Sinkron : jurnal dan penelitian teknik informatika
Not a member yet
1261 research outputs found
Sort by
Comparison of Tubercolosis Detection Using CNN Models (AlexNet and ResNet)
No full textThe bacterial infection caused by Mycobacterium tubercolosis, leading to tubercolosis is a prevalent contagious disease. This bacterium commonly targets the primary respiratory organs, particularly the lungs. Tuberculosis poses a significant global health challenge and necessitates early detection for effective management. In this context, to facilitate healthcare professionals in the early detection of patients, a technology capable of accurately identifying lung conditions is required. Therefore, CNN (Convolutional Neural Network) will be employed as the algorithm for detecting lung images. The research will utilize Convolutional Neural Network models, namely AlexNet and ResNet. The study aims to compare the performance of these two models in detecting TB through the analysis of chest X-ray images. The dataset comprises X-rays from both normal patients and TB patients, totaling 4.200 data points. The training process involves dividing the data into training and validation sets, with an 80% allocation for training and 20% for validation. The evaluation results indicate that the AlexNet model demonstrates higher detection accuracy, reaching 88.33% on the validation data, while ResNet achieves 83.10%. These findings suggest that the use of CNN models, especially AlexNet, can be an effective approach to enhancing early tuberculosis detection through the interpretation of chest X-ray images, with potential implications for improving global TB management and prevention efforts
Retail Marketing Strategy Optimization: Customer Segmentation with Artificial Intelligence Integration and K-Means Clustering
No full textThis study aims to optimize retail marketing strategies through customer segmentation using the K-Means clustering method and RFM (Recency, Frequency, Monetary) analysis. By utilizing transaction data from a large retail company, customers are categorized into six segments: VIP Customers, Loyal Customers, Potential Loyalists, New Customers, At-Risk Customers, and Dormant Customers. This segmentation allows for the implementation of more targeted marketing strategies for each customer group. For example, VIP Customers who represent 3.0% of total customers are very active with significant spending, so they deserve exclusive offers and premium services. Loyal Customers, which account for 7.0% of total customers, show high transaction frequency and loyalty, suitable for loyalty programs and recurring discounts. Potential Loyalists, which comprise 15.0%, show the potential for increased loyalty through retention campaigns. New customers representing 16.3% need a brand recognition and promotion strategy to increase their initial engagement. At-Risk Customers covering 30.7% indicated a decrease in transaction activity and required intervention to prevent churn, while Dormant Customers covering 28.1% required a strong reactivation strategy. The clustering evaluation showed an average Silhouette score of 0.3115, which indicates that the clusters that are formed are quite well defined, although there is still room for improvement. This research provides valuable insights to develop more effective and efficient marketing strategies, as well as increase customer satisfaction and loyalty
Sentiment Analysis on KPU Performance Post-2024 Election via YouTube Comments Using BERT
No full textThis research aims to analyze public sentiment regarding the performance of the General Election Commission after the 2024 presidential election using the BERT (Bidirectional Encoder Representations from Transformers) model. Given the General Election Commission's crucial role in maintaining election integrity and the importance of transparency in Indonesian democracy, understanding public opinion through sentiment analysis is essential. Data was collected from YouTube comments, a platform increasingly popular for public expression. The analysis process began with data preprocessing, including case folding, text cleaning, tokenization, and stop word removal. The BERT model was then applied to classify the sentiment of the comments, with the model's performance evaluated using 10-fold cross-validation. The evaluation results showed that the first fold (k=1) achieved the best performance with an accuracy of 96%, precision of 96%, recall of 96%, and an F1-score of 96%, indicating the model's effectiveness in accurately classifying sentiment. In contrast, the ninth fold (k=9) exhibited the lowest accuracy at 86% with other metrics also lower, suggesting performance instability potentially caused by data variability. Accuracy and loss graphs confirmed that the first fold experienced consistent accuracy improvements and significant loss reduction, while the ninth fold showed performance fluctuations. This study provides valuable insights into public sentiment regarding the General Election Commission performance, with BERT demonstrating significant potential for sentiment analysis on social media platforms like YouTube
Implementation Docker and Kubernetes Scaling Using Horizontal Scaler Method for Wordpress Services
No full textContainer is a technology that has recently been widely used because of the additional features that are very easy and convenient to use, especially for web hosting service developers, with Container making it easier for system admins to manage applications including building, processing and running applications on Container. With Container the process of creating and using the system will be easier but along with too many user requests so that the service does not run optimally. Therefore, the Container must have good scalability and performance. Scalability is needed for systems that can adjust to the needs of user demand and performance is needed to maintain the quality of services provided. This research aims to implement scaling using Docker and Kubernetes in terms of scalability and performance. The parameters of comparison between Docker and Kubernetes are for scalability, scaling up and scaling down time and for performance. The method in this research uses the Action Research methodology, which is a research model that is simultaneously practiced and theorized. With the initial steps of problem identification, action planning, action implementation, observation and evaluation. Based on the results that have been obtained, Docker consumes more CPU & Memory Usage Resources, namely at 500 Users Kubernetes consumes Resources with an average of 94.47%-4.70% while in Kubernetes 89.11%-4.50 because in Kubernetes itself has a complex system, especially special component components such as APIs, Metrics Server, Kubernetes manager to run the Container. While in Docker only has Docker Manager and Docker Compose components
Sentiment Analysis on BNI Mobile Application Review Using K- Nearest Neighbors Algorithm
No full textAdvances in Science and Technology continue to evolve in response to the demands of modern times, particularly in various fields such as banking. The development of information technology has transformed the way transactions are conducted from traditional to digital, accessible flexibly through Mobile Banking. BNI has created the BNI Mobile Banking application to facilitate customers in their transactions. The objective of this study is to investigate how the use of BNI Mobile can influence the ease of customers in conducting transactions. The data collection method used in this study is the K-Nearest Neighbors method, focusing on user experience with the BNI Mobile Banking applicatio
Optimizing the Blood Donation App with Gamification Using User-Centered Design
No full textIn today's digital era, motivating the younger generation to participate in routine voluntary blood donations is a significant challenge in the health sector. This research aimed to develop a gamified application called Gamified Blood Donation (G-BlooD), designed using the User Centered Design methodology. This application integrates gamification into the blood donation process, with features including donor location information, available blood stock data, and individual donation history. Using gamification elements such as challenges, ranking boards, and emblems enhanced user interactivity and motivation. Evaluation of G-BlooD demonstrated its effectiveness in achieving this goal; it scored 75 (Grade B) on the System Usability Scale (SUS), indicating good usability, while an average total index calculation from all responses on the Likert scale of 84.125% underscored its success in motivating younger generations towards regular blood donations. These results suggest combining digital technology with gamification can encourage recurring voluntary blood donation among younger generations. This research opens avenues for further exploration into leveraging digital technology to address other public health concerns
OPTIMIZATION NAÏVE BAYES ALGORITHM IN SENTIMENT ANALYSIS OF BUKALAPAK APP REVIEWS
No full textBukalapak application reviews on Google Play Store include useful information if processed correctly. The activity of analyzing application reviews is not enough to see the number of stars, it is necessary to see the entire contents of the review comments to be able to know the intent of the review. Sentiment analysis system is a system used to automatically analyze reviews. Review data is retrieved via the bukalapak application API and then classified using Naive Bayes Multinomial. A total of 1,000 reviews of bukalapak application users were collected to be used as dataset samples. The purpose of this research is to determine the accuracy level of sentiment analysis using the multinomial Naive Bayes algorithm. The stages of this research include, data collection, automatic labeling using python, pre-processing, sentiment classification, and evaluation. In the pre-processing stage there are 6 stages, namely Cleaning, Casefolding, Word Normalizer, Tokenizing, Stopword Removal and Stemming. TF-IDF (Term Frequency - Inverse Document Frequency) method is used for word weighting. The data will be grouped into two categories, namely negative and positive. The test results show an accuracy value of 90%, this result shows that the bukalapak application reviews tend to be negative. The research at this time only looks for accuracy values and provides an overview of the bulapak application to potential new users
Application of Data Mining for Clustering Human Development Index Based on West Java Province 2017-2022
No full textHuman development is used as a parameter to see development from the human side. The Human Development Index (HDI) explains how people get sufficient income, adequate health and education. Geographically, Indonesia is an archipelagic country where each province is spread across various islands separated by sea. Making the disparity in human development between provinces relatively high. The gap that occurs is still a problem that must be resolved immediately, because the gap in the human development index can hamper the government's goal of equalizing human welfare in Indonesia. −One of the problems related to population that West Java Province still has to face is the problem of imbalance in population distribution. Incomplete population distribution causes problems with population density and population pressure in an area. This research uses data sources from the West Java Province Central Statistics Agency (BPS). The data used in this research is data from 2017-2022 which consists of 27 regencies and cities of West Java Province. Therefore, researchers utilized the K-Means algorithm in clustering 27 Regencies and Cities of West Java Province. The data will be processed by clustering into 3 clusters, namely the high population area level cluster, the medium population area level cluster and the low population area level cluster. This research classifies population density using Ms. software. Excel and RapidMiner. The iteration process took place 3 times so that the results obtained were 8 regencies and cities with high population area clusters (C0), 0 regencies and cities with medium population area clusters (C1) and 19 regencies and cities with low population area clusters (C1). C2)
Enhancing Least Significant Bit Steganography Image Fidelity Using Brotli Compression
No full textThe rapid growth of technology has provided extensive convenience and openness in accessing information, yet this hasn't been balanced with an equivalent enhancement in information security. Steganography plays a crucial role in concealing and protecting data, with the Least Significant Bit (LSB) method being a commonly used algorithm that operates by substituting the least significant bits in the image pixels with the bits of the data to be hidden, aiming to preserve the image quality. This research aims to enhance the quality of the steganography result using LSB by employing the Brotli compression technique, coupled with increasing image's capacity. Brotli compression aims to reduce the size of the resulting stego image by combining embedded data that share identical values. Experiment results will be obtained by comparing the original image with the stego image using metrics like Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR), and Structural Similarity Index (SSIM). The experiments successfully demonstrated that the integration of LSB with Brotli compression outperformed the regular LSB method, showing a 6.64% increase in PSNR, followed by a decrease in MSE by approximately 63.79%, and an increase in SSIM by around 0.0039%. This was accompanied by a continuous increase in compression values depending on the input data size. These results indicate that the integration of LSB with Brotli compression was successfully implemented to enhance the fidelity of the stego imag
Optimizing Digital Image Steganography through Hybridization of LSB and Zstandard Compression
No full textIn response to the growing need for secure digital communication globally, this research delves into an innovative strategy for enhancing data transmission security through steganography. This inventive approach involves the integration of the conventional Least Significant Bit (LSB) method with Zstandard (Zstd) compression to elevate the quality of stego images. The study carefully explores how the synergistic use of LSB and Zstd contributes to an improved equilibrium between embedding capacity and visual quality in stego images. This hybrid methodology capitalizes on the efficiency of Zstd in reducing file size, thereby facilitating more effective data concealment using LSB. The experimental outcomes showcase a notable 51.2% increase in embedding capacity, a 4.70% elevation in PSNR value, accompanied by a substantial 51.03% decrease in MSE value. Additionally, SSIM values hover around 0.007%, indicating a perceptually minimal difference between the original and steganographically modified images. These compelling results underscore the efficacy of the proposed method, highlighting its proficiency in preserving and enhancing the quality of stego images generated through the embedding process. This research signifies a significant stride in the realm of secure digital communication, demonstrating a promising fusion of traditional LSB with advanced Zstd compression for optimizing digital image steganographic