Claim Missing Document
Check
Articles

Found 3 Documents
Search

Sara Detection on Social Media Using Deep Learning Algorithm Development M. Khairul Anam; Lucky Lhaura Van FC; Hamdani Hamdani; Rahmaddeni Rahmaddeni; Junadhi Junadhi; Muhammad Bambang Firdaus; Irwanda Syahputra; Yuda Irawan
Journal of Applied Engineering and Technological Science (JAETS) Vol. 6 No. 1 (2024): Journal of Applied Engineering and Technological Science (JAETS)
Publisher : Yayasan Riset dan Pengembangan Intelektual (YRPI)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37385/jaets.v6i1.5390

Abstract

Social media has become a key platform for disseminating information and opinions, particularly in Indonesia, where SARA (Ethnicity, Religion, Race, and Intergroup) issues can fuel social tensions. To address this, developing an automated system to detect and classify harmful content is essential. This study develops a deep learning model using Convolutional Neural Network (CNN) and Bidirectional Long Short-Term Memory (BiLSTM) to detect SARA-related comments on Twitter. The method involves data collection through web scraping, followed by cleaning, manual labeling, and text preprocessing. To address data imbalance, SMOTE (Synthetic Minority Over-sampling Technique) is applied, while early stopping prevents overfitting. Model performance is evaluated using precision, recall, and F1-score. The results demonstrate that SMOTE significantly improves model performance, particularly in detecting minority-class SARA comments. CNN+SMOTE achieves a accuracy of 93%, and BiLSTM+SMOTE records a recall of 88%, effectively capturing patterns in SARA and non-SARA data. With SMOTE and early stopping, the model successfully manages class imbalance and reduces overfitting. This research supports efforts to curtail hate speech on social media, especially in the Indonesian context, where SARA-related issues often dominate public discourse.
Klasterisasi Lagu Populer dan Eksplorasi Subgenre Spotify 2024 dengan K-Medoids Alfia Nurlaili Tahiyat; Bima Maulana; Ade Eka Saputra; Lusiana Efrizoni; Rahmaddeni Rahmaddeni
Jurnal Informatika Dan Tekonologi Komputer (JITEK) Vol. 5 No. 1 (2025): Maret : Jurnal Informatika dan Tekonologi Komputer
Publisher : Pusat Riset dan Inovasi Nasional

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55606/jitek.v5i1.5699

Abstract

Spotify's genre classification system remains too broad, often grouping songs with distinct characteristics into the same category. For example, Pop Ballads and Dance Pop are frequently classified under "Pop" despite significant differences in tempo, emotion, and production style. This leads to inaccurate song recommendations. This study applies the K-Medoids algorithm to enhance song classification based on Spotify Playlist Count, Spotify Playlist Reach, and Spotify Popularity. The CRISP- DM methodology guides business understanding, data understanding, data preparation, modeling, evaluation, and deployment. Clustering results without popularity ranking reveal three main groups: songs with low playlist count but high reach (dominated by light hip-hop), songs with high playlist count and reach (dominated by contemporary R&B), and songs with low popularity (dominated by dance). After ranking by popularity, clusters became more defined, with alternative pop dominating the high-reach cluster, contemporary R&B in the popular cluster, and dance pop in the less popular cluster. Evaluation using a Silhouette Score of 0.5014 indicates good cluster quality. Additionally, this study successfully identified the 15 most popular songs on Spotify in 2024. These findings can help Spotify refine its recommendation system by incorporating subgenre-based classification, ensuring more accurate search results aligned with user preferences and evolving music trends.
Penerapan Algoritma Support Vector Machine dan XGBoost Dalam Mengklasifikasikan Sentimen Opini Publik Terhadap Aplikasi Uber Rizky Rizaldi; M Ridho; Arraihan Tahta Ainullah; Lusiana Efrizoni; Rahmaddeni Rahmaddeni; M Fahrel Dea Putra
Jurnal Informatika Dan Tekonologi Komputer (JITEK) Vol. 5 No. 1 (2025): Maret : Jurnal Informatika dan Tekonologi Komputer
Publisher : Pusat Riset dan Inovasi Nasional

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55606/jitek.v5i1.5735

Abstract

The development of application-based transportation services such as Uber has driven an increase in the number of public opinions distributed through various digital platforms. Sentiment analysis of this public opinion is important to understand user perceptions of Uber services. This study applies the Support Vector Machine (SVM) and Extreme Gradient Boosting (XGBoost) algorithms to classify public opinion sentiment, by optimizing data imbalance using the Synthetic Minority Oversampling Technique (SMOTE). The data used comes from Uber reviews on public platforms, which are grouped into positive, negative, and neutral sentiments. The experimental results show that the SVM algorithm has superior performance with an accuracy of 94%, while XGBoost experienced an increase in accuracy of up to 93% after applying SMOTE. This study provides insight into the effectiveness of machine learning algorithms in sentiment analysis and its implementation in the development strategy of application-based transportation services. Abstrak: Perkembangan layanan transportasi berbasis aplikasi seperti Uber telah mendorong peningkatan jumlah opini publik yang disalurkan melalui berbagai platform digital. Analisis sentimen terhadap opini publik ini menjadi penting untuk memahami persepsi pengguna terhadap layanan Uber. Penelitian ini menerapkan algoritma Mesin Vektor Pendukung (SVM) dan Peningkatan Gradien Ekstrem (XGBoost) untuk mengklasifikasikan sentimen opini publik, dengan mengoptimalkan ketidakseimbangan data menggunakan Synthetic Minority Oversampling Technique (SMOTE). Data yang digunakan berasal dari ulasan Uber di platform publik, yang dikategorikan ke dalam sentimen positif, negatif, dan netral. Hasil eksperimen menunjukkan bahwa algoritma SVM memiliki performa lebih unggul dengan akurasi mencapai 94%, sementara XGBoost mengalami peningkatan akurasi hingga 93% setelah penerapan SMOTE. Penelitian ini memberikan wawasan mengenai efektivitas algoritma pembelajaran mesin dalam analisis sentimen serta implikasinya terhadap strategi pengembangan layanan transportasi berbasis aplikasi.