Analisis Sentimen terhadap Ulasan Hotel Berbahasa Indonesia pada Implementasi Data TripAdvisor Menggunakan Algoritma Convolutional LSTM (CO-LSTM)

Prastyo, Ilham Dwi (2021) Analisis Sentimen terhadap Ulasan Hotel Berbahasa Indonesia pada Implementasi Data TripAdvisor Menggunakan Algoritma Convolutional LSTM (CO-LSTM). Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
06111740000106-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 April 2024.

Download (2MB) | Request a copy
[img] Text
06111740000106-Undergraduate_Thesis.pdf
Restricted to Repository staff only

Download (2MB) | Request a copy

Abstract

Pertumbuhan pesat teknologi internet meningkatkan jumlah pengguna internet dan penggunaannya yang relatif cukup lama. Perilaku ini merupakan peluang besar bagi industri e-commerce, khususnya Online Travel Agent(OTA). Salah satu layanan dari OTA adalah reservasi hotel online yang dilengkapi dengan fasilitas untuk mengulas hotel secara online. Ulasan tersebut biasa digunakan sebagai tolak ukur kepuasan pelanggan dan sumber informasi terhadap pelayanan yang perlu ditingkatkan. Oleh karena itu, analisis sentimen diperlukan dalam hal mengekstraksi fitur atau teks ulasan kedalam sentimen positif dan negatif sebagai bahan evaluasi kedepannya untuk pelayanan yang lebih baik dengan tujuan untuk meningkatkan customer dengan mendapat feedback dari ulasan tersebut. Dalam tugas akhir ini, pendekatan hybrid dari dua arsitektur deep learning yaitu Convolutional Neural Network (CNN) dan Long Short Term Memory (LSTM) digunakan untuk klasifikasi sentimen pada review text yang diambil dari situs TripAdvisor. Convolutional Neural Network sangat efektif dalam pemilihan fitur lokal, sedangkan Long Short Term Memory sering kali memberikan hasil yang baik dalam analisis sekuensial dari teks yang panjang. Penerapan BERT sebagai Word Embedding dan model hybrid Co-LSTM mampu merepresentasikan tiap kata dengan sangat baik. Pada penelitian ini proses pembelajaran mesin melibatkan total sebanyak 86.496 ulasan, dengan hasil evaluasi model memiliki nilai performansi sebesar 96,23%. Model Convolutional LSTM dengan BERT sebagai fitur encoder dapat memprediksi sentimen atau kecenderungan opini pada ulasan pelanggan terhadap reservasi hotel online Tripadvisor dengan baik. ================================================================================================ The rapid growth of internet technology increases the number of internet users and their use is relatively long. This behavior is a big opportunity for the e-commerce industry, especially Online Travel Agents (OTA). One of the services from OTA is online hotel reservation which is equipped with facilities to review hotels online. These reviews are commonly used as a measure of customer satisfaction and a source of information on services that need to be improved. Therefore, sentiment analysis is needed in terms of extracting features or review texts into positive and negative sentiments as material for future evaluations for better service with the aim of increasing customers by getting feedback from these reviews. In this final project, a hybrid approach of two deep learning architectures, namely Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM) is used to classify sentiments in review texts taken from the TripAdvisor site. Convolutional Neural Networks are very effective in local feature selection, while Long Short Term Memory often gives good results in sequential analysis of long texts. The application of BERT as Word Embedding and the Co-LSTM hybrid model is able to represent each word very well. In this study, the machine learning process involved a total of 86,496 reviews, with the results of the model evaluation having a performance value of 96.23%. The Convolutional LSTM model with BERT as an encoder feature can predict sentiment or opinion trends in customer reviews of Tripadvisor online hotel reservations well

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: TripAdvisor, review text, convolutional neural network, long short term memory, analisis sentimen.
Subjects: Q Science > Q Science (General) > Q325.5 Machine learning.
Q Science > QA Mathematics > QA336 Artificial Intelligence
Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science)
R Medicine > R Medicine (General) > R858 Deep Learning
Divisions: Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User: Ilham Dwi Prastyo
Date Deposited: 03 Feb 2022 01:24
Last Modified: 03 Feb 2022 01:24
URI: https://repository.its.ac.id/id/eprint/92661

Actions (login required)

View Item View Item