ANALISIS SENTIMEN TERHADAP PEMILU PRESIDEN 2024 BERDASARKAN OPINI PENGGUNA TWITTER (X) MENGGUNAKAN ALGORITMA LONG SHORT TERM MEMORY DENGAN HYPERPARAMETER TUNING

TAUFIIQUL HAKIM, . (2025) ANALISIS SENTIMEN TERHADAP PEMILU PRESIDEN 2024 BERDASARKAN OPINI PENGGUNA TWITTER (X) MENGGUNAKAN ALGORITMA LONG SHORT TERM MEMORY DENGAN HYPERPARAMETER TUNING. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

[img] Text
cover.pdf

Download (890kB)
[img] Text
Bab 1 Pendahuluan.pdf

Download (376kB)
[img] Text
Bab 2 Landasan Teori.pdf
Restricted to Registered users only

Download (517kB) | Request a copy
[img] Text
Bab 3 Metodologi.pdf
Restricted to Registered users only

Download (466kB) | Request a copy
[img] Text
Bab 4 Hasil Penelitian.pdf
Restricted to Registered users only

Download (883kB) | Request a copy
[img] Text
Bab 5 Kesimpulan dan Saran.pdf
Restricted to Registered users only

Download (328kB) | Request a copy
[img] Text
Daftar Pustaka.pdf

Download (333kB)
[img] Text
Lampiran.pdf
Restricted to Registered users only

Download (2MB) | Request a copy

Abstract

Pemilihan umum merupakan perwujudan nyata demokrasi di Indonesia, di mana masyarakat memiliki peran penting dalam memilih pemimpin negara. Pemilu Presiden 2024 menjadi topik diskusi utama di media sosial, khususnya Twitter (X). Penelitian ini bertujuan untuk menganalisis sentimen masyarakat terhadap Pemilu Presiden 2024 menggunakan algoritma Long Short-Term Memory (LSTM) dengan metode hyperparameter tuning. Data diperoleh dengan teknik crawling data, rentang waktu data yang diambil mulai dari bulan September 2023 sampai bulan Agustus 2024 dan didapatkan data 1056 tweet. Hasil analisis sentimen dengan 1.056 data tweet berbahasa Indonesia dengan model LSTM tanpa hyperparameter tuning menghasilkan akurasi sebesar 73%, dengan presisi 78%, recall 83%, dan F1-score 81%, sedangkan penerapan hyperparameter tuning menghasilkan akurasi lebih rendah, yaitu 71%. Temuan ini menunjukkan bahwa hyperparameter tuning tidak selalu meningkatkan performa model, terutama pada dataset kecil dengan distribusi data yang tidak seimbang. Secara keseluruhan, algoritma LSTM terbukti mampu memberikan hasil yang cukup baik dalam menganalisis sentimen masyarakat terhadap Pemilu Presiden 2024 ***** General elections are a tangible manifestation of democracy in Indonesia, where people have an important role in choosing the country's leaders. The 2024 Presidential Election has become a major topic of discussion on social media, especially Twitter (X). This research aims to analyze public sentiment towards the 2024 Presidential Election using the Long Short-Term Memory (LSTM) algorithm with the hyperparameter tuning method. The data was obtained using the data crawling technique, the time span of the data taken was from September 2023 to June 2024 and 1056 tweets were obtained. The results of sentiment analysis with 1,056 Indonesian tweet data with the LSTM model without hyperparameter tuning resulted in an accuracy of 73%, with 78% precision, 83% recall, and 81% F1-score, while the application of hyperparameter tuning resulted in a lower accuracy of 71%. This finding shows that hyperparameter tuning does not always improve model performance, especially on small datasets with unbalanced data distribution. Overall, the LSTM algorithm proved to be able to provide quite good results in analyzing public sentiment towards the 2024 Presidential Election

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 25937 not found.
Date Deposited: 11 Feb 2025 08:54
Last Modified: 11 Feb 2025 08:54
URI: http://repository.unj.ac.id/id/eprint/52681

Actions (login required)

View Item View Item