with two graph-based components: an evidence graph and a semantic fusion graph. These components are processed through a dual-path architecture that integrates the outputs from a text encoder and a graph encoder, enabling deeper semantic alignment and cross-evidence reasoning. The PolitiFact dataset was used as the source of claims and evidence. The model was evaluated by using a data split of 70% for training, 20% for validation, and 10% for testing. The training process employed the AdamW optimizer, cross-entropy loss, and regularization techniques, including dropout and early stopping based on the F1-score. The evaluation results show that DiG-MFV with LaBSE achieved an accuracy of 85.80% and an F1-score of 85.70%, outperforming the mBERT and XLM-R variants, and proved to be more effective than the DGMFP baseline model (76.1% accuracy). The model also demonstrated stable convergence during training, indicating its robustness in cross-lingual political fact verification tasks. These findings encourage further exploration in graph-based multilingual fact verification systems.
ANALISIS RUNTUN WAKTU DALAM MEMPREDIKSI PENERIMAAN MAHASISWA BARU MENGGUNAKAN REGRESI LINEAR: Studi Kasus : UNIVERSITAS SATYA WIYATA MANDALA) Mawene, Brenstein; Kusrini; Hidayat, Tonny
Jurnal FATEKSA : Jurnal Teknologi dan Rekayasa Vol. 9 No. 1 (2024): Jurnal FATEKSA, Jurnal Teknologi dan Rekayasa
Publisher : Fakultas Teknologi dan Rekayasa

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Perlembangan teknologi menghasilkan modernitas dibidang pendidikan serta bidang lainnya dalan kehidupan sehari-hari. Tujuan dari penelitian ini yaitu untuk dapat memprediksi jumlah penerimaan mahasiswa baru ditahun mendatang, dengan menggunakan Metode Regresi Linear dan Gradient Boosting. Prediksi jumlah mahasiswa baru diperlukan untuk membantu Universitas Satya Wiyata Mandala untuk mengetahui jumlah mahasiswa baru di tahun yang akan datang sehingga pihak Universitas dapat melakukan evaluasi untuk mempersiapkan Sarana Dan Prasarana Perkuliahan. Dalam penelitian ini peneliti berfokus pada Jumlah Mahasiswa Baru. Berdasarkan evaluasi komparatif Model Regresi Linear menunjukan performa yang buruk dengan nilai MAE 32.83, RMSE 39.74, R² -0.11 sedangkan Model Gradient Boosting menunjukan performa yang lebih baik dengan nilai MAE 13.16, RMSE 17.13, R² 0.79. secara keseluruhan model Gradient Boosting memberikan hasil prediksi yang lebih baik dibandingkan Model Regresi Linear.
Optimizing 4-ary Huffman Trees and Normalizing Binary Code Structures to Minimize Redundancy and Level Reduction Hidayat, Tonny; Kurniawan , Hendra; Mustopa , Ali; Kuswanto, Jeki
Elinvo (Electronics, Informatics, and Vocational Education) Vol. 10 No. 1 (2025): May 2025
Publisher : Department of Electronic and Informatic Engineering Education, Faculty of Engineering, UNY

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21831/elinvo.v10i1.78722

Abstract

Since the present data expansion and increase are occurring at an increasingly rapid pace, the solution of adding storage space is not sustainable in the long run. The growing need for storage media can be addressed with lossless compression, which reduces stored data while allowing complete restoration. Huffman remains a potent method for data compression, functioning as a "back end" process and serving as the foundational algorithm in applications, among others, Monkey's PKZIP, WinZip, 7-Zip, and Monkey's Audio. Lossless compression of 16-bit audio requires binary structure adjustments to balance speed and optimal compression ratio. The use of a 4-ary Huffman tree (4-ary) branching procedure to generate binary code generation and to insert a maximum of 2 dummy data symbol variables that are given a binary value of 0 with the condition that if the number of MOD 3 data variables = remaining 2, then two dummy data are added, if the result is the remainder 0 = 1 dummy data, and if the remainder = 1 then it is not required. This process effectively maintains a high ratio level while speeding up the 4-ary Huffman code algorithm's performance in compression time. The results show that the efficiency reaches 95.94%, the ratio is 38%, and the comparison is 1/3 of the Level based on calculations, testing, and comparison with other generations of the Huffman code. The 4-ary algorithm significantly optimizes archived data storage, reducing redundancy to 0.124 and achieving an entropy value of 2.91 across various data types.