AKBAR, MUHAMMAD FALDIAN (2025) ANALISIS SENTIMEN PENGGUNA TERHADAP APLIKASI AI DI GOOGLE PLAYSTORE MENGGUNAKAN ALGORITMA SUPPORT VECTOR MACHINE DAN NAIVE BAYES (STUDI KASUS: ChatGPT, Google Gemini, dan Microsoft Copilot). S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 COVER.pdf Download (526kB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
02 BAB 1.pdf Restricted to Registered users only Download (38kB) |
|
|
Text (BAB II)
03 BAB 2.pdf Restricted to Registered users only Download (224kB) |
|
|
Text (BAB III)
04 BAB 3.pdf Restricted to Registered users only Download (144kB) |
|
|
Text (BAB IV)
05 BAB 4.pdf Restricted to Registered users only Download (1MB) |
|
|
Text (BAB V)
06 BAB 5.pdf Restricted to Registered users only Download (42kB) |
|
|
Text (DAFTAR PUSTAKA)
07 DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (97kB) |
|
|
Text (LAMPIRAN)
08 LAMPIRAN.pdf Restricted to Registered users only Download (894kB) |
Abstract
The development of Artificial Intelligence (AI)-based applications such as ChatGPT, Google Gemini, and Microsoft Copilot has gained significant public attention, particularly in Indonesia. User reviews on the Google Play Store serve as an important data source to understand public perception of these applications. This study aims to perform sentiment analysis on user reviews of AI applications from the Google Play Store using two classification algorithms: Support Vector Machine (SVM) and Naïve Bayes, as well as to compare the performance of both algorithms. The research process begins with data collection through web scraping, gathering 6,000 reviews for each application, followed by manual data labeling into two sentiment categories: positive and negative. The data is then processed through several pre-processing stages, including cleaning text, case folding, tokenizing, stopword removal, and filtering non-alphabet characters. Feature weighting is carried out using the TF-IDF method. To address data imbalance issues, SMOTE (Synthetic Minority Over-sampling Technique) and Random Oversampling techniques are applied before building sentiment classification models using SVM and Naïve Bayes algorithms. The evaluation results show that the SVM algorithm consistently delivers better performance compared to Naïve Bayes, with the best results obtained from the combination of SVM and Random Oversampling on the Microsoft Copilot dataset. Furthermore, the sentiment distribution analysis reveals that positive sentiment reviews generally dominate negative sentiment reviews across all three applications. This indicates that most users provide positive feedback regarding the performance and features offered by these AI-based applications. This research is expected to contribute to the development of sentiment analysis related to AI applications and serve as a reference for developers in improving the quality and services of their products. Kata kunci: Sentiment Analysis, Artificial Intelligence, Support Vector Machine, Naïve Bayes, SMOTE, Random Oversampling, Google Play Store. Perkembangan aplikasi berbasis Artificial Intelligence (AI) seperti ChatGPT, Google Gemini, dan Microsoft Copilot telah menarik perhatian masyarakat, khususnya di Indonesia. Ulasan pengguna di Google Play Store menjadi sumber data penting untuk memahami persepsi masyarakat terhadap aplikasi-aplikasi tersebut. Penelitian ini bertujuan untuk melakukan analisis sentimen terhadap ulasan pengguna aplikasi AI di Google Play Store menggunakan dua algoritma klasifikasi, yaitu Support Vector Machine (SVM) dan Naïve Bayes, serta membandingkan performa kedua algoritma tersebut. Proses penelitian dimulai dengan pengumpulan data ulasan melalui teknik web scraping sebanyak 6.000 data untuk masing-masing aplikasi, kemudian dilakukan pelabelan data secara manual ke dalam dua kategori sentimen, yaitu positif dan negatif. Selanjutnya, data diolah melalui tahapan pra-pemrosesan yang meliputi cleaning text, case folding, tokenizing, stopword removal, dan filtering non-alphabet. Setelah itu, dilakukan pembobotan fitur menggunakan metode TF-IDF. Untuk mengatasi ketidakseimbangan data (class imbalance), diterapkan teknik SMOTE dan Random Oversampling sebelum membangun model klasifikasi dengan algoritma SVM dan Naïve Bayes. Hasil evaluasi menunjukkan bahwa algoritma SVM secara konsisten menghasilkan performa yang lebih baik dibandingkan Naïve Bayes, dengan hasil terbaik diperoleh pada kombinasi SVM dan Random Oversampling di dataset aplikasi Microsoft Copilot. Selain itu, dari hasil analisis distribusi sentimen, ditemukan bahwa ulasan dengan sentimen positif secara umum lebih dominan dibandingkan dengan sentimen negatif pada ketiga aplikasi yang diteliti. Hal ini menunjukkan bahwa mayoritas pengguna memberikan tanggapan positif terhadap aplikasi-aplikasi AI tersebut. Penelitian ini diharapkan dapat memberikan kontribusi dalam pengembangan analisis sentimen pada aplikasi berbasis AI serta menjadi referensi bagi pengembang dalam meningkatkan kualitas dan pelayanan produk mereka. Kata kunci: Analisis Sentimen, Artificial Intelligence, Support Vector Machine, Naïve Bayes, SMOTE, Random Oversampling, Google Play Store.
Actions (login required)
 |
View Item |
