INDAH RELISIAWATI, . (2026) DETEKSI EMOSI PADA TEKS BERBAHASA INDONESIA DI APLIKASI X MENGGUNAKAN ALGORITMA RANDOM FOREST. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.
|
Text
COVER.pdf Download (538kB) |
|
|
Text
BAB I.pdf Download (608kB) |
|
|
Text
BAB II.pdf Restricted to Registered users only Download (1MB) | Request a copy |
|
|
Text
BAB III.pdf Restricted to Registered users only Download (1MB) | Request a copy |
|
|
Text
BAB IV.pdf Restricted to Registered users only Download (1MB) | Request a copy |
|
|
Text
BAB V.pdf Restricted to Registered users only Download (568kB) | Request a copy |
|
|
Text
DAFTAR PUSTAKA.pdf Download (568kB) |
|
|
Text
LAMPIRAN.pdf Restricted to Registered users only Download (2MB) | Request a copy |
Abstract
Peningkatan penggunaan media sosial X di Indonesia menghasilkan data teks emosional yang melimpah, namun penelitian deteksi emosi berbasis teks berbahasa Indonesia masih terbatas. Penelitian ini bertujuan untuk menganalisis kinerja kombinasi algoritma Random Forest dengan representasi fitur TF-IDF dalam mendeteksi emosi pada teks berbahasa Indonesia dari media sosial X dan membandingkannya dengan metode baseline (SVM dan KNN yang dikombinasikan dengan TF-IDF dan Word2Vec). Data yang digunakan sebanyak 7.687 tweet berbahasa Indonesia yang mencakup enam kategori emosi, yaitu Senang, Sedih, Marah, Takut, Jijik, dan Kaget. Data melalui tahapan pre-processing meliputi cleansing, case folding, stopword removal, tokenisasi, dan stemming, kemudian direpresentasikan menggunakan TF-IDF dan Word2Vec, lalu diklasifikasikan dengan enam model yang dievaluasi secara sistematis. Hasil penelitian menunjukkan bahwa model utama Random Forest + TF-IDF mencapai akurasi tertinggi sebesar 95,83%, dengan precision weighted 95,98%, recall 95,83%, dan F1-Score 95,86%, mengungguli semua model baseline termasuk SVM + TF-IDF (93,47%) sebagai baseline terbaik. Kesimpulan penelitian ini adalah kombinasi Random Forest dengan TF-IDF merupakan pendekatan yang efektif untuk deteksi emosi teks berbahasa Indonesia dari media sosial, dengan keunggulan pada mekanisme ensemble learning yang robust dan representasi fitur TF-IDF yang kaya. ***** The growing use of social media X in Indonesia produces an abundance of emotional text data; however, research on text-based emotion detection in Indonesian remains limited. This study aims to analyze the performance of the Random Forest algorithm combined with Term Frequency-Inverse Document Frequency (TF-IDF) feature representation in detecting emotions from Indonesian-language text on social media X, and to compare it with baseline methods (SVM and KNN combined with TF-IDF and Word2Vec). The dataset consisted of 7,687 Indonesian-language tweets covering six emotion categories: Happy, Sad, Angry, Fear, Disgust, and Surprise. Data underwent pre-processing steps including cleansing, case folding, stopword removal, tokenization, and stemming, then represented using TF-IDF and Word2Vec, and classified using six models evaluated systematically. Results show that the main model, Random Forest + TF-IDF, achieved the highest accuracy of 95.83%, with a weighted precision of 95.98%, recall of 95.83%, and F1-Score of 95.86%, outperforming all baseline models including SVM + TF-IDF (93.47%) as the best baseline. This study concludes that the combination of Random Forest with TF-IDF is an effective approach for emotion detection in Indonesian-language text from social media, with advantages in its robust ensemble learning mechanism and the rich feature representation produced by TF-IDF.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Additional Information: | 1). Dr. Widodo, S.Kom., M.Kom; 2). Murien Nugraheni, S.T., M.Cs. |
| Subjects: | Teknologi dan Ilmu Terapan > Teknik Komputer |
| Divisions: | FT > S1 Pendidikan Teknik Informatika Komputer |
| Depositing User: | Users 34540 not found. |
| Date Deposited: | 07 Aug 2026 04:01 |
| Last Modified: | 07 Aug 2026 04:01 |
| URI: | http://repository.unj.ac.id/id/eprint/68978 |
Actions (login required)
![]() |
View Item |
