Publication Search

This research aims to analyze the best-selling products at Toko Hartati using the K-Means Clustering method. K-Means Clustering is an unsupervised learning algorithm that is effective in grouping data based on certain similar characteristics. In this context, the data used includes the number of sales, product prices, and product categories. Through this analysis, it is hoped that insight can be gained regarding products that have the best sales performance, as well as sales patterns that can be used as a reference in stock management and marketing strategies. The data used in this research includes sales transactions during a certain period, with the aim of identifying product clusters based on sales patterns. The analysis results show the existence of two main product groups, where the first cluster contains products with high sales numbers, which can be classified as best-selling products, while the second cluster includes products with lower sales. These findings provide valuable information for the management of Toko Hartati in determining more targeted marketing strategies and more efficient stock management. This research suggests using the K-Means Clustering method in data-based decision making to improve sales performance in retail stores.

Topic modeling is an integral text mining component, employing diverse algorithms to uncover hidden themes within texts. This study examines the comparative performance of prominent topic modeling techniques on news headlines, which is characterized by brevity and specific linguistic style. Given the corpus originates from a non-native English-speaking country, an additional layer of complexity is introduced to the task. Our research explores the feasibility of employing a committee approach for topic modeling, evaluating the efficacy and challenges of various methods in practical settings. We applied three techniques—Latent Dirichlet Allocation (LDA), Non-negative Matrix Factorization (NMF), and BERTopic—to create models with a fixed number of topics (n=40). These models were then tested on approximately 150,000 news headlines. To assess topic coherence, we utilized Word2Vec, human evaluators, and two large language models. Statistical tests confirmed the significance and impact of our findings. BERTopic demonstrated superior coherence compared to NMF, though slightly, but consistently outperformed NMF and LDA according to human and LLM evaluations. The notable disparity in LDA's performance relative to BERTopic and NMF underscores the importance of carefully selecting a topic modeling technique, as the choice can significantly influence the outcome of the analysis.

In today’s era, social media has become a driving force for increasing digital entrepreneurship. Businesses are utilizing social media sites such as Instagram, TikTok, LinkedIn, or even Facebook to brand their companies or products and interact with clients. This is great news for businesses, especially SMEs, to have low-cost access to key markets worldwide. One evident trend is the emergence of social commerce – business-to-consumer commerce without intermediaries, exclusive of other e-commerce models. However, the adoption of social media in digital entrepreneurship comes with several challenges, such as changes in algorithms that can affect content visibility and risks related to data security and user privacy. Nevertheless, social media remains useful in terms of analytics to support strategic decisions. This study shows the value of social media for entrepreneurship and technologies that help improve content personalization and consumer behavior analysis, such as artificial intelligence and big data. This study attempts to fill the gap in the literature by looking at the differences in the outcomes of social media use in developing and developed countries and the outcomes of new technologies on digital business ventures.

This study presents the development and evaluation of an automatic passenger counting system for public buses using the YOLOv8 algorithm based on Convolutional Neural Networks (CNN). Accurate passenger counting plays a crucial role in optimizing public transportation operations, as it enables effective capacity management, reduces operational costs, and improves overall passenger comfort. Conventional manual counting methods are often inefficient, time-consuming, and prone to human error, particularly in high-density urban transportation environments. Therefore, an automated and intelligent solution is required to support real-time monitoring and operational decision-making. The proposed system employs deep learning-based object detection to identify and count passengers from video streams captured by cameras installed inside buses. Two camera positions, namely front and rear views, were evaluated to assess system performance under different visual conditions. The experimental results show that the system achieves high detection accuracy in the front camera view, with a confidence score of 0.82, indicating reliable performance in scenarios with minimal object occlusion. In contrast, the rear camera view demonstrates slightly lower accuracy, with a confidence score of 0.76, mainly due to increased object overlap and variations in lighting conditions. These findings emphasize the importance of appropriate camera placement and environmental consideration in improving detection reliability. In addition, the implementation of the proposed system enables real-time monitoring of passenger flow, which supports dynamic scheduling, demand-based route planning, and efficient fleet management. Accurate passenger data allows transportation operators to optimize service allocation, reduce congestion, and enhance overall service quality. Overall, this study contributes to the development of intelligent transportation systems by demonstrating the practical applicability of deep learning-based passenger counting solutions. The proposed approach offers strong potential for real-world deployment in smart city environments, supporting the creation of more sustainable, efficient, and passenger-oriented public transportation services.

This research focuses on designing and simulating a makeup robot control system with a Human-Robot Interaction (HRI) approach. The main goal is to develop a robot that is not only efficient in applying makeup, but also able to interact directly with users. The design of this robot is designed with attention to anthropomorphism and non-verbal interaction, in order to improve user comfort and experience during the use process. This control system leverages cutting-edge sensor technology, such as facial recognition and expression analysis, to detect user emotions and adjust the robot's response in real-time. The simulation process is carried out using the Robot Operating System (ROS) to develop an algorithm that supports task coordination between robots and humans as well as interactive feedback. The results of the simulation show that the robot is able to recognize the user's emotions and adjust their actions, thus creating a more intuitive and responsive interaction experience. This research has made a significant contribution to the development of robotics technology in the field of beauty, thereby improving the user experience in personal care. The findings also pave the way for further research into more complex human-robot interactions that are responsive to individual needs.

This study develops an OpenCV-based color recognition system to support individuals with color blindness, aiming to enhance their independence in recognizing colors. The User-Centered Design (UCD) methodology was employed, allowing direct user feedback for development tailored to their needs. The developed system showed significant capability in detecting and identifying colors with a camera and analyzing them using OpenCV. The system provides text output to facilitate use by color-blind individuals and has been tested across various lighting conditions and devices. However, there is still potential for improvement, especially in dealing with fluctuations in extreme lighting conditions and integrating voice output for better auditory assistance. Recommendations for further development include the integration of voice recognition, improvements to the color detection algorithm, and further testing with a diverse user group. This will enhance the system's functionality, accuracy, and accessibility for everyone, particularly those suffering from color blindness.

This research focuses on the development of automatic waste classification by applying the Convolutional Neural Network (CNN) method in a web-based application. This system is designed to help the waste management process through automatic sorting between organic and inorganic waste, so that it can support recycling efforts and reduce environmental impacts. In its application, this application utilizes the CNN algorithm to analyze images and recognize the type of waste with good accuracy. The development uses technologies such as Python and OpenCV to ensure efficient processing of image data, with the CNN model trained using a dataset of 22,564 images. Test results show excellent accuracy, reaching 99.27% for organic waste and 98.72% for inorganic waste.

This study discusses the implementation of computer vision technology for face detection in photos using two sample images with variations in lighting and face pose. The developed system combines the Viola-Jones algorithm and Convolutional Neural Networks (CNN) to enhance resilience against lighting and face orientation variations. Experimental results show high accuracy even with only two sample images. This research also develops preprocessing techniques to handle extreme lighting conditions and demonstrates efficient implementation using Python and OpenCV.  

Artificial Intelligence (AI)plays a vital role in criminal investigations, offering innovative solutions to challenges faced by law enforcement. With its fast and accurate data analysis capabilities, AI can identify behavioral patterns, detect anomalies, and predict potential crimes. Technologies such as facial recognition, social network analysis, and natural language processing help speed up the investigation process and improve prosecution effectiveness. However, the application of AI also raises ethical challenges, including privacy issues and potential bias in algorithms. Therefore, it is important to develop a framework that ensures the responsible use of AI in a legal context.    

This research focuses on the use of hand tracking technology in a drawing program based on the MediaPipe framework. The aim of this study is to develop a digital drawing system that can track hand movements in real-time without additional input devices like a mouse or stylus. This technology utilizes computer vision algorithms to detect and track the user's hand movements, which are then translated into strokes on the screen. The study employs a descriptive-qualitative method with a software experimentation approach. The results show that the system has a high level of accuracy and is responsive to hand movements, providing a more natural and intuitive user experience. The implications of this research are significant in supporting technology-based educational and creative applications.

Human resources are a company's most dominant asset, because they can play an important role in the company's business development. An outsourcing company is a legal entity and is obliged to comply with business licenses issued by the Central Government. The eligibility process for new employees is supported based on the level of ability and competency determined by the company. The existence of difficulties for companies in determining the eligibility of new employees which makes the reason for the ineffectiveness of the processes carried out by the company at this time, is used as a goal for the authors for the purposes of a study. By using the classification method in a data mining with the C4.5 algorithm (Decision Tree) and a RapidMiner application as a tool in the analysis process carried out to find a factor supporting the process of a new employee eligibility. With the data of 960 applicants used as a sample, this data was taken from 2021-2022. From the data divided into several attributes used, the highest Gain value obtained from these attributes through the results of Test 2 of 0.417152421 which will be used as the root in the process of determining employee eligibility and has the highest accuracy value of 98.44%.

This study presents an advanced approach to multi-class skin lesion classification by leveraging an ensemble model comprising the Inception-V3, ResNet-50, and VGG16 architectures. The classification task focuses on categorizing skin lesions into distinct classes, including Melanoma, basal cell carcinoma (BCC), and squamous cell carcinoma (SCC), using the ISIC dataset, a comprehensive collection of dermoscopic images. In order to properly balance the dataset, the oversampling strategy is utilized, as some lesion types are underrepresented due to inherent imbalances in the dataset. By ensuring that the model is trained on a more representative dataset, this balancing improves the algorithm's capacity to categorize all lesion types properly and impartially. By combining the complementary features of ResNet-50, Inception-V3, and VGG16, the ensemble technique improves the overall classification performance. ResNet-50 is chosen for its deep feature extraction capabilities, which help capture fine details in lesion patterns. Inception-V3 is selected for its multi-scale processing, allowing it to effectively analyze lesions at varying resolutions and sizes. VGG16 is included due to its simple yet highly effective architecture for image classification tasks. The ensemble model with data augmentation significantly outperforms individual models in skin lesion classification for both the original and balanced ISIC datasets regarding accuracy, precision, recall, and F1-score. This method offers a robust solution for skin lesion classification, contributing to more accurate and reliable diagnostic tools in dermatology.

In the midst of increasingly rapid technological developments, Artificial Intelligence (AI) technology has also been formed as a form of development. The presence of AI technology helps many people complete their work. Including strategic activities in the decision-making process for business development. This research aims to discuss the role of artificial intelligence technology in improving decision making in the business development process by focusing on the Management Information Systems (SIM), Micro, Small and Medium Enterprises (MSMEs), and finance sectors. With a literature review used as a comprehensive research method, it involves collecting and analyzing articles related to the topic from various academic sources. The research results show that Artificial Intelligence (AI) can improve the efficiency and accuracy of decision making through in-depth analysis and algorithm-based predictions. In the Management Information Systems (MIS) sector, artificial intelligence contributes to business automation processes and information management, in the MSME sector, artificial intelligence helps in understanding consumer behavior and market trends. Meanwhile, in the financial sector, AI plays an important role as a risk analyst, financial manager and investment manager. This research is expected to provide knowledge for corporate organizations about the role of artificial intelligence technology in improving the decision-making process.

KPRI KOKARNABA Baturraden faces challenges in managing increasingly complex sales data, particularly in identifying the most in-demand products to maximize profit. This study aims to analyze sales patterns using the Naïve Bayes algorithm as a probability-based classification method. The collected sales data were analyzed to identify categories of best-selling and less popular products within the cooperative. The results indicate that the Naïve Bayes algorithm has an accuracy rate of 77.56% in predicting product categories. This research is expected to assist the cooperative in optimizing stock management and improving member satisfaction.

The design of a village fund assistance information system using the C4.5 algorithm aims to optimize the selection process for prospective aid recipients at the Ogan Ilir Regency Social Service (DINSOS). Currently, the village fund assistance selection process often takes a long time and is prone to inaccuracy and unfairness due to the limitations of the manual system. The C4.5 algorithm was chosen to build an effective decision tree in classifying based on predetermined criteria, such as income, number of dependents, employment, housing status, and expenses. By utilizing the Gain or Gain Ratio value of each attribute, the C4.5 algorithm is able to produce a clear decision tree, which makes it easier for DINSOS to make decisions objectively and transparently. This information system is designed with an easy-to-use user interface and a structured database to facilitate the management of aid recipient data. The results of the implementation of this system show increased accuracy in determining prospective aid recipients and time efficiency in data processing, thus supporting efforts to evenly distribute village fund assistance in Ogan Ilir Regency in a targeted manner.

Victory in Mobile Legends is influenced by various factors, such as player skills, strategy, and character selection. To predict game outcomes, the backpropagation algorithm is applied to process historical gameplay data and create an accurate predictive model. This study aims to apply the backpropagation algorithm to predict victory based on player attributes, including team role, experience level, and past performance. The research method involves training and testing the model using data from multiple gameplay sessions with varied outcomes. Findings show that the backpropagation algorithm can predict game results with high accuracy, especially when the data includes a more comprehensive range of attributes. The implications of this study suggest that a backpropagation-based predictive model can help players understand their chances of winning and optimize their gameplay strategies. Furthermore, future developments in this algorithm could provide benefits for similar applications in other digital gaming fields.

Credit approval prediction is one of the critical challenges in the financial industry, where the accuracy and efficiency of credit decision-making can significantly affect business risk. This study proposes an outlier detection method using the Gaussian Mixture Model (GMM) combined with Extreme Gradient Boosting (XGBoost) to improve prediction accuracy. GMM is used to detect outliers with a probabilistic approach, allowing for finer-grained anomaly identification compared to distance- or density-based methods. Furthermore, the data cleaned through GMM is processed using XGBoost, a decision tree-based boosting algorithm that efficiently handles complex datasets. This study compares the performance of XGBoost with various outlier detection methods, such as LOF, CBLOF, DBSCAN, IF, and K-Means, as well as various other classification algorithms based on machine learning and deep learning. Experimental results show that the combination of GMM and XGBoost provides the best performance with an accuracy of 95.493%, a recall of 91.650%, and an AUC of 95.145%, outperforming other models in the context of credit approval prediction on an imbalanced dataset. The proposed method has been proven to reduce prediction errors and improve the model's reliability in detecting eligible credit applications.

Foodstuffs are raw materials in the form of agricultural, vegetable and animal products that are used by the food processing industry to produce a food product. Food ingredients consist of plant foods and animal foods. Food is the most basic need for human resources in a country. Food prices sometimes experience erratic increases or decreases. The aim of this research is to determine the results of food price predictions in the Deli Serdang Regency area using the Backpropagation algorithm. The data used in this research is food price data from 2020 to 2023 which comes from the official National Food Ingredients website. This research uses the Backpropagation algorithm artificial neural network method which uses several architectural models and the results of this test will produce the best accuracy values. The test results show that the best architecture for research on implementing the backpropagation algorithm in predicting food prices in Deli Serdang Regency is 2-10-1 with an accuracy of 87.5% and the 2-3-8-1 architecture with an accuracy of 87.5%.

This study investigates optimal retrofit strategies for buildings in tropical climates, focusing on energy efficiency, thermal comfort, and indoor air quality (IAQ). Given the unique challenges of high temperatures, humidity, and energy demands in tropical regions, traditional retrofitting methods often fall short of achieving a balance between these critical factors. By employing a multi-objective optimization approach, this research identifies the most effective combination of retrofit solutions, including insulation, natural ventilation, and high-performance window treatments. The results show that the proposed retrofit strategy significantly reduces cooling energy consumption, while maintaining or improving occupant comfort and IAQ. Insulation, particularly external insulation, proved to be the most effective in reducing heat transfer, while natural ventilation strategies and advanced materials further contributed to improving thermal regulation. The study demonstrates that integrating passive and active retrofit measures, tailored specifically to tropical climates, leads to optimal building performance. The multi-objective optimization algorithm (NSGA-II) allowed for the generation of Pareto-optimal solutions, offering a set of trade-offs between energy efficiency, thermal comfort, and IAQ. These findings are particularly relevant for policymakers and building professionals seeking sustainable retrofit solutions in tropical regions. The study also highlights the importance of integrating energy efficiency and IAQ considerations in retrofit strategies to avoid compromising occupant health. Further research is recommended to explore the integration of advanced materials, such as phase change materials (PCMs), and to enhance IAQ management in retrofitted buildings, ensuring long-term sustainability and occupant well-being in tropical environments.

Blockchain and cryptocurrencies have changed the way we transact and interact in the digital age. However, the rapid advancement of these technologies has resulted in major environmental impacts. Efficient implementation of blockchain requires the use of large amounts of energy and computing power, with consensus algorithms as the foundation. The purpose of this study is to investigate the environmental implications of blockchain and cryptocurrency implementation, as well as initiatives to mitigate these issues. The type of research is a library study (library research) using a qualitative method, namely by combining, collecting information or previous scientific papers on relevant topics. Along with the growing popularity of cryptocurrencies, the continuous mining process often results in large energy consumption and carbon emissions, sparking concerns about their long-term viability and environmental impact. Based on the results of previous research and research sources, the author found a solution to the problem of high energy consumption from the use of blockchain, namely: Proof of Stake (PoS), Proof of Authority (PoA), Sidechains and Layer-2 Solutions, Hardware Optimization, Implementation of Consensus Algorithms Based on RUST.