Jurnal Politeknik Negeri Batam (PoliBatam)
Not a member yet
3001 research outputs found
Sort by
Enhancing Web Security and Performance with Hybrid Stateless Authentication
No full textEnsuring operational integrity across industries and protecting sensitive data require strong authentication systems. This paper presents a novel hybrid stateless authentication method that integrates binary payloads, token specifications, and database solutions. By employing a distinctive expiration policy, our proposed approach overcomes limitations inherent in traditional token revocation strategies while achieving token verification speeds that are up to 86 times faster than conventional statefull session-based methods. Overall, through uniformed benchmarking experiments and a comprehensive review of the literature substantiate the performance and security advantages of our method. Ultimately, this hybrid technique offers a more scalable and secure framework for authentication management, enabling efficient and flexible deployment in high-demand distributed environments
Optimization of Random Forest Algorithm with Backward Elimination Method in Classification of Academic Stress Levels
No full textStress is a phenomenon experienced by all individuals as a natural response to pressure, which can impact mental and physical health. In an academic setting, the stress experienced by students is known as academic stress, which can affect their performance and mental well-being. Therefore, there is a need for effective prediction methods to aid in the management and prevention of academic stress. Therefore, there is a need to predict the level of academic stress to aid more effective management and prevention. This study uses a public dataset categorized based on the Student-life Stress Inventory (SSI), which includes psychological, physiological, social, environmental, and academic factors. Data mining is often used to detect diseases, one of which is the Random Forest algorithm. The Random Forest algorithm is applied as a classification technique for academic stress levels, with optimization using the Backward Elimination method for feature selection to improve model accuracy. The results showed that the accuracy of the Random Forest algorithm without feature selection obtained an accuracy of 86%, compared to the random forest algorithm with feature selection using the Backward Elimination method obtained a higher accuracy of 88%. This increase shows that the feature selection method can optimize model performance by selecting more relevant features. Thus, this research is expected to contribute to the management of student academic stress against the risk of academic stress
Web-Based Makeup Recommendation System Using Hybrid Filtering
No full textThe increasing use of makeup products in the modern era, driven by evolving beauty trends and e-commerce accessibility, presents challenges in selecting products suited to individual skin types and conditions. A recommendation system addresses this issue by enhancing selection efficiency. This study explores the implementation of Content-Based Filtering (CBF) using TF-IDF and Cosine Similarity, Collaborative Filtering (CF) with Singular Value Decomposition (SVD), and a Hybrid Filtering approach integrating both methods through Weighted Hybrid techniques. The system\u27s performance is evaluated across two user scenarios: new users (without prior ratings) and old users (with rating history). The evaluation method includes Precision, Normalized Discounted Cumulative Gain (NDCG), and accumulation of the best scenario based on user opinion. Results show that Hybrid Filtering outperforms CBF and CF, with notable differences between user groups. For new users, 32% prefer Scenario 1, which emphasizes CBF, achieving 80.8% Precision and 89.73% NDCG. For old users, 23% favor Scenario 2, attaining 83.4% Precision and 90.31% NDCG
Hamming Code in JPEG Image Steganography within the Discrete Cosine Transform Domain
No full textThis study proposes a novel JPEG image steganography method that combines Least Significant Bit (LSB) embedding in the Discrete Cosine Transform (DCT) domain with Hamming Code (2k − 1, 2k − k − 1) to minimize the number of modified DCT coefficients. Experiments were conducted on images with varying resolutions (512×512, 1024×1024, 2048×2048) and JPEG quality factor of 75, where PSNR (Peak Signal to Noise Ratio) and SSIM (Structural Similarity Index) parameters are used to measure the quality and similarity between the original image and the stego image. The method achieved an embedding capacity of up to 524,288 bits, with an average PSNR of 39–41 dB and SSIM above 0.98. Compared to conventional techniques such as JSteg and F5, the proposed approach demonstrates improved embedding capacity, better visual quality, and higher resistance to statistical steganalysis, making it suitable for secure and efficient data hiding applications.This study proposes a novel JPEG image steganography method that combines Least Significant Bit (LSB) embedding in the Discrete Cosine Transform (DCT) domain with Hamming Code (2k − 1, 2k − k − 1) to minimize the number of modified DCT coefficients. Experiments were conducted on images with varying resolutions (512×512, 1024×1024, 2048×2048) and JPEG quality factor of 75, where PSNR (Peak Signal to Noise Ratio) and SSIM (Structural Similarity Index) parameters are used to measure the quality and similarity between the original image and the stego image. The method achieved an embedding capacity of up to 524,288 bits, with an average PSNR of 39–41 dB and SSIM above 0.98. Compared to conventional techniques such as JSteg and F5, the proposed approach demonstrates improved embedding capacity, better visual quality, and higher resistance to statistical steganalysis, making it suitable for secure and efficient data hiding applications
Sentiment Analysis of Public Comments on X Social Media Related to Israeli Product Boycotts Using The Long Short-Term Memory (LSTM) Method
No full textThe boycott of Israeli products is a widely discussed issue on social media, particularly on X. This study aims to analyze public sentiment regarding the boycott using the Long Short-Term Memory (LSTM) method. Data was collected via the X API, resulting in 800 comments after cleaning and removing duplicates from initially 980 crawled datasets. LSTM was chosen for this analysis due to its superior ability to process sequential data like text and effectively capture long-term dependencies in natural language, which is crucial for accurate sentiment classification. Data was processed through preprocessing steps, sentiment labeling, and Term Frequency-Inverse Document Frequency (TF-IDF) weighting before being fed into the LSTM model. Sentiment was classified into three categories: positive, negative, and neutral. Model evaluation was conducted using accuracy, precision, recall, and F1-score metrics. The results show that the LSTM model achieved an accuracy of 80.62%, with negative sentiment dominating, followed by neutral and positive. This study demonstrates that the LSTM method effectively classifies public sentiment and can be applied to inform public policy decisions, map public opinion trends, and monitor responses to foreign policy issues related to the Israeli-Palestinian conflict.The boycott of Israeli products is a widely discussed issue on social media, particularly on X. This study aims to analyze public sentiment regarding the boycott using the Long Short-Term Memory (LSTM) method. Data was collected via the X API, resulting in 800 comments after cleaning and removing duplicates from initially 980 crawled datasets. LSTM was chosen for this analysis due to its superior ability to process sequential data like text and effectively capture long-term dependencies in natural language, which is crucial for accurate sentiment classification. Data was processed through preprocessing steps, sentiment labeling, and Term Frequency-Inverse Document Frequency (TF-IDF) weighting before being fed into the LSTM model. Sentiment was classified into three categories: positive, negative, and neutral. Model evaluation was conducted using accuracy, precision, recall, and F1-score metrics. The results show that the LSTM model achieved an accuracy of 80.62%, with negative sentiment dominating, followed by neutral and positive. This study demonstrates that the LSTM method effectively classifies public sentiment and can be applied to inform public policy decisions, map public opinion trends, and monitor responses to foreign policy issues related to the Israeli-Palestinian conflict
Analysis of docker container Implementation in SIEM infrastructure
No full textIt is known that configuring system information and event management (SIEM) infrastructure using conventional virtualization still provides essential functions. However, if a problem occurs such as a configuration error during the staging process or application service failure, the recovery process from the error requires quite a long time. This research aims to explore and analyze the implementation of container technology in the SIEM Infrastructure using the Wazuh platform. The analysis focuses on a Docker-based architecture running Wazuh\u27s core components: the wazuh-indexer, wazuh-manager, and wazuh-dashboard, each in its own container. This approach is evaluated to see how containerization affects SIEM effectiveness and efficiency, particularly in resource utilization and fault recovery. Performance testing carried out on systems using Docker Containers shows lower Memory and CPU usage compared to Conventional Virtualization. The results demonstrate that Docker not only enhances resource efficiency but also improves system resilience, directly impacting SIEM operational functionality
Comparative Study of SVM, KNN, and Naïve Bayes for Sentiment Analysis of Religious Application Reviews
No full textThis study aims to evaluate and compare the performance of three machine learning algorithms—Support Vector Machine (SVM), K-Nearest Neighbors (K-NN), and Naïve Bayes—for sentiment classification of user reviews on the NU Online application in the Google Play Store. NU Online is a religious digital platform providing Islamic content such as articles, prayers, and worship schedules. A total of 1,500 user reviews were collected using web scraping, and 1,491 were retained after data cleaning. Preprocessing steps included punctuation removal, case folding, normalization, stopword removal, stemming, and tokenization. Sentiment labels (positive or negative) were automatically assigned using a lexicon-based approach. The performance of the models was assessed using accuracy, precision, recall, and F1-score, calculated via confusion matrix with a training-testing data split. The results show that the SVM with a linear kernel achieved the best accuracy (81.6%), followed by Naïve Bayes (73.2%) and K-NN (66.9%). These findings indicate that SVM is the most effective algorithm in this context, providing practical contributions for developers of the NU Online digital religious platform and contributing to research in Indonesian natural language processing.This study aims to evaluate and compare the performance of three machine learning algorithms—Support Vector Machine (SVM), K-Nearest Neighbors (K-NN), and Naïve Bayes—for sentiment classification of user reviews on the NU Online application in the Google Play Store. NU Online is a religious digital platform providing Islamic content such as articles, prayers, and worship schedules. A total of 1,500 user reviews were collected using web scraping, and 1,491 were retained after data cleaning. Preprocessing steps included punctuation removal, case folding, normalization, stopword removal, stemming, and tokenization. Sentiment labels (positive or negative) were automatically assigned using a lexicon-based approach. The performance of the models was assessed using accuracy, precision, recall, and F1-score, calculated via confusion matrix with a training-testing data split. The results show that the SVM with a linear kernel achieved the best accuracy (81.6%), followed by Naïve Bayes (73.2%) and K-NN (66.9%). These findings indicate that SVM is the most effective algorithm in this context, providing practical contributions for developers of the NU Online digital religious platform and contributing to research in Indonesian natural language processing
Sentiment Analysis on the Relocation of the National Capital (IKN) on Social Media X Using Naive Bayes and K-Nearest Neighbor (KNN) Methods
No full textThis study investigates public sentiment toward the relocation of Indonesia’s capital from Jakarta to East Kalimantan, focusing on reactions from social media platforms such as X (formerly Twitter). Understanding these sentiments is crucial for the government to gauge support for this significant policy shift. The study compares the performance of two classification algorithms, Naïve Bayes and K-Nearest Neighbor (K-NN), in sentiment analysis. A total of 1.277 comments were collected using the tweet-harvest library through a crawling process. The data underwent preprocessing, including cleaning, case folding, normalization, stopword removal, tokenization, and stemming. Sentiment labels were assigned through both manual and automated methods, while feature extraction was performed using the TF-IDF technique. The algorithms\u27 performance was assessed using accuracy, precision, recall, and F1-score metrics. The results revealed that Naïve Bayes outperformed K-NN, with an accuracy of 70%, precision of 72%, recall of 70%, and an F1-score of 69%. In contrast, K-NN achieved an accuracy of 60%, precision of 62%, recall of 60%, and an F1-score of 59%. These results suggest that Naïve Bayes is more effective in classifying sentiment related to the capital relocation. The findings offer valuable insights for policymakers and highlight the potential of automated sentiment analysis as a tool for monitoring public opinion on major governmental policies.This study investigates public sentiment toward the relocation of Indonesia’s capital from Jakarta to East Kalimantan, focusing on reactions from social media platforms such as X (formerly Twitter). Understanding these sentiments is crucial for the government to gauge support for this significant policy shift. The study compares the performance of two classification algorithms, Naïve Bayes and K-Nearest Neighbor (K-NN), in sentiment analysis. A total of 810 comments were collected using the tweet-harvest library through a crawling process. The data underwent preprocessing, including cleaning, case folding, normalization, stopword removal, tokenization, and stemming. Sentiment labels were assigned through both manual and automated methods, while feature extraction was performed using the TF-IDF technique. The algorithms\u27 performance was assessed using accuracy, precision, recall, and F1-score metrics. The results revealed that Naïve Bayes outperformed K-NN, with an accuracy of 73%, precision of 76%, recall of 73%, and an F1-score of 70%. In contrast, K-NN achieved an accuracy of 66%, precision of 67%, recall of 66%, and an F1-score of 63%. These results suggest that Naïve Bayes is more effective in classifying sentiment related to the capital relocation. The findings offer valuable insights for policymakers and highlight the potential of automated sentiment analysis as a tool for monitoring public opinion on major governmental policies
Performance of Machine Learning Algorithms on Imbalanced Sentiment Datasets Without Balancing Techniques
No full textThis study explores the performance of five sentiment classification algorithms—Naïve Bayes, Logistic Regression, Support Vector Machine, Decision Tree, and Random Forest—on an imbalanced sentiment dataset, with the SMOTE technique applied as a comparison. The research follows the Knowledge Discovery in Databases (KDD) framework, which includes data selection, preprocessing, transformation, data mining, and evaluation. The evaluation uses metrics such as accuracy, precision, recall, F1-score, and macro average F1-score. Initial results show that all five algorithms performed fairly well even without using a balancing technique, with Naïve Bayes achieving the highest F1-score of 0.84 and recall of 0.81. After applying SMOTE, only small improvements were observed in some models, such as Random Forest (F1-score increased from 0.81 to 0.85), while other models like Naïve Bayes experienced a decrease in performance, dropping to 0.77. This suggests that the effect of balancing techniques like SMOTE can vary depending on the algorithm. Thus, this study provides empirical contributions that highlight the importance of selecting appropriate approaches and the need for a deep understanding of each algorithm\u27s behavior in the context of imbalanced data. Researchers are encouraged to carefully consider these aspects when designing experiments and interpreting results.This study explores the performance of five sentiment classification algorithms—Naïve Bayes, Logistic Regression, Support Vector Machine, Decision Tree, and Random Forest—on an imbalanced sentiment dataset, with the SMOTE technique applied as a comparison. The research follows the Knowledge Discovery in Databases (KDD) framework, which includes data selection, preprocessing, transformation, data mining, and evaluation. The evaluation uses metrics such as accuracy, precision, recall, F1-score, and macro average F1-score. Initial results show that all five algorithms performed fairly well even without using a balancing technique, with Naïve Bayes achieving the highest F1-score of 0.84 and recall of 0.81. After applying SMOTE, only small improvements were observed in some models, such as Random Forest (F1-score increased from 0.81 to 0.85), while other models like Naïve Bayes experienced a decrease in performance, dropping to 0.77. This suggests that the effect of balancing techniques like SMOTE can vary depending on the algorithm. Thus, this study provides empirical contributions that highlight the importance of selecting appropriate approaches and the need for a deep understanding of each algorithm\u27s behavior in the context of imbalanced data. Researchers are encouraged to carefully consider these aspects when designing experiments and interpreting results
An Analysis of the Relationship between Job Autonomy, Feedback, and Empowering Leadership on Innovative Work Behavior: The Role of Resilience as a Mediating Factor
No full textThis research examines the relationship between job autonomy, feedback, empowering leadership, and innovative work behavior, with resilience as a mediator. Job autonomy refers to employees\u27 freedom in performing tasks, while feedback provides performance-related information. Empowering leadership encourages employees to take initiative and make independent decisions. A quantitative approach was used, and data was collected through questionnaires from employees with over one year of experience in manufacturing and shipyard companies in Batam. Structural Equation Modeling (SEM) analyzed direct and mediating relationships. Findings indicate that job autonomy and feedback significantly enhance innovative work behavior while empowering leadership has a positive but insignificant effect. Resilience mediates these relationships, meaning employees with autonomy, constructive feedback, and empowering leaders are more innovative, especially if they have high resilience. This study offers practical insights for organizations to foster innovation by enhancing job autonomy, providing effective feedback, implementing empowering leadership, and strengthening employee resilience programs