Abstract:Penggunaan dompet digital yang terus meningkat menghasilkan banyak ulasan pengguna yang dapat dimanfaatkan untuk mengevaluasi kualitas layanan. Penelitian ini bertujuan meningkatkan akurasi klasifikasi sentimen pengguna…
dompet digital menggunakan metode Stacking Ensemble Machine Learning yang mengombinasikan Support Vector Machine (SVM), Multinomial Naïve Bayes (MNB), dan AdaBoost dengan Logistic Regression sebagai meta-learner. Data ulasan diproses melalui tahapan text preprocessing meliputi case folding, cleaning, tokenizing, stopword removal, stemming, dan pembobotan fitur menggunakan TF-IDF. Penyeimbangan data dilakukan dengan SMOTE, sedangkan evaluasi model menggunakan 5-Fold Cross-Validation. Hasil penelitian menunjukkan bahwa model Stacking Ensemble memperoleh akurasi rata-rata 80,55%, lebih tinggi dibandingkan algoritma dasar. Evaluasi menggunakan Confusion Matrix, Classification Report, dan ROC Curve juga menunjukkan peningkatan nilai precision, recall, F1-score, dan kemampuan diskriminasi model. Hasil ini menunjukkan bahwa pendekatan Stacking Ensemble Machine Learning efektif untuk meningkatkan akurasi klasifikasi sentimen pengguna dompet digital serta mendukung evaluasi kualitas layanan berbasis opini pengguna.
The rapid growth of digital wallet usage has generated a large volume of user reviews that can be utilized to evaluate service quality. This study aims to improve the accuracy of digital wallet user sentiment classification using a Stacking Ensemble Machine Learning approach that combines Support Vector Machine (SVM), Multinomial Naïve Bayes (MNB), and AdaBoost with Logistic Regression as the meta-learner. User reviews were processed through text preprocessing stages, including case folding, text cleaning, tokenization, stopword removal, stemming, and TF-IDF feature weighting. Synthetic Minority Over-sampling Technique (SMOTE) was employed to address class imbalance, while model performance was evaluated using 5-Fold Cross-Validation. The experimental results show that the proposed Stacking Ensemble model achieved an average accuracy of 80.55%, outperforming the individual base learners. Furthermore, evaluations based on the Confusion Matrix, Classification Report, and Receiver Operating Characteristic (ROC) Curve demonstrated improvements in precision, recall, F1-score, and the model's discriminative capability. These findings indicate that the proposed Stacking Ensemble Machine Learning approach is effective in improving the accuracy of digital wallet user sentiment classification and can serve as a reliable tool for supporting service quality evaluation based on user opinions.
Abstract:PayLater services are one of the rapidly growing digital financial innovations widely utilised in fintech apps in Indonesia, including Kredivo and Akulaku. User reviews on the Google Play Store reflect a range of experiences,…
nces, from satisfaction with the ease of use of the service to complaints regarding bills, interest rates, late payment fees, credit limits, and app performance. This study aims to classify the sentiment of user reviews regarding PayLater services on the Kredivo and Akulaku apps using the Multinomial Naïve Bayes algorithm. Data was collected via web scraping from the Google Play Store and automatically labelled based on user ratings, with ratings of 1-2 classified as negative sentiment and ratings of 4-5 as positive sentiment, whilst a rating of 3 was excluded as it was considered ambiguous. Following a preprocessing stage comprising cleaning, case folding, tokenisation, stopword removal, and stemming, as well as feature extraction using TF-IDF, 3,652 reviews were obtained with a training-to-test data split ratio of 80:20. The results indicate that positive sentiment dominates the dataset at 56.49%, whilst negative sentiment accounts for 43.51%. Analysis by application revealed that Kredivo was dominated by positive sentiment (68.20%), whilst Akulaku was dominated by negative sentiment (51.70%). The Naïve Bayes multinomial model achieved an accuracy of 84.13%, with average precision, recall, and F1-score values of 0.84, demonstrating good and balanced classification performance across both sentiment classes.
Abstract:The learning motivation of slow learner students tends to be low due to cognitive limitations, less engaging learning experiences, and the lack of positive reinforcement, thus requiring adaptive and sustainable interventions.…
ions. This study focuses on how the implementation of the token economy method can enhance motivation, engagement, and learning behavior of slow learner students with speech delay in elementary school, as well as examining its effectiveness in influencing behavioral change. The study employed a quantitative approach using a single case study A-B-A’ design, which allows systematic observation of behavioral changes through pre-intervention, intervention, and post-intervention phases. The findings indicate a significant improvement in participation, persistence, and task completion, along with a reduction in negative behaviors such as task refusal and crying during the intervention phase. These improvements were maintained in the post-intervention phase despite the reduced intensity of token reinforcement, indicating a shift from extrinsic to intrinsic motivation. Trend analysis of the data reveals consistent and stable improvement across all phases, confirming the effectiveness of the token economy method as a behavioral intervention. The novelty of this study lies in the specific application of the token economy to slow learner students with speech delay through a single-subject approach, providing an in-depth and contextual understanding of motivational changes. The findings contribute theoretically to the reinforcement of the behavioristic approach and practically offer an alternative adaptive learning strategy for inclusive education at the elementary level.
Abstract:This research discusses the Economic Token Technique Method for Improving Student Discipline at Muhammadiyah Gresik Elementary School. The aim of this research is to determine the effectiveness of the economic token intervention…
rvention model to help students be more disciplined during the teaching and learning process for grade 2 students at Muhammadiyah Elementary School. The method used in this research is a quantitative method. The problem in this research is that students have low disciplinary attitudes when in class and also during the teaching and learning process. This is because the subject is a person who is not comfortable studying and remaining silent in class for a long time. In this problem, to overcome this problem, behavior modification is carried out using the token economy technique. The subject used in this research was 1 boy from class 2 of SD Muhammadiyah Gresik who had a low disciplinary attitude. This research aims to determine the increase in student discipline by modifying behavior using the token economy technique. In the implementation, research subjects were given intervention or treatment in the form of rewards by sticking stickers for 21 days with pretest, intervention and posttest stages, then their disciplinary attitudes were observed and at the end of the session if the target behavior that was changed showed good changes then the subject would be given a reward according to the agreement at the beginning. where the prize is exchanged for the number of stickers obtained during the session. This can be seen from the pretest results, the significance value obtained was 10 and for the posttest results the significance value obtained was 13. Based on the gain score calculation, it showed that the subject experienced an increase of 3 scores from the pretest score of 10 to 13. During the posttest this showed an increase but it was not significant.
Abstract:Abstract: In this study, an analysis of the use of the Naive Bayes algorithm for sentiment analysis of reviews from Shopee app users on the Google Play Store was conducted, with classification divided into three categories:…
es: positive, negative, and neutral. To improve data quality, a preprocessing process was carried out with stages of cleaning, case folding, normalization, stop word removal, stemming, and tokenizing. Next, the text is formatted using the TF-IDF method to facilitate classification. For this data, the Naive Bayes model is used, which has an accuracy rate of 87% in detecting sentiment. Positive and negative categories can be easily identified compared to neutral sentiments due to the smaller amount of neutral data. Overall, the Naive Bayes algorithm successfully analyzed user sentiments well. The research can be developed with other algorithm methods, such as SVM, K-NN, or Decision Tree, in order to compare the performance of various algorithms.
Keywords: sentiment analysis; naive bayes; user reviews; e-commerce; shopee
Abstrak: Dalam penelitian ini dilakukan analisis penggunaan algoritma Naive Bayes untuk analisis sentimen review dari pengguna aplikasi Shopee di Google Play Store, klasifikasi dibagai menjadi 3 kategori yaitu positif, negatif, dan netral. Untuk meningkatkan kualitas data, dilakukam proses preprocessing dengan tahap cleanimg, case folding, normalisasi, stopword removal, stemming, dan tekonezing. Selanjutnya, teks diformat menggunakan metode TF-IDF untuk memudahkan klasifikasi. Untuk data ini, model Naive Bayes digunakan, yang memiliki tingkat akurasi 87% dalam mendeteksi sentimen. Kategori positif dan negatif dapat dengan mudah diidentifikasi dibadingkan sentiemen netral karena jumlah data netral yang lebih sedikit. Secara keseluruhan, algoritma Naive Bayes berhasil menganalisis perasaan pengguna dengan baik. Penelitian dapat dikembangkan dengan algoritma metode lain, seperti SVM, K-NN, atau Decision Tree, guna membandingkan kinerja berbagai algoritma.
Kata kunci: analisis sentiment; naive bayes; ulasan pengguna; e-commerce; shopee
Abstract:Abstract: The rapid growth of e-commerce mobile applications has generated large volumes of user reviews, making manual sentiment analysis increasingly impractical. This study aims to compare the effectiveness of three machine…
achine learning algorithms Support Vector Machine (SVM), Random Forest, and Naive Bayes for automated sentiment classification of Indonesian-language mobile application reviews. A dataset of 3,000 user reviews from the RupaRupa application on the Google Play Store was collected and preprocessed through normalization, tokenization, stopword removal, and stemming. TF-IDF vectorization was applied for feature extraction, while the Synthetic Minority Over-sampling Technique (SMOTE) was used to address class imbalance across three sentiment categories: positive, negative, and neutral. The results show that SVM achieved the highest accuracy of 90.02%, while Random Forest obtained the best F1-score of 88.08% when sufficient training data were available. Naive Bayes demonstrated relatively stable performance across varying training data sizes. Furthermore, TF-IDF keyword analysis revealed that negative reviews were primarily associated with delivery issues, technical problems, and pricing concerns. These findings demonstrate the effectiveness of machine learning approaches for sentiment classification and provide practical insights for improving mobile application services.
Keywords: sentiment analysis; machine learning; SMOTE; TF-IDF; text classification
Abstrak: Pertumbuhan pesat aplikasi mobile e-commerce telah menghasilkan volume ulasan pengguna yang sangat besar, sehingga analisis sentimen secara manual menjadi semakin tidak praktis. Penelitian ini bertujuan untuk membandingkan efektivitas tiga algoritma machine learning Support Vector Machine (SVM), Random Forest, dan Naive Bayes dalam melakukan klasifikasi sentimen otomatis terhadap ulasan aplikasi mobile berbahasa Indonesia. Dataset yang digunakan terdiri dari 3.000 ulasan pengguna aplikasi RupaRupa yang dikumpulkan dari Google Play Store. Data kemudian diproses melalui tahapan preprocessing yang meliputi normalisasi, tokenisasi, penghapusan stopword, dan stemming. Ekstraksi fitur dilakukan menggunakan metode Term Frequency–Inverse Document Frequency (TF-IDF), sedangkan ketidakseimbangan kelas ditangani menggunakan Synthetic Minority Over-sampling Technique (SMOTE) pada tiga kategori sentimen, yaitu positif, negatif, dan netral. Hasil penelitian menunjukkan bahwa SVM mencapai tingkat akurasi tertinggi sebesar 90,02%, sementara Random Forest memperoleh nilai F1-score terbaik sebesar 88,08% ketika tersedia data pelatihan yang memadai. Naive Bayes menunjukkan performa yang relatif stabil pada berbagai ukuran data pelatihan. Selain itu, analisis kata kunci berbasis TF-IDF mengungkapkan bahwa ulasan negatif terutama berkaitan dengan masalah pengiriman, kendala teknis aplikasi, dan isu harga. Temuan ini menunjukkan bahwa pendekatan machine learning efektif untuk klasifikasi sentimen serta memberikan wawasan yang bermanfaat dalam meningkatkan kualitas layanan aplikasi mobile.
Kata Kunci: analisis sentimen; pembelajaran mesin; SMOTE; TF-IDF; klasifikasi teks.
Abstract:Deteksi emosi secara fine-grained pada teks media sosial merupakan salah satu tantangan dalam bidang pemrosesan bahasa alami (Natural Language Processing/NLP), terutama karena sifat data yang tidak terstruktur dan multi-label.…
label. Penelitian ini bertujuan untuk mengevaluasi performa tiga model berbasis arsitektur Transformer, yaitu EmoBERT, RoBERTa, dan EmoRoBERTa, dalam tugas klasifikasi emosi pada teks dari dataset SenWave. Dataset ini terdiri dari 10.001 tweet berbahasa Inggris yang telah dilabeli ke dalam sepuluh kategori emosi, namun penelitian ini berfokus pada empat label utama: anxious, annoyed, empathetic, dan sad. Proses penelitian meliputi prapemrosesan data, tokenisasi, pembagian data latih dan uji, pelatihan model, serta evaluasi menggunakan metrik akurasi, presisi, recall, dan f1-score. Hasil evaluasi menunjukkan bahwa model EmoBERT dan EmoRoBERTa memiliki performa terbaik dengan nilai f1-score sebesar 0,81, sedangkan RoBERTa memperoleh nilai f1-score sebesar 0,73. Temuan ini menunjukkan bahwa penyesuaian arsitektur Transformer khusus untuk emosi dapat meningkatkan akurasi klasifikasi emosi pada teks media sosial.
Abstract:Abstract: The global transition towards sustainable food systems faces significant challenges in vegetarian food supply chains, including transparency issues, distribution inefficiencies, and quality verification problems.…
s. This research proposes VegeChain development, a decentralized marketplace ecosystem based on smart contracts designed to transform vegetarian food supply chains and accelerate Meatless, Balanced, Green (MBG) program adoption. Using mixed-method methodology integrating blockchain system design, stakeholder analysis, and economic simulation, this research develops a comprehensive technology framework combining blockchain transparency, smart contract automation, and sustainable tokenomics with novel mathematical models. The system implements dynamic pricing algorithms based on Automated Market Maker (AMM) mechanisms, multi-objective optimization for supply chain efficiency, and reputation-based consensus protocols. Simulation results demonstrate that VegeChain implementation can improve supply chain efficiency by 35%, reduce food waste by 28%, and increase consumer trust by 42% measured through validated stakeholder satisfaction surveys (n=456) using 5-point Likert scales with statistical significance p<0.001. Technical innovations include Byzantine Fault Tolerant consensus with 99.9% reliability, gas optimization achieving 67% cost reduction, and real-time quality verification algorithms with 98.7% accuracy.
Keywords: smart contracts; supply chain optimization; automated market makers; blockchain technology; sustainable tokenomics
Abstract:Abstract: Twitter occupies the top position of the most popular social media platform in Indonesia. Police and other related issues were the subject of much discussion. The aim of this research is to analyze public sentiment…
ment towards the National Police Agency using Twitter with the support vector machine method. The research started by crawling Twitter data. The data contains a total of 6,925 entries for three keywords. Next, we move on to the preprocessing stage consisting of (cleaning, case folding, tokenization, and filtering). Next is the tf-idf feature extraction stage, finally the classification and evaluation stage. The results of manual data inspection (73:27) showed accuracy of 70.66%, precision of 70.68%, and recall of 99.76%. Testing the second data (82:18), found accuracy 86%, precision 86.21%, recall 99.71%. The results of manual data checking (82:18) showed accuracy of 70.66%, precision of 70.68%, recall of 99.76%. Testing the second data (82:18), found accuracy 86%, precision 86.21%, recall 99.71%. From the data system testing results (80:20), accuracy was 87.55%, positive precision 87.53%, negative precision 88.24%, positive recall 99.48%, and negative recall. the rate is 99.48.% – The result is 21.43%. Data testing results (60:40) showed accuracy of 86.89%, positive precision of 86.84%, negative precision of 88.46%, positive recall of 99.61%, and negative recall of 16.43%. Single test data validation system (80:20), accuracy 87.55, overall test cross validation system (k fold 5 accuracy) 86.673%.
Keywords: data mining;police agencies;support vector machines
Abstrak: Twitter menduduki posisi teratas platform media sosial terpopuler di Indonesia. Polisi dan masalah terkait lainnya menjadi pokok bahasan banyak pembicaraan. Tujuan penelitian ini untuk menganalisis sentimen masyarakat terhadap Badan Kepolisian Nasional menggunakan Twitter dengan metode support vector machine. Penelitian dimulai dengan crawling data Twitter. Data memuat total 6.925 entri dari tiga kata kunci. Selanjutnya beralih ke tahap preprocessing terdiri dari (pembersihan, pelipatan kasus, tokenisasi, dan pemfilteran). Selanjutnya tahap ekstraksi fitur tf-idf, terakhir tahap klasifikasi dan evaluasi. Hasil pemeriksaan data manual (73:27) menunjukkan akurasi 70,66%, presisi 70,68%, dan recall 99,76%. Menguji data kedua (82:18), menemukan akurasi 86%, presisi 86,21%, recall 99,71%. Hasil pemeriksaan data secara manual (82:18) menunjukkan akurasi 70,66%, presisi 70,68%, recall 99,76%. Menguji data kedua (82:18), menemukan akurasi 86%, presisi 86,21%, recall 99,71%. Dari hasil pengujian sistem data (80:20), akurasi 87,55%, presisi positif 87,53%, presisi negatif 88,24%, recall positif 99,48%, dan recall negatif. tarifnya adalah 99,48.% – Hasilnya 21,43%. Hasil pengujian data (60:40) menunjukkan akurasi 86,89%, presisi positif 86,84%, presisi negatif 88,46%, recall positif 99,61%, dan recall negatif 16,43%. Uji tunggal sistem validasi data (80:20), akurasi 87,55, uji keseluruhan sistem validasi silang (akurasi k fold 5) 86,673%.
Kata Kunci: data mining;instansi kepolisian;mesin vektor pendukung
Abstract:Konflik geopolitik di kawasan Timur Tengah, khususnya konflik Israel-Gaza yang meningkat eskalasi sejak Oktober 2023, menjadi salah satu isu internasional yang paling banyak diberitakan oleh media online Indonesia. Penelitian…
itian ini bertujuan untuk mengidentifikasi frame dominan yang digunakan media online Indonesia dalam memberitakan konflik tersebut menggunakan pendekatan komputasional berbasis topic modeling Latent Dirichlet Allocation (LDA). Sebanyak 127 artikel berita dikumpulkan dari berbagai media online Indonesia melalui metode web scraping pada platform Google News RSS, kemudian diproses melalui tahapan preprocessing teks meliputi cleaning, tokenisasi, stopword removal, dan stemming menggunakan library Sastrawi. Model LDA dengan K=4 topik diidentifikasi sebagai konfigurasi optimal berdasarkan coherence score tertinggi sebesar 0,5620. Hasil analisis menemukan empat frame dominan dalam pemberitaan, yaitu: (1) Frame Geopolitik dan Eskalasi Regional (21,3%), (2) Frame Diplomasi dan Negosiasi (32,3%), (3) Frame Dampak Ekonomi Global (17,3%), dan (4) Frame Kemanusiaan dan Korban Sipil (29,1%). Temuan ini menunjukkan bahwa media online Indonesia cenderung membingkai konflik Timur Tengah melalui lensa diplomatik dan kemanusiaan, dengan relatif sedikit perhatian pada dimensi ekonomi. Penelitian ini berkontribusi pada pengembangan metodologi analisis framing berbasis komputasional untuk konteks bahasa Indonesia.