BEES: Bulletin of Electrical and Electronics Engineering

Analisis Sentimen Ulasan Pengguna Aplikasi DANA pada Google Play Store Menggunakan TF-IDF dan Naïve Bayes

Puspita Wanny — 2026-07-27

The DANA application is a digital wallet service widely used by the public to support various digital financial transaction activities. The high number of users results in numerous reviews on the Google Play Store containing various responses, experiences, and opinions regarding the quality of the DANA application service. However, the large number of review data makes the manual process of identifying and grouping opinions less effective and takes a relatively long time. This study aims to analyze and classify the sentiment of DANA application user reviews on the Google Play Store into positive, negative, and neutral categories and to determine the performance of the algorithm used in the classification process. The solution implemented is sentiment analysis using a text mining approach and Natural Language Processing (NLP) to process user reviews automatically. The research data was obtained through a scraping process and resulted in 3,509 DANA application user reviews. The data then went through preprocessing stages including cleaning, case folding, normalization, tokenizing, stopword removal, and stemming. Then, sentiment labeling and word weighting were carried out using the Term Frequency-Inverse Document Frequency (TF-IDF) method. The data was then divided into 80% training data and 20% testing data. Classification was then performed using the Multinomial Naïve Bayes algorithm. Model performance was evaluated using a Confusion Matrix with Accuracy, Precision, Recall, and F1-Score metrics. The results showed that the Naïve Bayes model produced an Accuracy value of 80.48%, Precision of 76.75%, Recall of 80.48%, and F1-Score of 78.13%. These results indicate that the combination of the TF-IDF method and the Naïve Bayes algorithm is capable of classifying the sentiment of DANA app user reviews with quite good performance and can be used to help obtain an overview of user perceptions of the DANA app based on reviews provided on the Google Play Store

Analisis Pola Asosiasi Penjualan Toko Bangunan Menggunakan Algoritma Apriori Untuk Strategi Penempatan Barang

Viany Berliana Jelita Koraag — 2026-07-29

Small and medium-scale hardware stores generally still rely on intuition when determining product layout without systematically considering customer purchasing patterns, causing cross-selling opportunities to remain underutilized. This study aims to analyze product association patterns in hardware store transaction data using the Apriori algorithm as the basis for a data-driven product placement strategy. The dataset consists of 90 transaction rows representing 30 unique transactions involving 9 product types. Research stages include data cleaning, one-hot encoding transformation, and application of the Apriori algorithm with a minimum support of 15% and minimum confidence of 40%. The analysis identified 14 frequent itemsets and 10 association rules, all with lift values above 1.0, indicating positive associations. The strongest rules were Wall Paint → Brush with a confidence of 77.78% and lift of 1.46, and Sand → Brick with a confidence of 50.00% and lift of 1.50. These findings provide an empirical foundation for shelf zone arrangement recommendations, product bundling packages, and stock management prioritization in hardware stores. The contribution of this research is to provide a data-driven analytical framework that can be directly adopted by small and medium-scale hardware store managers without requiring complex technological infrastructure, while also extending the application of the Apriori algorithm to the hardware store domain with a specific focus on physical product placement strategies.

Klasifikasi Sentimen Publik terhadap Isu Toleransi Agama Menggunakan Algoritma Random Forest

Flienschy Faith Maxy Tamaka — 2026-07-31

Social media, particularly Twitter, has evolved into a dynamic arena for discussions on religious issues in Indonesia. Interfaith tolerance is one of the topics that most frequently elicits a wide range of responses, from support to hate speech. This study designs a five-class sentiment scheme, consisting of Positive/Neutral, Neutral-Abusive, and Negative tweets divided into three intensity levels (Weak, Moderate, and Strong), and classifies them using the Random Forest algorithm. The dataset used is the Indonesian Abusive and Hate Speech Twitter Text available on the Kaggle platform, consisting of 13,169 tweets with dual labels. Sentiment labels were created based on a combination of the HS and HS_Religion columns and hate speech intensity levels: Weak, Moderate, and Strong. Tweets without hate speech and unrelated to religion are considered positive or neutral, while tweets with HS_Religion=1 are classified as negative and grouped into three intensity levels. Prior to modeling, the text undergoes slang normalization, removal of inappropriate words, Nazief-Adriani stemming, and feature extraction using TF-IDF bigrams. Results from 10-fold cross-validation show an accuracy of 66.0%, macro precision of 52.1%, macro recall of 56.1%, and macro F1-Score of 51.3%, comparable to SVM (F1 52.7%) and Naive Bayes (31.8%), with differences between models assessed statistically using the McNemar test.

Optimasi Load Balancing Trafik Jaringan LTE Melalui Implementasi Antena Pro Sectoral 1800 Mhz & 2100 Mhz

Almurozy Mursidan — 2026-07-31

This study aims to analyze the effectiveness of implementing a sectoral pro antenna on LTE 1800 MHz and 2100 MHz frequencies in improving LTE network performance at the Kota Wisata site, Bogor, West Java. A quantitative comparative method was employed by comparing network performance before and after the implementation of the pro antenna. Data were collected from the operator’s network monitoring system during the pre-implementation period (Mei 11–12, 2026) and the post-implementation period (Mei 18–19, 2026). The analyzed parameters included Main Site Payload, Cluster Payload, and RSRP Coverage. The results indicate that the implementation of the sectoral pro antenna significantly improved LTE network performance. The Main Site Payload increased from 335.42 GB to 726.13 GB, representing an improvement of 116.48%, while the Cluster Payload increased from 6,308.82 GB to 6,431.93 GB, representing an increase of 1.95%. In addition, network coverage quality improved, as indicated by the increase in the percentage of RSRP values greater than -100 dBm from 99.09% to 99.85%, as well as an overall RSRP coverage improvement of 2.11%. These results demonstrate that the implementation of the sectoral pro antenna on LTE 1800 MHz and 2100 MHz frequencies is effective in distributing traffic more evenly, reducing the potential for network congestion, improving signal coverage quality, and enhancing overall LTE network performance.

Prototype Pengendali Lampu Pada Rumah Pintar dengan Tinkercad

Nober Yohanis — 2026-08-29

The development of automation technology has encouraged the implementation of smart home concepts aimed at improving comfort, efficiency, and convenience in controlling electronic devices, particularly lighting systems. This study aims to design and implement an automatic lamp control prototype through simulation using Tinkercad by utilizing Arduino Uno, an LDR sensor as a light intensity detector, PIR sensors as motion detectors, and a relay as the lamp controller. The method used in this research is the Prototyping Process Model, which is carried out iteratively through the stages of communication, planning, design, prototype construction, evaluation, and system refinement. The results show that the system operates according to the designed logic, where the lamp turns on only under dark conditions when motion is detected, and remains off under bright conditions even when activity is present. The use of multiple PIR sensors has proven effective in expanding motion detection coverage, thereby increasing system responsiveness, while the LDR sensor plays an important role in improving energy efficiency by preventing unnecessary electricity consumption. Simulation using Tinkercad has also proven effective in simplifying the design and testing process before real-world implementation. The contribution of this research is the development of a smart home lighting control prototype that has been validated through Tinkercad simulation as an initial testing platform, thereby reducing hardware development costs and risks while providing a reference for the future development of Arduino-based smart home automation systems.