Garuda - Garba Rujukan Digital

Implementasi Algoritma Convolutional Neural Network Untuk Klasifikasi Citra Kemasan Kardus Defect dan No Defect Antoni, Alan; Rohana, Tatang; Pratama, Adi Rizky
Building of Informatics, Technology and Science (BITS) Vol 4 No 4 (2023): March 2023
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v4i4.3270

Packaging is an important aspect of a product, because packaging can affect the quality and competitiveness of the product. Damaged packaging can result in decreased product quality. One popular packaging used is corrugated cardboard type box. To visually distinguish defect and no defect cardboard packaging, there are tears, holes and dents on the defect cardboard packaging. Whereas the no defect cardboard packaging has a visual that there are no tears, holes or dents. To simplify the classification, technology is needed that can distinguish between defect and no-defect cardboard packaging. In this study the total images used as a dataset are 1300 images, which are then divided into 2 with a percentage of 80% for training data and 20% for test data. The dataset first goes through the preprocessing stage before being used. Preprocessing consists of cropping, augmentation and resizing. And also do the segmentation process using Grabcut method. Then feature extraction is also performed using Local Binary Pattern (LBP). This study uses the Convolutional Neural Network algorithm with a total of 3 convolution layers, namely 16.32 and 64 and the Adam optimizer. Four experiments were carried out by differentiating the hyperparameter epoch, the input image size and the learning rate. The results showed that the model produced with an epoch hyperparameter of 30, an input image size of 300x300 and a learning rate of 0.001 obtained the best performance with an accuracy value of 95.77%, 96% precision, 96% recall, 96% f1-score and loss of 0.1478.

COMPARISON OF DIABETES DISEASE CLASSIFICATION MODELS USING LOGISTIC REGRESSION AND RANDOM FOREST ALGORITHMS nabila, putri; Mutoi Siregar, Amril; Faisal, Sutan; Pratama, Adi Rizky
Faktor Exacta Vol 17, No 3 (2024)
Publisher : LPPM

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30998/faktorexacta.v17i3.24388

Diabetes is a lifelong chronic disease that disrupts blood sugar regulation. Diabetes is a life-threatening condition that, if left untreated, can lead to death and other health problems. Several medical tests, including the glycated hemoglobin (A1C) test, blood sugar test, oral glucose tolerance test, and fasting blood sugar test, can be used to detect diabetes. According to statistics, high glucose levels are one of the problems associated with diabetes. This study aims to categorize patients into diabetic and non-diabetic groups using specific diagnostic metrics included in the dataset. 1500 patient records with 9 attributes and 2 classes were used by the researchers. The study used machine learning techniques, including Logistic Regression and Random Forest, along with Confusion Matrix and Receiver Operating Characteristics (ROC) assessment. The Random Forest method produced results of 97% accuracy, 97% precision, 100% recall, and 98% f1-score, indicating that the accuracy level seems good but can still be improved. Based on the accuracy findings, Random Forest is the most effective strategy of Logistic Regression.

DETECTION OF THE SIZE OF PLASTIC MINERAL WATER BOTTLE WASTE USING THE YOLOV5 METHOD Karyanto, Dony Dwi; Indra, Jamaludin; Pratama, Adi Rizky; Rohana, Tatang
JIKO (Jurnal Informatika dan Komputer) Vol 7, No 2 (2024)
Publisher : Universitas Khairun

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33387/jiko.v7i2.8535

The use of plastic bottles for various needs is increasingly massive, especially in consumption needs such as mineral water bottles. The use of plastic bottles is used to reduce costs and be effective in maintaining the quality of mineral water, but its impact can affect natural conditions if not managed properly. Plastic bottle waste if left buried in the ground will have difficulty expanding, which can cause environmental pollution. Therefore, we can take advantage of technology to sort plastic bottle waste using a camera based on the size of plastic bottles. Differentiating the size of bottles aims to distinguish the economic value when exchanged at the waste bank. This technology utilizes object detection and recognition functions such as the YOLO (You Only Look Once) method. YOLO is a detection method that is a development of the CNN (Convolutional Neural Network) algorithm. By using YOLOv5, we can detect objects in the form of plastic bottle waste of various different sizes. To maximize object detection according to size, data annotation is done by creating a Bounding Box on each dataset according to its size. The test was carried out with several different distance configurations including 40cm, 80cm and 1m. Detection results using YOLOv5 produce up to 84% accuracy in real-time.

Prediksi Pola Pergerakan Saham Adro.Jk Melalui Model LSTM Berbasis Data Historis Iskandar, Muhammad Irsyad; Mudzakir, Tohirin Al; Cahyana, Yana; Pratama, Adi Rizky
Bulletin of Computer Science Research Vol. 5 No. 4 (2025): June 2025
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bulletincsr.v5i4.554

The fluctuating nature of stock price movements presents a significant challenge in investment decision-making. To address this issue, a predictive model capable of capturing historical patterns and accurately forecasting stock prices is required. This study aims to develop a stock price prediction model for PT Alamtri Resources Indonesia Tbk (ADRO.JK) using the Long Short-Term Memory (LSTM) algorithm. The dataset comprises daily closing prices from January 1, 2020, to December 30, 2024, obtained from Yahoo Finance. The data was processed in a time series format using a sliding window approach, employing 30 historical data points to predict the next price point. The model was constructed using two LSTM layers, one Dense layer, and techniques such as Dropout and EarlyStopping to prevent overfitting.The training and testing results indicate that the model performs exceptionally well, achieving a Mean Absolute Percentage Error (MAPE) of 0.0341 or 3.41%, corresponding to a prediction accuracy of 96.59%. In a short-term prediction scenario over seven days, the model achieved an accuracy of 99.07% (MAPE = 0.0093), while in a medium-term scenario up to May 19, 2025, it achieved an accuracy of 98.76% (MAPE = 0.0124). The predicted stock price on May 19, 2025, is estimated at IDR 1,913.76. With its high accuracy and low error rate, the LSTM model has proven to be a reliable tool for forecasting stock prices based on historical data.

Optimasi AdaBoost dan XGBoost untuk Klasifikasi Obesitas Menggunakan SMOTE Sukmawati, Cici Emilia; Pratama, Adi Rizky; Hikmayanti, Hanny; Juwita, Ayu Ratna
Jurnal Informatika: Jurnal Pengembangan IT Vol 10, No 3 (2025)
Publisher : Politeknik Harapan Bersama

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30591/jpit.v10i3.8536

Obesity is a condition in which a person's weight exceeds the normal limit due to excessive accumulation of fat tissue. Thus, obesity is considered a global public health challenge. This is evidenced by the latest data from the World Health Organization (WHO) in 2022, namely that 2.5 billion adults aged 18 years and over are overweight and 890 million of them are obese. Therefore, it is very important to accurately identify these risk factors in order to implement effective interventions in the prevention and management of obesity. However, in previous studies there has been no application of SMOTE with the AdaBoost and XGBoost algorithms, so this study aims to compare the performance of the AdaBoost and XGBoost algorithms with SMOTE. The stages of this research begin with problem identification, data collection, preprocessing and model evaluation and model comparison. This study also applies the SMOTE technique to balance unbalanced data. Based on the results of the research that has been carried out, it shows that the accuracy and recall values of the XGBoost algorithm with SMOTE are 0.945 and precision 0.947. Meanwhile, the accuracy and recall values on AdaBoost with SMOTE are 0.388. Then, the precision is 0.371. Thus, it is expected that the results of the XGBoost model with SMOTE can be a source for other research and can help in efforts to prevent and manage obesity.

EXIF Metadata Feature Extraction to Improve Source Device Identification Accuracy in Digital Images within a Digital Forensics Approach Bahreisy, Muhammad Naufal; Pratama, Adi Rizky; Munzi, Gugy Guztaman; Wicaksana, Yusuf Eka
JURNAL SISFOTEK GLOBAL Vol 15, No 2 (2025): JURNAL SISFOTEK GLOBAL
Publisher : Institut Teknologi dan Bisnis Bina Sarana Global

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.38101/sisfotek.v15i2.15998

This study aims to develop and evaluate methods for digital image source device identification through three main approaches, namely EXIF metadata feature extraction, visual analysis using Convolutional Neural Networks (CNN), and Photo Response Non-Uniformity (PRNU). The dataset consists of 500 original images captured from five different devices, with 100 images per device containing intact metadata. The EXIF-only model was built using the Random Forest algorithm, the CNN model employed a ResNet18 architecture, while PRNU utilized high-pass filtering to construct sensor noise templates for each device. Evaluation was carried out using accuracy, precision, recall, and f1-score metrics. The results show that EXIF-only achieved perfect accuracy (100%) on the dataset with complete metadata, CNN reached 21% accuracy with imbalanced recall across classes, and PRNU demonstrated low performance due to the limited number of templates and image quality. These findings indicate that EXIF-only excels under intact metadata conditions but is vulnerable to manipulation, CNN can be applied when metadata is unavailable but requires optimization, while PRNU has potential resilience against metadata manipulation but demands higher-quality data. The novelty of this study lies in its comparative multi-method approach that integrates metadata-based, visual-based, and sensor fingerprint-based analyses, along with the proposal of a multimodal integration framework to enhance the reliability of device identification systems in digital forensic practice.

INTRODUCTION NATIONAL IDENTIFICATION NUMBER AND NAME ON ID CARD USING OCR (OPTICAL CHARACTER RECOGNITION) METHOD Holila, Holila; Pratama, Adi Rizky; Lestari, Santi Arum Puspita; Indra, Jamaludin
Jurnal Teknik Informatika (Jutif) Vol. 5 No. 4 (2024): JUTIF Volume 5, Number 4, August 2024
Publisher : Informatika, Universitas Jenderal Soedirman

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52436/1.jutif.2024.5.4.2242

This study examines the use of Optical Character Recognition (OCR) methods for the automatic recognition and extraction of text from images of Identity Cards (KTP). The aim is to provide an effective solution to the problems of document forgery and duplication, particularly in the use of KTP as an identity verification tool. Utilizing the Tesseract library, this research involves preprocessing steps such as conversion to grayscale, perspective transformation, and noise reduction to enhance OCR accuracy. Testing was conducted with 50 different KTP images using Python programming, achieving an Optical Character Recognition accuracy rate of 91%. Additionally, tests conducted with a dataset of 50 KTP images containing NIK and name variables showed that all images were successfully detected with an accuracy rate of 90%. This study confirms that the OCR method is effective in reading text from KTP images in real-time, thus it can be implemented for automatic identity verification.

Analisis Sentimen Pemboikotan Produk dengan Pendekatan Algoritma Naïve Bayes Media Sosial X Rifaldi, Rizky; Indra, Jamaludin; Pratama, Adi Rizky; Juwita, Ayu Ratna
Journal of Information System Research (JOSH) Vol 5 No 4 (2024): Juli 2024
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/josh.v5i4.5420

This research aims to analyze sentiment regarding the problem of product boycotting by the public using the Naive Bayes algorithm. 1426 data were collected from social media x to study consumer behavior towards certain products. Through the application of the Naive Bayes algorithm, sentiment analysis was carried out to identify patterns in consumer opinions regarding boycotting the products studied. Experimental results show that the Naive Bayes algorithm succeeded in achieving 81% accuracy in classifying sentiment towards products. This shows the algorithm's ability to analyze consumer sentiment effectively, which can provide valuable insights for companies in understanding public perception and managing the reputation of their products. The practical implication of this research is the importance of utilizing sentiment analysis techniques in marketing strategy and brand management to increase product competitiveness in a competitive market.

DETECTION OF THE SIZE OF PLASTIC MINERAL WATER BOTTLE WASTE USING THE YOLOV5 METHOD Karyanto, Dony Dwi; Indra, Jamaludin; Pratama, Adi Rizky; Rohana, Tatang
JIKO (Jurnal Informatika dan Komputer) Vol 7, No 2 (2024)
Publisher : Universitas Khairun

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33387/jiko.v7i2.8535

The use of plastic bottles for various needs is increasingly massive, especially in consumption needs such as mineral water bottles. The use of plastic bottles is used to reduce costs and be effective in maintaining the quality of mineral water, but its impact can affect natural conditions if not managed properly. Plastic bottle waste if left buried in the ground will have difficulty expanding, which can cause environmental pollution. Therefore, we can take advantage of technology to sort plastic bottle waste using a camera based on the size of plastic bottles. Differentiating the size of bottles aims to distinguish the economic value when exchanged at the waste bank. This technology utilizes object detection and recognition functions such as the YOLO (You Only Look Once) method. YOLO is a detection method that is a development of the CNN (Convolutional Neural Network) algorithm. By using YOLOv5, we can detect objects in the form of plastic bottle waste of various different sizes. To maximize object detection according to size, data annotation is done by creating a Bounding Box on each dataset according to its size. The test was carried out with several different distance configurations including 40cm, 80cm and 1m. Detection results using YOLOv5 produce up to 84% accuracy in real-time.

Title

Found 9 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 9 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 9 Documents
Search