IRPI Publisher Journals (Institute of Research and Publication Indonesia)
Not a member yet
1012 research outputs found
Sort by
Predictive Sales Analysis in Coffee Shops Using the Random Forest Algorithm
Full text linkThe coffee shop industry has experienced significant growth, evolving into a highly competitive marketplace demanding specialty coffee and personalized experiences. While data-driven strategies are crucial for optimizing operations, many owners still struggle to effectively leverage their sales data to understand dynamic customer behavior and enhance decision-making. Addressing this gap, this study explores the application of machine learning (ML) techniques, specifically the Random Forest Regressor model, to predict sales performance within the coffee shop business environment. By analyzing factors such as transaction timing, store location, product type, and day of the week, this research aims to uncover patterns that can enhance inventory management and customer engagement. The Random Forest model was evaluated through cross-validation, yielding a mean Mean Squared Error (MSE) of 80.97, which indicates moderate predictive accuracy and represents an improvement over traditional forecasting methods commonly employed in the industry. Feature importance analysis revealed that Premium Beans is the most influential predictor, followed by seasonal trends (month), time of day, and weekend sales patterns. These findings underscore the importance of incorporating temporal and contextual factors into forecasting models.
Analisis Sentimen Coretax: Perbandingan Pelabelan Data Manual, Transformers-Based, dan Lexicon-Based pada Performa IndoBERT: Sentiment Analysis of Coretax: A Comparison of Manual, Transformers-Based, and Lexicon-Based Data Labeling on IndoBERT Performance
No full textAnalisis sentimen terhadap opini publik di media sosial menjadi tantangan signifikan karena kompleksitas bahasa informal dan volume data yang besar. Penelitian ini bertujuan untuk mengevaluasi pengaruh lima pendekatan pelabelan data manual, IndoBERT , IndoBERT weet, RoBERTa , dan InSet Lexicon terhadap performa model Indonesian Bidirectional Encoder Representations from Transformers (IndoBERT) dalam klasifikasi sentimen terkait isu Coretax. Sebanyak 8.035 tweet dikumpulkan, diproses, dan dilabeli menggunakan masing-masing pendekatan. Dataset hasil pelabelan kemudian digunakan untuk melatih ulang model IndoBERT, yang dievaluasi menggunakan metrik akurasi, F1-score, confusion matrix, dan kurva Receiver Operating Characteristic-Area Under the Curve (ROC-AUC). Hasil menunjukkan bahwa pelabelan otomatis menggunakan Indonesian Bidirectional Encoder Representations from Transformers for Tweet (IndoBERTweet) menghasilkan metrik tertinggi F1-Score (0,9802), tetapi mengalami dominasi kelas netral yang menunjukkan overfitting. Pelabelan manual menghasilkan distribusi kelas yang lebih merata meskipun dengan metrik lebih rendah F1-Score (0,8684), sedangkan Robustly Optimized BERT Pretraining Approach (RoBERTa) menunjukkan keseimbangan terbaik antara performa metrik dan distribusi label. InSet Lexicon dan IndoBERT menunjukkan kecenderungan bias terhadap kelas tertentu. Simpulan dari penelitian ini menegaskan bahwa efektivitas pelabelan tidak hanya ditentukan oleh skor metrik, tetapi juga oleh distribusi kelas yang seimbang untuk menghasilkan model yang adil dan dapat digeneralisasi
Analisis Time Series Pembangkit Listrik Tenaga Surya Berdasarkan Data Historis dan Iradiansi Menggunakan Metode ARIMA: Time Series Analysis Pembangkit Listrik Tenaga Surya Berdasarkan Data Historis dan Iradiansi Menggunakan Metode ARIMA
No full textThe demand for renewable energy in Indonesia continues to increase in line with the government's efforts to promote a sustainable energy transition. One of the rapidly growing technologies is On-Grid Solar Power Plants (PLTS), which rely on solar energy as their primary source. However, variations in solar irradiation and environmental factors cause fluctuations in the system's performance, potentially affecting its efficiency and reliability. Therefore, a robust method is needed to accurately predict system performance, supporting maintenance and operational optimization. This study applies the Seasonal Autoregressive Integrated Moving Average (SARIMA) method as a time series analysis approach to predict the Performance Ratio (PR) of PLTS based on historical data and solar irradiation variables. SARIMA was chosen because stationarity tests revealed a significant seasonal pattern that conventional ARIMA models cannot effectively handle. By considering seasonal factors, SARIMA provides a more accurate estimation of PR trends and fluctuations. This research aims to detect potential anomalies early, identify recurring operational patterns, and improve PLTS system monitoring efficiency. Model evaluation results show that SARIMA has higher accuracy than ARIMA in capturing seasonal patterns in PR data. Implementing this model can assist PLTS operators in making more data-driven decisions, optimizing maintenance strategies, and ensuring the reliability of renewable energy systems. These findings contribute to the development of more efficient energy management strategies and support the sustainability of solar energy utilization in Indonesia.Permintaan energi terbarukan di Indonesia terus meningkat seiring dengan upaya transisi energi berkelanjutan. Salah satu teknologi yang berkembang pesat adalah Pembangkit Listrik Tenaga Surya (PLTS) On-Grid, yang mengandalkan energi matahari sebagai sumber utama. Namun, variasi iradiasi matahari serta faktor lingkungan lainnya menyebabkan fluktuasi pada kinerja sistem PLTS, yang dapat berdampak pada efisiensi dan keandalannya. Oleh karena itu, diperlukan metode yang mampu memprediksi kinerja sistem dengan akurat untuk mendukung pemeliharaan serta optimalisasi operasional. Dalam studi ini, diterapkan metode Seasonal Autoregressive Integrated Moving Average (SARIMA) sebagai pendekatan analisis deret waktu guna memprediksi Performance Ratio (PR) dari PLTS berbasis data historis dan variabel iradiasi matahari. Model SARIMA dipilih karena hasil uji stasioneritas menunjukkan adanya pola musiman yang signifikan, yang tidak dapat ditangani secara optimal oleh model ARIMA konvensional. Dengan mempertimbangkan faktor musiman, SARIMA mampu memberikan estimasi yang lebih akurat terhadap tren dan fluktuasi PR. Penelitian ini bertujuan untuk mendeteksi potensi anomali sejak dini, mengidentifikasi pola operasional yang berulang, serta meningkatkan efisiensi pemantauan sistem PLTS. Hasil evaluasi model menunjukkan bahwa SARIMA memiliki tingkat keakuratan yang lebih tinggi dibandingkan dengan ARIMA dalam menangkap pola musiman pada data PR. Implementasi model ini dapat membantu operator PLTS dalam mengambil keputusan berbasis data yang lebih tepat, mengoptimalkan strategi perawatan, serta memastikan keandalan sistem energi terbarukan. Temuan ini memberikan kontribusi terhadap pengembangan strategi manajemen energi yang lebih efisien serta mendukung keberlanjutan dalam pemanfaatan tenaga surya di Indonesia..
 
Design and Development of a Makeup Artist Data Management Application : Design and Development of a Makeup Artist Data Management Application
No full textThe growing beauty industry requires efficient management tools for Makeup Artist (MUA) teams, who often face operational challenges due to the use of manual or non-integrated digital tools. This study contributes by providing an integrated, mobile-first workflow that directly links bookings, schedules, and financial records with role-based access and automated reminders, reducing fragmentation from manual or separate tools and offering consolidated insights for MUAs’ decision-making. The application's quality was validated through Black-Box testing, achieving a 100% success rate across 35 functional test cases that covered all main features. Usability was then evaluated in a case study with a professional MUA and two team members (N=3) who used the application for one week before completing the Computer System Usability Questionnaire (CSUQ). The results yielded an overall mean score of 6.52 out of 7, indicating very high user satisfaction, with subscale scores showing that users found the system both helpful and easy to navigate. Future development may include expanding the dashboard's analytical capabilities or integrating a direct client payment gateway
Transformation of Village Entrepreneurship: Financial Literacy as a Driver of Economic Independence for Borrowers of BUMDesa Bersama in Serang Regency: Transformasi Kewirausahaan Desa: Literasi Keuangan sebagai Pendorong Kemandirian Ekonomi Bagi Debitur BUMDesa Bersama di Kabupaten Serang
Full text linkThe aim of this community service (PKM) initiative is to facilitate the improvement of financial literacy among debtors of BUMDesma LKD across Serang Regency. The execution of this PKM activity was conducted in three phases: Planning, Implementation, and Evaluation to attain best outcomes. During the planning phase, the team formulated the PKM activity idea, which consists of three primary steps: assessing the financial management attributes of the borrowers, delivering financial management training to the debtors, and offering post-training mentorship. In the implementation phase, debtor characteristics were discerned through observations and interviews of 22 debtors from 18 BUMDesma LKD throughout 17 sub-districts in Serang Regency. The second stage entailed administering financial management instruction for the debtors utilizing a practical financial management methodology known as "GoDisKo." The concluding phase of the implementation emphasized post-training mentorship. The evaluation, as the last activity, demonstrated that the beginning phase substantially facilitated the following training phase, allowing for successful and efficient training execution. Nevertheless, the mentoring phase was suboptimal owing to the inherent preferences of the debtors, who exhibited a diminished propensity to engage with online media for mentoring purposes
Implementation of Machine Learning Algorithm for Heart Attack Disease Prediction
No full textHeart attack disease is one of the leading causes of death worldwide, making early detection a critical factor in reducing mortality. However, manual prediction is often inaccurate due to the complexity of medical data. To address this issue, this study evaluates five machine learning algorithms K-Nearest Neighbors (KNN), Decision Tree, Naïve Bayes, Random Forest, and Support Vector Machine (SVM) for predicting heart attack risk. The dataset, obtained from Kaggle, was preprocessed and divided into training and testing sets using 70:30 and 80:20 ratios. Algorithm performance was assessed using accuracy, precision, recall, and F1-score. The results showed that Decision Tree and Random Forest achieved the best performance with accuracy up to 97.98%, while KNN recorded the lowest accuracy at around 61.36%. This study not only demonstrates the comparative effectiveness of these algorithms on the same dataset, contributing to the growing body of research on AI in healthcare, but also highlights their potential clinical utility. In particular, Decision Tree and Random Forest can support the development of AI-based clinical decision support systems to assist healthcare professionals in early diagnosis and risk managemen
Classification of A Credit Card Fraud Detection Model Using XGBoost with Smote and Gridsearchcv Optimization
No full textThe development of digital technology has motivated rapid growth in online transactions, so the increase in the volume of digital transactions also increases the risk of credit card fraud, particularly in transactions where a card is not present. By employing the Extreme Gradient Boosting (XGBoost) method in conjunction with the Synthetic Minority Over-sampling Technique (SMOTE) to solve class imbalance and fine-tuning model parameters using GridSearchCV, this study aims to improve a fraud detection system. The dataset, which consists of anonymized credit card transactions, presents a stark imbalance with fraudulent cases accounting for only 0.172% of the data. The study involves several stages: preprocessing the data, balancing class distribution, training the model, and evaluating its performance through metrics such as F1-score, precision, recall, accuracy, and AUC-ROC. Implementation of SMOTE proved effective in enhancing the representation of rare fraud cases without introducing overfitting, while GridSearchCV identified the most effective parameter configuration. The resulting model achieved top-tier performance with 100% accuracy, 0.81 precision, 0.85 recall, an F1-score of 0.83, and an AUC-ROC of 0.979, indicating strong capability in distinguishing fraudulent from legitimate transactions. The novelty of this study lies in the systematic integration of SMOTE, XGBoost, and GridSearchCV into a unified pipeline designed to address extreme class imbalance in real-world credit card transactions. Unlike previous studies that focused solely on algorithm comparison or hyperparameter tuning, this research emphasizes reducing false negatives, which pose the greatest financial and reputational risks. The findings not only demonstrate superior performance metrics but also provide practical contributions for financial institutions, regulators, and e-commerce platforms in developing scalable, reliable, and adaptive fraud detection system
A Simulation of Student Study Group Formation Design Using K-Means Clustering
No full textThis research focuses on developing a simulation model for forming student study groups using an enhanced K-Means algorithm, addressing the challenge of optimizing group dynamics to improve learning outcomes. By analyzing the effectiveness of the formed study groups through RMSE (Root Mean Square Error) after dimensionality reduction with various regression models—including Linear Regression, Ridge Regression, Lasso Regression, Elastic Net, Random Forest Regressor, Gradient Boosting Regressor, and XGBoost Regressor—we aim to provide educators with a robust tool for assessing group configurations. The study identifies four distinct clusters, revealing that "Previous_Score" and "Attendance" are critical variables, achieving a highest Silhouette Score of 0.64 with five selected features. The ridge regression model also yielded a low RMSE of 0.045, explaining 72.39% of the variance in "Exam_Score." The findings suggest that targeted interventions tailored to each cluster—yellow, purple, blue, and green—can enhance academic outcomes by addressing specific student needs. This data-driven approach optimizes group dynamics and fosters a more inclusive learning environment, enhancing academic performance and cultivating essential social skills. The study underscores the potential of machine learning techniques in education and suggests avenues for future research into alternative clustering methods and their long-term impact on student engagement and success
Real-Time Road Damage Detection on Mobile Devices using TensorFlow Lite and Teachable Machine
Full text linkThis study presents a mobile-based road damage detection system using Teachable Machine and TensorFlow Lite to support real-time monitoring and efficient infrastructure maintenance. The system identifies road damage types such as cracks, potholes, and uneven surfaces. The RDD2020 dataset is used for model training, with preprocessing steps including augmentation, normalization, and resizing. A Convolutional Neural Network (CNN) model is trained through Teachable Machine for ease of customization. TensorFlow Lite is employed for on-device inference, with optimization techniques like quantization and pruning applied to improve speed and reduce model size. The system is evaluated using precision, recall, F1-score, and accuracy metrics under varying lighting and weather conditions. The final model is deployed in a mobile app using TensorFlow Lite Interpreter for efficient performance. Experimental results show high detection accuracy, with a precision of X% and F1-score of Y% (insert actual values). This approach offers a lightweight, cost-effective solution for road maintenance authorities and urban planners. Future enhancements include dataset expansion, integration with mapping tools, and improved robustness in diverse environments. Overall, the proposed system enables real-time, accurate road damage detection and supports smarter, eco-friendly infrastructure management
Smart Prescription Reader: Enhancing Accuracy in Medical Prescriptions
Full text linkReading a doctor's handwritten prescription is a challenge faced by most patients and some pharmacists, which in some cases can lead to negative consequences due to misinterpretation of the prescription. The "Doctor's Handwritten Prescription BD Dataset" on Kaggle contains segmented images of handwritten prescription words from BD (Bangladesh) doctors. This dataset, intended for machine learning applications, includes 4,680 individual words segmented from prescription images. This study introduces a Handwriting Recognition System using Convolutional Neural Network (CNN) developed to identify text in prescription images written by doctors and convert the cursive handwriting into readable text. Two models were evaluated in this study: CNN and MobileNet. Based on the experiments, MobileNet showed better results compared to CNN alone. From the dataset of 4,680 words, 3,120 were used for training, 780 for testing, and 780 for validation. The study achieved a training accuracy of 97%, a testing accuracy of 88%, and a validation accuracy of 83%. The developed model was successfully implemented in a web applicatio