Analisis sentimen terhadap penurunan tingkat pernikahan di Indonesia menggunakan algoritma Bidirectional Encoder Representations from Transformers (BERT)

Hamidah, Sarah Yusti (2025) Analisis sentimen terhadap penurunan tingkat pernikahan di Indonesia menggunakan algoritma Bidirectional Encoder Representations from Transformers (BERT). Sarjana thesis, UIN Sunan Gunung Djati Bandung.

Preview	Text 1_cover.pdf Download (190kB) | Preview
Preview	Text 2_abstrak.pdf Download (178kB) | Preview
Preview	Text 3_skbebasplagiarism.pdf Download (155kB) | Preview
Preview	Text 4_daftarisi.pdf Download (197kB) | Preview
Preview	Text 5_bab1.pdf Download (304kB) | Preview
	Text 6_bab2.pdf Restricted to Registered users only Download (391kB) | Request a copy
	Text 7_bab3.pdf Restricted to Registered users only Download (583kB) | Request a copy
	Text 8_bab4.pdf Restricted to Registered users only Download (9MB) | Request a copy
	Text 9_bab5.pdf Restricted to Registered users only Download (190kB) | Request a copy
	Text 10_daftarpustaka.pdf Restricted to Registered users only Download (216kB) | Request a copy

Abstract

INDONESIA: Penurunan angka pernikahan di Indonesia menjadi isu sosial yang menarik untuk dianalisis melalui opini publik di media sosial. Penelitian ini menggunakan metode Sample, Explore, Modify, Model, and Assess (SEMMA) dengan algoritma Bidirectional Encoder Representations from Transformers (BERT) untuk melakukan analisis sentimen komentar YouTube terkait topik tersebut. Data dikumpulkan menggunakan teknik crawling dan diolah menggunakan text preprocessing serta Easy Data Augmentation (EDA) untuk penyeimbangan kelas. Model dilatih menggunakan delapan skenario, dengan skenario kedua memperoleh akurasi tertinggi sebesar 98.52%, F1-score 98.52%, precision 98.55%, dan recall 98.52% pada data validasi. Hasil analisis sentimen mengungkapkan kata kunci dominan yang berkaitan dengan nilai keluarga, ekonomi, dan pandangan sosial. Temuan ini diharapkan dapat menjadi rujukan dalam perumusan kebijakan dan program literasi pernikahan yang relevan dengan kondisi masyarakat Indonesia. ENGLISH: The decline in marriage rates in Indonesia has become a significant social issue that warrants analysis through public opinion on social media. This study employs the Sample, Explore, Modify, Model, and Assess (SEMMA) methodology combined with the Bidirectional Encoder Representations from Transformers (BERT) algorithm to perform sentiment analysis on YouTube comments related to the topic. Data were collected using crawling techniques and processed through text preprocessing as well as Easy Data Augmentation (EDA) to balance class distributions. The model was trained using eight scenarios, with the second scenario achieving the highest performance: 98.52% accuracy, 98.52% F1-score, 98.55% precision, and 98.52% recall on the validation set. Sentiment analysis results revealed dominant keywords associated with family values, economic factors, and social perspectives. These findings are expected to serve as a reference for formulating policies and marriage literacy programs relevant to Indonesian society.

Actions (login required)

View Item