Jurnal Politeknik Negeri Batam (PoliBatam)
Not a member yet
3001 research outputs found
Sort by
PERSPECTIVE ON INVESTING IN JAKARTA ISLAMIC INDEX WITH FINTECH AS AN INTERVENING VARIABLE (Case Study on Generation Z)
No full textABSTRACT
This study aims to analyze what factors are influential in investing in the Sharia Capital Market focusing on the Jakarta Islamic Index from the perspective of generation Z. The research variables used in this study are the independent variables are financial inclusion, environment, and risk. Furthermore, for the dependent variable, investment interest is used and moderation uses fintech as an intervening variable. The research methodology used is quantitative with causality studies, this is to determine the cause and effect of the hypothesis proposed. The respondents used in the study were generation Z in East Java Province with a research sample using purposive sampling techniques. Data analysis techniques with validity tests and reliability tests and path analysis is used to test the influence between variables directly and indirectly between the variables used. The result of the research is that the influence of risk and fintech on investment interest has a significant positive influence. For the environment of awaraness and financial inclusion on investment interest, it does not have a significant positive influence. In Islamic Financial Inclusion, risk and environmental awareness of investment interest when mediated by fintech produce a higher value than the direct influence of each variable on investment interest.
Keywords
Pengembangan Sistem Manajemen Inventaris Berbasis IoT dengan Teknologi Pick to Light dan Sistem Identifikasi Barang untuk Meningkatkan Akurasi Pengambilan Barang
No full textSistem penyimpanan konvensional sering menimbulkan kesalahan dan memperlambat proses pengambilan barang. Untuk mengatasinya, dikembangkan sistem manajemen inventaris berbasis Internet of Things (IoT) yang mengintegrasikan aplikasi inventaris berbasis LabVIEW, sistem pick to light dengan ESP32 sebagai panduan visual, serta sistem identifikasi barang menggunakan ESP32-CAM dan QR-Code untuk verifikasi. Hasil pengujian menunjukkan efisiensi meningkat dengan rata-rata waktu pengambilan berkurang 34,3% dari 136,3 detik menjadi 89,6 detik. Sistem juga mampu mendeteksi kesalahan pengambilan secara real-time melalui sensor rak dan meningkatkan akurasi dengan identifikasi QR-Code. Dengan dukungan jaringan WiFi 5G, sistem ini terbukti mampu mempercepat pertukaran data, meningkatkan akurasi, efisiensi, serta mendukung pencatatan inventaris secara waktu nyat
A Comparative Study of Naïve Bayes and K-Nearest Neighbors (KNN) Algorithms in Sentiment Analysis of ChatGPT Usage Among Students
No full textThis study compares the performance of the Naïve Bayes and K-Nearest Neighbors (KNN) algorithms in sentiment analysis of Lhokseumawe State Polytechnic students toward the use of ChatGPT. The comparison is conducted due to the varied results of previous research, where the effectiveness of both algorithms largely depends on the data type and context. The model was developed using 9.800 external data collected from Twitter and Google Play Store, which were processed through text preprocessing and TF-IDF transformation stages, and then tested on 237 student questionnaire data as a case study. The initial evaluation showed that Naïve Bayes achieved an accuracy of 88% with a prediction time of 0,0063 seconds, while KNN recorded an accuracy of 83% with a prediction time of 0,4760 seconds. In the student questionnaire test, Naïve Bayes again outperformed with 79,75% accuracy compared to KNN’s 49,37%.Penelitian ini membandingkan kinerja algoritma Naïve Bayes dan K-Nearest Neighbors (KNN) dalam analisis sentimen mahasiswa Politeknik Negeri Lhokseumawe terhadap penggunaan ChatGPT. Perbandingan ini dilakukan karena penelitian sebelumnya menunjukkan hasil yang bervariasi, di mana efektivitas kedua algoritma sangat dipengaruhi oleh konteks dan jenis data. Model dikembangkan menggunakan 9.800 data eksternal dari Twitter dan Google Play Store yang diproses melalui tahapan praproses teks dan transformasi TF-IDF, kemudian diuji pada 237 data kuesioner mahasiswa sebagai studi kasus. Evaluasi awal memperlihatkan bahwa Naïve Bayes memperoleh akurasi 88% dengan waktu prediksi 0,0063 detik, sedangkan KNN mencatatkan akurasi 83% dengan waktu prediksi 0,4760 detik. Pada pengujian dengan data kuesioner, Naïve Bayes kembali unggul dengan akurasi 79,75% dibandingkan KNN yang hanya 49,37%. Temuan ini menegaskan bahwa Naïve Bayes lebih optimal untuk klasifikasi teks berbasis opini dalam konteks ini, serta dapat dijadikan pertimbangan dalam pengembangan kebijakan akademik terkait pemanfaatan kecerdasan buatan di pendidikan tinggi
Sentiment-Based Knowledge Discovery of Wondr by BNI App Reviews Using SVM, KNN, and Naive Bayes for CRM Enhancement
No full textThe rapid development of digital banking services has necessitated a deeper understanding of user perceptions and satisfaction levels. This study analyzes sentiment from user reviews of the Wondr by BNI app using a Knowledge Discovery approach and machine learning methods. Three classification algorithms were compared: Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Naive Bayes, evaluated with accuracy, precision, recall, and f1-score. The results show that SVM and Naive Bayes achieved the best performance with F1-scores of 0.88 and 0.87, while KNN lagged behind with 0.77. An ANOVA test further confirmed that the performance differences were statistically significant (p < 0.05), with SVM and Naive Bayes consistently outperforming KNN. Word Cloud analysis revealed dominant positive terms such as "easy," "fast," and "transaction," alongside negative terms like "login," "difficult," and "verification." These findings highlight user appreciation for simplicity and speed, while pointing out functional issues that require attention. This research not only enriches the literature on Indonesian-language sentiment analysis in the financial sector but also provides practical insights for Customer Relationship Management (CRM), particularly in strengthening customer retention strategies and guiding UX redesign for digital banking services.The rapid development of digital banking services has necessitated a deeper understanding of user perceptions and satisfaction levels. This study analyzes sentiment from user reviews of the Wondr by BNI app using a Knowledge Discovery approach and machine learning methods. Three classification algorithms were compared: Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Naive Bayes, evaluated with accuracy, precision, recall, and f1-score. The results show that SVM and Naive Bayes achieved the best performance with F1-scores of 0.88 and 0.87, while KNN lagged behind with 0.77. An ANOVA test further confirmed that the performance differences were statistically significant (p < 0.05), with SVM and Naive Bayes consistently outperforming KNN. Word Cloud analysis revealed dominant positive terms such as "easy," "fast," and "transaction," alongside negative terms like "login," "difficult," and "verification." These findings highlight user appreciation for simplicity and speed, while pointing out functional issues that require attention. This research not only enriches the literature on Indonesian-language sentiment analysis in the financial sector but also provides practical insights for Customer Relationship Management (CRM), particularly in strengthening customer retention strategies and guiding UX redesign for digital banking services
Utilizing IndoBERT and BERTopic to Explore Public Opinion on BPS Instagram Posts
No full textThis study aims to analyze public sentiment and topics of opinion toward the Central Statistics Agency (BPS) through comments on the Instagram account @bps_statistics. A total of 3,075 comments collected from January 1 to July 24, 2025, were analyzed using the IndoBERT model for sentiment classification and BERTopic for topic modeling. The IndoBERT model was developed using a semi-supervised learning approach, achieving an 88% classification accuracy with high precision and recall across all sentiment categories. The analysis results show that neutral comments dominate (52.78%), followed by negative comments (31.54%) and positive comments (15.69%). Topic modeling on negative sentiment revealed two main issues: distrust of poverty data and preference for international institution indicators such as the World Bank. Positive sentiment reflects appreciation for the quality of statistical data and moral support for BPS. Neutral comments mostly contain informative discussions about socioeconomic conditions and access to digital services. These findings emphasize the importance of improving BPS public communication, particularly in bridging the gap in public perception of official data. The social media-based approach has proven effective as a complement to formal surveys in capturing public opinion in a broad and dynamic manner.This study aims to analyze public sentiment and topics of opinion toward the Central Statistics Agency (BPS) through comments on the Instagram account @bps_statistics. A total of 3,075 comments collected from January 1 to July 24, 2025, were analyzed using the IndoBERT model for sentiment classification and BERTopic for topic modeling. The IndoBERT model was developed using a semi-supervised learning approach, achieving an 88% classification accuracy with high precision and recall across all sentiment categories. The analysis results show that neutral comments dominate (52.78%), followed by negative comments (31.54%) and positive comments (15.69%). Topic modeling on negative sentiment revealed two main issues: distrust of poverty data and preference for international institution indicators such as the World Bank. Positive sentiment reflects appreciation for the quality of statistical data and moral support for BPS. Neutral comments mostly contain informative discussions about socioeconomic conditions and access to digital services. These findings emphasize the importance of improving BPS public communication, particularly in bridging the gap in public perception of official data. The social media-based approach has proven effective as a complement to formal surveys in capturing public opinion in a broad and dynamic manner
Balancing Student Specialization Class Placement Based on Interests and Talents Using K-Means Clustering and Genetic Algorithm
No full textStudent specialization placement in Indonesian secondary schools often produces imbalanced class distributions and misalignment between student interests and assigned tracks. This study develops a hybrid optimization system combining K-Means clustering and Genetic Algorithm (GA) to allocate 133 tenth-grade students from SMAN 1 Ngimbang into four specialization classes (Science, Mixed-Science, Mixed-Social, Social) while balancing operational constraints. Initial K-Means clustering (k=4, n_init=100) achieved a Silhouette Score of 0.287 but yielded severely imbalanced distribution (10, 51, 48, 24 students). GA optimization (population=300, generations=150, crossover=70%, mutation=10%, elitism=10%) with multi-component fitness function incorporating cosine similarity, distribution penalty, movement penalty, and entropy produced balanced classes (31, 35, 35, 32 students) within the 30-35 target range. Post-optimization metrics showed 73.7% retention rate, average match score of 0.792, entropy of 0.482, and execution time of 47.8 seconds. The Silhouette Score decreased to 0.080, reflecting an acceptable trade-off between cluster purity and operational feasibility. Sensitivity analysis confirmed weight configuration robustness. This system demonstrates practical applicability for real-time school implementation, reducing distribution gap by 90.2% while maintaining individual-class compatibility
Forecasting the Number of Passengers for the Jakarta-Bandung High-Speed Rail using SARIMA and SSA Models
No full textTime series forecasting is essential for analyzing past data to predict future trends, supporting planning, and decision-making. The SARIMA model is widely used for seasonal data but may be less effective for highly fluctuating or non-stationary data, which can impact forecast accuracy. As an alternative, Singular Spectrum Analysis (SSA) offers a flexible approach, decomposing time series into trend, seasonal, and noise components without strict parametric assumptions, making it effective for complex data patterns. This study compares SARIMA and SSA models in forecasting daily passenger counts on the Jakarta-Bandung high-speed rail, using data from November 1, 2023, to September 30, 2024. The results show that the performance of SSA is more stable compared to SARIMA in the term of MAPE, where SSA provides lower MAPE then SARIMA in all three scenarios of data splits. These results are expected due to the non-linear pattern that appears in the data. Moreover, the predictions on both methods show that slight increment of passengers in the end of 2024 to the beginning of 2025. This finding suggests that the government needs to consider implementing interventions if they wish to change the current trend, such as offering discounts or year-end holiday promotions
Mobile-Based Multi-Output Animal Taxonomy Classification Using CNN with Edge and Cloud Deployment
No full textDistinguishing animals that appear visually similar but belong to different species or taxonomic groups, such as Eurasian and house sparrows, koi and common carp, or leopard cat and domestic cat, remains challenging and hinders biodiversity education. This study develops a Convolutional Neural Network (CNN)-based multi-output, multi-class taxonomy classification system capable of identifying seven animal species across five taxonomic levels (class, order, family, genus, species), producing 35 possible outputs. The dataset comprised 6,998 images from public sources. Among various configurations, the best-performing model (D3-M2), trained using the High Dataset with 256×256 input size, 0.2 dropout, and four hidden layers, achieved 90.15% average accuracy, the highest F1-score at the family level (98.11%), and 95.99% at the species level. Slightly lower species-level performance was due to high visual similarity among particular species. Edge AI deployment offered faster inference (0.17s) and offline capability, making it ideal for field use. Real-world testing under bright and low light at 30, 60, and 100 cm showed higher accuracy (64.8%) than low light (57.1%), with the most stable performance at 60 cm. However, limitations include an imbalanced dataset and limited environmental variation affecting species-level accuracy. Future work will focus on expanding dataset diversity and employing advanced architectures to improve fine-grained classification. This system offers a practical tool for biodiversity education and species identification, particularly in field environments where rapid, offline, and accurate classification is essential
Performance Comparison of Multilayer Perceptron (MLP) and Random Forest for Early Detection of Cardiovascular Disease
No full textCardiovascular disease is a disorder of the heart and blood vessels that can lead to heart attacks, strokes, and heart failure, so early detection is essential. This study compares Multilayer Perceptron (MLP) and Random Forest for risk classification in a Kaggle dataset containing 70,000 samples with balanced targets. Pre-processing included age conversion, outlier cleaning, standardization, and feature selection based on feature importance. Both models were optimized using RandomizedSearchCV and evaluated using accuracy, precision, recall, F1-score, AUC-ROC, confusion matrix, and k-fold cross-validation. The results show that the accuracy of MLP is 73.90% and Random Forest is 74.23% with an AUC of 0.80 for both. Random Forest is more stable across all folds and performs better on the negative class, while MLP is slightly more sensitive to the positive class. Independent t-test and Mann-Whitney U tests show p>0.05, indicating that the difference in performance is not significant. The most influential features were diastolic blood pressure, age, cholesterol, and systolic blood pressure. The non-clinical Streamlit prototype demonstrated the model\u27s potential for education and initial decision support
Comparison of Linkage Methods in Hierarchical Clustering for Grouping Districts/Cities in East Java Based on Stunting Determinants
No full textStunting is a long-term nutritional problem that generally occurs in children under five years old and is characterized by a shorter body than other children of the same age due to continuous dietary deficiencies. As a result of the Indonesian Health Survey (SKI) conducted in 2023, the stunting rate in East Java decreased to 17.7%. In 2024, the target is to reduce it to 14%. This study aims to group regencies and cities in East Java based on indicators of child nutritional status by using five linkage approaches in the hierarchical clustering method. This study found areas with similar causes of stunting so that intervention programs can be more targeted. The analysis showed that the centroid linkage methods formed two clusters with the highest cophenetic correlation coefficient of 0.8619. The first cluster consists of 37 regencies/cities with a low stunting category, and the second cluster consists of one regency/city with a high stunting category. The model in this clustering has a silhouette value of 0.6155, which indicates that the model is in the good category