Shafirra, Nadhifa Ayu (2020) Klasifikasi Sentimen Ulasan Film Indonesia Menggunakan Metode Convolutional Neural Network (CNN). Other thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
06211640000064-Undergraduate_Theses.pdf Download (3MB) | Preview |
Abstract
Ulasan film adalah sebuah opini yang bersifat subjektif. Ulasan film memiliki media yang beragam, seperti tulisan, audio, dan video. Ulasan film dapat diolah dengan menggunakan klasifikasi sentimen, agar ucapan seseorang terkait film dapat ditentukan sebagai sentimen tertentu. Di masa sekarang, data memiliki berbagai bentuk, pemilihan jenis data yang lebih baik juga dapat mem-pengaruhi klasifikasi sentimen. Data video dapat diekstraksi menggunakan Mel-Frequency Cepstral Coefficients (MFCC) dan dikonversi menjadi data teks dengan bantuan Speech-to-Text (STT). Fitur Ekstraksi MFCC digunakan karena keunikan skala Mel memungkinkan audio yang ada dibedakan berdasarkan suaranya. Data teks digunakan karena kata atau kalimat dapat dibedakan secara negatif atau positif. Sehingga, kedua jenis data tersebut dibandingkan dan dipilih model dengan jenis data yang memiliki hasil klasifikasi terbaik. Dengan menggunakan metode Convolutional Neural Network, didapatkan bahwa data teks memiliki nilai AUC lebih baik dibandingkan data MFCC. Model terbaik yang dipilih adalah model klasifikasi sentimen dengan data teks yang dimodelkan berdasarkan aspek dari penilaian film.
=================================================================================================================================
Movie review is an opinion that is subjective. Movie reviews have a variety of media, such as writing, audio, and video. Movie reviews can be processed using sentiment classification, so that someone's sayings related to a particular film can be determined as certain sentiments. At present, data takes a variety of forms, and selection of better data types can also affect sentiment classification. Video data can be extracted using Mel-Frequency Cepstral Coefficients (MFCC) and converted to text data with the help of Speech-to-Text (STT). The MFCC extraction feature is used because the uniqueness of the Mel scale allows existing audio to be differentiated based on sound. Text data is used because words or sentences can be distinguished negatively or positively. Thus, the two types of data are compared and the model with the best classification results is selected. By using the Convolutional Neural Network method, it was found that the text data has better AUC scores than MFCC data. The best model chosen is the sentiment classification model with text data that are modeled based on aspects of movie evaluation.
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSSt 519.53 Sha k-1 2020 |
| Uncontrolled Keywords: | Convolutional Neural Network, MFCC, Speech-to-Text, Ulasan Film |
| Subjects: | H Social Sciences > HA Statistics Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Nadhifa Ayu Shafirra |
| Date Deposited: | 25 Apr 2024 11:16 |
| Last Modified: | 25 Apr 2024 11:16 |
| URI: | http://repository.its.ac.id/id/eprint/73683 |
Actions (login required)
 |
View Item |