FEATURE ANALYSIS OF PUBG MOBILE APPLICATION BY USING SENTIMENT ANALYSIS AND TOPIC MODELLING: COMPARISON OF PRE AND POST UPDATE

NAOMI BONTOR MARIA GIOVANNI

Informasi Umum

Kode

20.04.3803

Klasifikasi

793.4 - Games of actions

Jenis

Karya Ilmiah - Skripsi (S1) - Reference

Subjek

Games

Dilihat

50 kali

Informasi Lainnya

Abstraksi

User review is one of medium for customer to express their thoughts about a particular product. The reviews are useful for the company to determine customer insight about their product as the comment load customer opinion about the product from various angles. This research examines the opinion around the user in form of sentiment and topic toward development of PUBG season 10 and season 11 from review section on Google Play. Data for this study gathered by scraping tools on Google Play website using Octoparse application. The method of this study is quantitative descriptive using sentiment analysis and topic modelling. The results of this study obtained by sentiment analysis through the Naïve Bayes classifier method that can be used to classify consumer sentiment precisely and the Latent Dirichlet Allocation (LDA) proven to be able extracting customer insight about the topics within the comments. The results of this research expected to obtain valuable information for the company to solve problems that occurs, develop preventive strategies and plan future development based on customer voice for the next product launch in order to avoid the same mistakes.

Keywords: Sentiment Analysis, Topic Modelling, Latent Dirichlet Allocation, Big Data, Content Analysis, Google Play.