PERANCANGAN PROTOTIPE SISTEM ANALISIS SENTIMEN ULASAN KONSUMEN APLIKASI TIKET.COM

• Main Author: Raihan Dzaky Musyaffa, Muhammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/77763
• Institution: Institut Teknologi Bandung

PT Global Tiket Network, widely known as tiket.com is a pioneering online travel agent (OTA) company in Indonesia. Established in 2011, tiket.com has become one of the largest OTA companies in Indonesia. The company provides its services through mobile applications (available on Android and iOS). From mid-2022 to mid-2023, the tiket.com app had an average rating of 2.33 out of a maximum scale of 5. The low rating reflects consumers' perception of the app. However, the company did not focus on consumer reviews of their app as a decision-making basis for service development and improvement. As a result, there is a gap between consumers' expectations of the app and its actual condition. Therefore, this research aims to produce a sentiment analysis model and a prototype system that can perform sentiment analysis on text data of consumer reviews of the tiket.com app. This research utilizes the CRISP-DM framework. The outcome of this research is the development of a sentiment analysis model and a prototype application. In order to build the model, three algorithms were utilized, namely Naïve Bayes (NB), Support Vector Machine (SVM), and Bidirectional Long Short-Term Memory (Bi-LSTM). The data used for this research was sourced from consumer reviews of the tiket.com app on Google Play. Upon evaluation, the SVM model proved to perform the best, with an accuracy rate of 88.86%, surpassing both the NB and Bi-LSTM models. As a result, the SVM model was selected as the analysis model for the prototype design process. The sentiment analysis application prototype is a tool that assists users in capturing, preprocessing, classifying, and visualizing app review data. The data visualization component includes a word cloud, word frequency graph, and N-gram frequency graph. The prototype was developed using the Streamlit library in Python programming language. The company can use this prototype to aid in decision-making related to app development and future improvements. By incorporating the prototype into the decision-making process, businesses can create applications that better meet the needs of their consumers.