Jurnal Politeknik Negeri Batam (PoliBatam)
Not a member yet
3001 research outputs found
Sort by
Implementation of Braille-Mobile Device to Help Visually and Speech-Impaired Persons Communicate Based on the Blynk IoT
No full textVisual and speech impairment is a condition in which an individual is unable to communicate with family or society due to the inability to see and speak. The objective of this study is to develop a Braille-Mobile device that assists individuals with visual and speech disabilities in communicating remotely with family members or society using Internet of Things (IoT) technology. In this study, the method used to generate messages from the Braille-Mobile device is based on the combination of six buttons pressed on the device, which are translated into letters using the Braille code concept. The messages are then transmitted via the Blynk IoT platform from the Braille-Mobile device to the mobile devices of family members or society through the Internet network. The results of this study show that the developed Braille-Mobile device can be used to send messages in the form of the words HELP, EAT, DRINK, and DRUG to family smartphones using IoT technology with a success rate of up to 76.25% and a message transmission time ranging from 4 to 8 seconds. Furthermore, the Braille-Mobile device is also capable of receiving confirmation from family smartphones in the form of voice responses
Abstract Syntax Tree Model for Minimizing False Negative in Semantic Evaluation of Python Fill-in-the-Blank
No full textThis study develops and evaluates an automated assessment model using Abstract Syntax Trees (AST) with a view to overcoming the limitations of string-matching techniques in the assessment of Fill-in-the-Blank (FIB) programming answers. Traditional string-matching techniques have a relatively high False Negative Rate (FNR) of 21.5% within the context of detecting semantic equivalence. The current model uses semantic structural triangulation to ascertain the semantic similarity of student answers. Technical assessment shows that the AST approach markedly reduces the FNR to 4.5%. The model demonstrates high reliability (ϰ = 0.83) with high classification accuracy (F1 Score = 0.966) which attests to its inferential validity. From a pedagogical perspective, system implementation leads to substantial learning gains, evidenced by a large effect size (Cohen’s d = 1.82) and a high normalized gain (Normalized Gain = 0.90). Multiple regression analysis confirms that semantic accuracy is the primary causal factor driving improved student comprehension. Ontologically, while AST is valid as a partial representation, its limitations—particularly tree isomorphism in recursive structures—highlight the need for further exploration of graph isomorphism approaches. Control Flow Graphs (CFG) and Data Flow Graphs (DFG) offer more expressive relational models for capturing control and data dependencies. The model demonstrates functional feasibility with a System Usability Scale (SUS) score of 76.47. Overall, the AST Triangulation Model is validated as pedagogically effective, inferentially robust, and supportive of evaluative transparency. Future research recommends validating the model on more complex tasks and releasing it as open-source to support reproducibility.This study develops and evaluates an automated assessment model using Abstract Syntax Trees (AST) with a view to overcoming the limitations of string-matching techniques in the assessment of Fill-in-the-Blank (FIB) programming answers. Traditional string-matching techniques have a relatively high False Negative Rate (FNR) of 21.5% within the context of detecting semantic equivalence. The current model uses semantic structural triangulation to ascertain the semantic similarity of student answers. Technical assessment shows that the AST approach markedly reduces the FNR to 4.5%. The model demonstrates high reliability (ϰ = 0.83) with high classification accuracy (F1 Score = 0.966) which attests to its inferential validity. From a pedagogical perspective, system implementation leads to substantial learning gains, evidenced by a large effect size (Cohen’s d = 1.82) and a high normalized gain (Normalized Gain = 0.90). Multiple regression analysis confirms that semantic accuracy is the primary causal factor driving improved student comprehension. Ontologically, while AST is valid as a partial representation, its limitations—particularly tree isomorphism in recursive structures—highlight the need for further exploration of graph isomorphism approaches. Control Flow Graphs (CFG) and Data Flow Graphs (DFG) offer more expressive relational models for capturing control and data dependencies. The model demonstrates functional feasibility with a System Usability Scale (SUS) score of 76.47. Overall, the AST Triangulation Model is validated as pedagogically effective, inferentially robust, and supportive of evaluative transparency. Future research recommends validating the model on more complex tasks and releasing it as open-source to support reproducibility
Proboboost: A Hybrid Model for Sentiment Analysis of Kitabisa Reviews
No full textThe rapid advancement of digital technology has significantly transformed public behavior in social activities, particularly in online donations and zakat payments. The Kitabisa application was selected in this study not only for its popularity but also due to its high user engagement and large volume of reviews on the Google Play Store, making it an ideal representation of public trust in Indonesia’s digital philanthropy ecosystem. This research aims to analyze user sentiment toward the Kitabisa application using a hybrid Proboboost model, which combines Multinomial Naive Bayes (MNB) and Gradient Boosting Classifier through a soft voting mechanism. The model is designed to address class imbalance and improve accuracy in short-text sentiment analysis for the Indonesian language. The study employed preprocessing techniques including case folding, text cleaning, stopword removal, and stemming using the Sastrawi algorithm. Feature extraction was performed using TF-IDF, with an 80:20 train-test split and 5-fold cross-validation to ensure model reliability. Experimental results indicate that the Proboboost model achieved an accuracy of 89.51% and an F1-score of 87.4%, outperforming the Naive Bayes baseline with 87.98% accuracy. The sentiment distribution demonstrates a dominance of positive sentiment (87.24%), followed by negative (8.53%) and neutral (4.23%) reviews. These findings suggest that users generally express satisfaction and trust toward the Kitabisa platform. The results also confirm that the hybrid Proboboost model effectively balances classification performance between majority and minority sentiment classes, offering deeper insights into user perceptions of digital philanthropic services.The rapid advancement of digital technology has significantly transformed public behavior in social activities, particularly in online donations and zakat payments. The Kitabisa application was selected in this study not only for its popularity but also due to its high user engagement and large volume of reviews on the Google Play Store, making it an ideal representation of public trust in Indonesia’s digital philanthropy ecosystem. This research aims to analyze user sentiment toward the Kitabisa application using a hybrid Proboboost model, which combines Multinomial Naive Bayes (MNB) and Gradient Boosting Classifier through a soft voting mechanism. The model is designed to address class imbalance and improve accuracy in short-text sentiment analysis for the Indonesian language. The study employed preprocessing techniques including case folding, text cleaning, stopword removal, and stemming using the Sastrawi algorithm. Feature extraction was performed using TF-IDF, with an 80:20 train-test split and 5-fold cross-validation to ensure model reliability. Experimental results indicate that the Proboboost model achieved an accuracy of 89.51% and an F1-score of 87.4%, outperforming the Naive Bayes baseline with 87.98% accuracy. The sentiment distribution demonstrates a dominance of positive sentiment (87.24%), followed by negative (8.53%) and neutral (4.23%) reviews. These findings suggest that users generally express satisfaction and trust toward the Kitabisa platform. The results also confirm that the hybrid Proboboost model effectively balances classification performance between majority and minority sentiment classes, offering deeper insights into user perceptions of digital philanthropic services
Comparative Analysis of Random Forest, SVM, and Naive Bayes for Cardiovascular Disease Prediction
No full textCardiovascular disease is one of the leading causes of death worldwide; therefore, accurate early detection is essential to reduce fatal risks. This study aims to compare the performance of three machine learning algorithms — Random Forest, Support Vector Machine (SVM), and Naïve Bayes — in predicting cardiovascular disease risk using the Mendeley Cardiovascular Disease Dataset, which contains 1,000 patient records and 14 clinical attributes. The models were evaluated using accuracy, precision, recall, and F1-score metrics, and their performance differences were statistically tested using the paired t-test. The experimental results indicate that the Random Forest algorithm achieved the best performance with 99% accuracy, 100% recall, 98% precision, and an F1-score of 99%. The SVM model followed with 98% accuracy and 100% recall, while the Naïve Bayes algorithm obtained 94.5% accuracy and an F1-score of 95%. The p-value < 0.05 confirmed that the performance differences among the three models were statistically significant. From a clinical perspective, a model with high recall, such as Random Forest, is more desirable because it reduces the likelihood of false negatives, which are critical in heart disease diagnosis. The feature importance analysis also revealed that age, resting blood pressure, and cholesterol level were the most influential factors in predicting cardiovascular risk. These findings suggest that machine learning algorithms, particularly Random Forest, have strong potential to be implemented in Clinical Decision Support Systems (CDSS) for accurate and efficient early detection of cardiovascular disease.Cardiovascular disease is one of the leading causes of death worldwide; therefore, accurate early detection is essential to reduce fatal risks. This study aims to compare the performance of three machine learning algorithms — Random Forest, Support Vector Machine (SVM), and Naïve Bayes — in predicting cardiovascular disease risk using the Mendeley Cardiovascular Disease Dataset, which contains 1,000 patient records and 14 clinical attributes. The models were evaluated using accuracy, precision, recall, and F1-score metrics, and their performance differences were statistically tested using the paired t-test. The experimental results indicate that the Random Forest algorithm achieved the best performance with 99% accuracy, 100% recall, 98% precision, and an F1-score of 99%. The SVM model followed with 98% accuracy and 100% recall, while the Naïve Bayes algorithm obtained 94.5% accuracy and an F1-score of 95%. The p-value < 0.05 confirmed that the performance differences among the three models were statistically significant. From a clinical perspective, a model with high recall, such as Random Forest, is more desirable because it reduces the likelihood of false negatives, which are critical in heart disease diagnosis. The feature importance analysis also revealed that age, resting blood pressure, and cholesterol level were the most influential factors in predicting cardiovascular risk. These findings suggest that machine learning algorithms, particularly Random Forest, have strong potential to be implemented in Clinical Decision Support Systems (CDSS) for accurate and efficient early detection of cardiovascular disease
Evaluating the Impact of Random Over Sampling on IndoBERT Performance for Indonesian Sentiment Analysis
No full textSentiment analysis is a prominent research area in natural language processing (NLP). For the Indonesian language, IndoBERT has emerged as a leading model due to its competitive performance. However, its effectiveness is strongly influenced by balanced class distribution. A common challenge arises because user reviews on digital platforms, such as the Google Play Store, often exhibit imbalanced classes. This study investigates the effectiveness of the Random Over Sampler (ROS) technique in improving IndoBERT’s performance under imbalanced data conditions. The dataset consists of 13,821 user reviews of the IDN App collected from the Google Play Store between 2015 and July 2025. Prior to modeling, data preprocessing was performed, including punctuation removal, case folding, stopword removal, tokenizing, normalization, and stemming to ensure textual consistency. Reviews were categorized into two sentiment classes: positive (3–5 stars) and negative (1–2 stars). Two experimental scenarios were conducted: (1) IndoBERT without ROS and (2) IndoBERT with a balanced dataset using ROS. Model performance was evaluated using accuracy, precision, recall, and F1-score, with data split into 70% training, 20% validation, and 10% testing. Results showed a significant improvement after ROS implementation: 94.55% accuracy, 94.61% precision, 94.53% recall, and 94.54% F1-score. Confusion matrix analysis indicated improved classification of the minority class, reducing the error rate by 49%. However, learning curve analysis revealed potential overfitting due to ROS. Further research is needed to optimize ROS strategies for better performance and generalization.Sentiment analysis is a prominent research area in natural language processing (NLP). For the Indonesian language, IndoBERT has emerged as a leading model due to its competitive performance. However, its effectiveness is strongly influenced by balanced class distribution. A common challenge arises because user reviews on digital platforms, such as the Google Play Store, often exhibit imbalanced classes. This study investigates the effectiveness of the Random Over Sampler (ROS) technique in improving IndoBERT’s performance under imbalanced data conditions. The dataset consists of 13,821 user reviews of the IDN App collected from the Google Play Store between 2015 and July 2025. Prior to modeling, data preprocessing was performed, including punctuation removal, case folding, stopword removal, tokenizing, normalization, and stemming to ensure textual consistency. Reviews were categorized into two sentiment classes: positive (3–5 stars) and negative (1–2 stars). Two experimental scenarios were conducted: (1) IndoBERT without ROS and (2) IndoBERT with a balanced dataset using ROS. Model performance was evaluated using accuracy, precision, recall, and F1-score, with data split into 70% training, 20% validation, and 10% testing. Results showed a significant improvement after ROS implementation: 94.55% accuracy, 94.61% precision, 94.53% recall, and 94.54% F1-score. Confusion matrix analysis indicated improved classification of the minority class, reducing the error rate by 49%. However, learning curve analysis revealed potential overfitting due to ROS. Further research is needed to optimize ROS strategies for better performance and generalization
The Role of Profitability in Moderating the Relationship between Sustainability Reporting and ESG Disclosure onFirm Value of Energy Sector in Indonesia
No full textThis study aims to test and analyze the effects of Sustainability Reporting and ESG Disclosure on firm Value, moderated by profitability, among energy sector companies listed on the Indonesia Stock Exchange in 2019-2023. The research is associative and quantitative. The population of this study comprised 87 companies listed on the Indonesia Stock Exchange, which were purposively sampled, yielding 11 companies in the 2019-2023 period. The collected data were analyzed using SPSS version 27. The results of this study indicate that only Sustainability Reporting affects firm value, as measured by Price-to-Book Value (PBV). Then, Profitability as measured by ROA is also unable to moderate either Sustainability Reporting or ESG Disclosure on firm Value
Traffic Accident Spatial Modeling Using Adaptive Kernel Density Estimation Method Based on Geographical Information Systems For Road Sections In Brebes District, Brebes Regency
No full textBased on information from the Satlantas Polres Bebes, in 2022 the number of accidents in the Brebes Regency area reached 1.088 incidents that cause fatality damage and material losses. Data shows that in the last three years, Brebes District has recorded as the district with the highest accident statistics in this region. These incidents have a tendency to occur in certain sections. This research using a Geographic Information System (GIS) based approach using the Adaptive Kernel Density Estimation method to analyze the density of accidents on the road sections. The road network is divided into segments of 1000 meters and sub-segments iterated every 20 meters to obtain more accurate results. Vulnerability maps are classified based on the weighting of accident frequency, blackspot maps are classified based on Equivalent Accident Number (EAN) calculation. The results of the accuracy test comparison show that the Adaptive Kernel Density Estimation method can produce a vulnerability maps model with suitability level accuracy of 71,13%. In blackspot modeling, the CPAI calculation results show that the Adaptive Kernel Density Estimation method can produce a CPAI index of 71,73%
Peningkatan Kompetensi Siswa SMK melalui Pelatihan Dasar Programmable Logic Controller
No full textThis community service program was carried out at SMK Negeri 1 Minas to improve students’ competence in understanding and operating control systems based on Programmable Logic Controllers. The target participants were 20 students from the Electrical Power Installation and Air Conditioning and Refrigeration Engineering Departments who had limited skills in basic programming and industrial automation applications. The program adopted an educational and training approach through a structured workshop covering sequential logic, basic programming instructions such as timer, counter, comparison, arithmetic, and the introduction of human-machine interface systems. Evaluation was conducted through pretest and posttest, as well as a Likert-scale feedback questionnaire. The results indicated a significant improvement in student performance, with the average score increasing from 43.46 to 79.49, representing an 82.87% improvement. Participants also expressed a very high level of satisfaction with the materials, instructors, and hands-on sessions. Overall, this training program effectively enhanced students’ fundamental understanding and practical skills in automation control, thereby strengthening their readiness to enter the technology-driven industrial workforce.Kegiatan pengabdian kepada masyarakat ini dilaksanakan di SMK Negeri 1 Minas dengan tujuan meningkatkan kompetensi siswa dalam memahami dan mengoperasikan sistem kendali berbasis Programmable Logic Controller. Mitra sasaran adalah 20 siswa dari Jurusan Teknik Instalasi Tenaga Listrik dan Teknik Pendinginan dan Tata Udara yang memiliki keterbatasan dalam pemrograman dasar dan penerapan sistem otomatisasi industri. Metode pelaksanaan menggunakan pendekatan pendidikan dan pelatihan melalui lokakarya terstruktur yang mencakup pengenalan logika sekuensial, pemrograman instruksi dasar seperti timer, counter, comparison, arithmetic, serta pengenalan sistem antarmuka manusia. Evaluasi dilakukan melalui pretest dan posttest, serta kuesioner berbasis skala Likert. Hasil kegiatan menunjukkan peningkatan signifikan pada kemampuan siswa, dengan rata-rata nilai meningkat dari 43,46 menjadi 79,49 atau terjadi kenaikan sebesar 82,87%. Umpan balik peserta juga menunjukkan tingkat kepuasan yang sangat tinggi terhadap aspek materi, instruktur, dan kegiatan praktik. Secara keseluruhan, program pelatihan ini berhasil meningkatkan pemahaman dan keterampilan dasar siswa SMK dalam pengendalian otomatis, sekaligus memperkuat kesiapan mereka menghadapi dunia kerja berbasis teknologi industri
DESIGN AND PROTOTYPE DEVELOPMENT OF AN IOT-BASED TEMPERATURE AND HUMIDITY MONITORING SYSTEM WITH REAL-TIME DATA AND AUTOMATED ALERTS
No full textThis study presents the design and implementation of an Internet of Things (IoT)-based system for real-time monitoring of room temperature and humidity. The system integrates a DHT11 sensor with an ESP32-C3-WROOM-2 microcontroller to measure environmental parameters within ranges of 0–50 °C and 20–90% relative humidity, achieving accuracies of ±2°C and ±5% RH, respectively. Measurement data are displayed on a 20×4 I2C LCD and transmitted wirelessly via Wi-Fi to Google Sheets for cloud-based storage and analysis. The system features an automatic notification mechanism that sends alerts through WhatsApp when environmental conditions exceed predefined thresholds. Calibration testing against the HTC-2 reference standard device demonstrated a total average measurement error of 1.594%, confirming compliance with the DHT11 sensors specified tolerance limits. Consistency tests conducted in both air-conditioned and non-air-conditioned environments revealed that the system effectively captures genuine environmental variations, with higher variability observed in mechanically controlled spaces due to HVAC operational cycles. The developed prototype successfully addresses limitations of conventional monitoring methods by eliminating manual intervention requirements, enabling continuous real-time surveillance, and facilitating rapid preventive actions. Results indicate that the system provides sufficient measurement accuracy and stability for practical applications in residential, commercial, and industrial environmental monitoring, while offering a cost-effective and accessible solution for IoT-based climate control systems
Sentiment and Emotional Analysis of The Public Housing Savings Program (TAPERA) using Orange Data Mining
No full textThis study employs a text analysis methodology to assess public perception of the People\u27s Housing Savings Program (TAPERA), by examining 3.078 tweets containing the keyword "tapera" using the Orange Data Mining application with two analytical approaches: the Valence Aware Dictionary and Sentiment Reasoner (VADER) for sentiment analysis and the Profile of Mood States (POMS) for emotional analysis. The sentiment analysis results indicate 1.481 tweets (48,2%) expressed negative sentiment, 830 tweets (27%) were neutral, and 767 tweets (24,8%) conveyed positive sentiment. These findings suggest that although there is a portion of positive responses toward the TAPERA policy, most of the public tends to express dissatisfaction or scepticism about the program. Furthermore, the emotional analysis identified depression as the most dominant emotion expressed by the public, appearing in 2.019 tweets (65,6%), followed by confusion (14,7%) and anger (9,6%). Positive emotions such as vigour and tension were recorded in significantly lower proportions, at 2,9% and 1,8%, respectively. These results illustrate that the public feels frustrated, confused, and anxious regarding the TAPERA policy, with minimal expressions of optimism or enthusiasm. This analysis highlights the need for a more transparent, educational, and data-driven communication approach to enhance public understanding, trust, and participation in the TAPERA policy. Therefore, the government must design more effective outreach strategies to address public concerns and ensure the successful implementation of this program