UTOMO, SATRIO TRI (2025) ANALISIS SENTIMEN PADA MEDIA SOSIAL TWITTER MENGENAI QUICK COUNT MENGGUNAKAN ALGORITMA MACHINE LEARNING DAN DEEP LEARNING. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 COVER.pdf Download (467kB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
02 BAB I.pdf Restricted to Registered users only Download (231kB) |
|
|
Text (BAB II)
03 BAB II.pdf Restricted to Registered users only Download (696kB) |
|
|
Text (BAB III)
04 BAB III.pdf Restricted to Registered users only Download (280kB) |
|
|
Text (BAB IV)
05 BAB IV.pdf Restricted to Registered users only Download (732kB) |
|
|
Text (BAB V)
06 BAB V.pdf Restricted to Registered users only Download (156kB) |
|
|
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (234kB) |
|
|
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (673kB) |
Abstract
Sentiment analysis on social media, especially Twitter, can provide valuable insights into public perceptions of certain topics, including quick counts during the 2024 Indonesian elections. This study aims to analyze public sentiment based on tweets related to quick counts using various machine learning and deep learning algorithms, including Naive Bayes, Support Vector Machine (SVM), Logistic Regression, Long Short-Term Memory (LSTM), Recurrent Neural Network (RNN), and Bidirectional LSTM (BiLSTM). Data were collected using the tweet harvest library with the keyword quick count. Experiments were conducted with two approaches, namely preprocessing with and without stemming, to evaluate their impact on model performance. The performance of each model was evaluated using accuracy, precision, recall, and F1-score metrics. The results showed that the machine learning model, namely the SVM model with the stemming technique, gave the best results with an accuracy value of 81.66%. The deep learning model, namely Bidirectional LSTM with the stemming technique, gave the best accuracy of 81.07%. Other models, such as Naive Bayes, Logistic Regression, RNN, and LSTM, improved accuracy when stemming was applied. This study highlights the importance of preprocessing stages, including stemming, in improving model performance, especially on unstructured text data such as tweets. Keywords: Analysis Sentiment, Quick Count, Machine Learning, Deep Learning, Stemming Analisis sentimen di media sosial, khususnya Twitter, dapat memberikan wawasan berharga mengenai persepsi publik terhadap topik tertentu, termasuk tentang quick count selama pemilu di Indonesia pada tahun 2024. Penelitian ini bertujuan untuk menganalisis sentimen publik berdasarkan tweet terkait quick count menggunakan berbagai algoritma machine learning dan deep learning, yaitu Naive Bayes, Logistic Regression, Support Vector Machine (SVM), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), dan Bidirectional LSTM (BiLSTM). Data dikumpulkan menggunakan library tweet harvest dengan kata kunci quick count. Eksperimen dilakukan dengan dua pendekatan, yaitu preprocessing dengan dan tanpa stemming, untuk mengevaluasi dampaknya terhadap performa model. Kinerja dari masing-masing model di evaluasi menggunakan metrik accuracy, precision, recall, dan f1-score. Hasil menunjukkan bahwa model machine learning, yaitu model SVM dengan teknik stemming memberikan hasil terbaik dengan nilai akurasi sebesar 81.66%. Sedangkan model deep learning, yaitu Bidirectional LSTM dengan teknik stemming memberikan akurasi terbaik dengan akurasi sebesar 81.07%. Model-model lainnya seperti Naive Bayes, Logistic Regression, RNN, dan LSTM juga mengalami peningkatan akurasi saat teknik stemming diterapkan. Studi ini menegaskan pentingnya tahapan preprocessing, termasuk stemming, dalam meningkatkan performa model, khususnya pada data teks yang tidak terstruktur seperti tweet. Kata Kunci: Analisis Sentimen, Quick Count, Machine Learning, Deep Learning, Stemming
Actions (login required)
 |
View Item |
