SENTIMENT ANALYSIS OF GENSHIN IMPACT ON TWITTER USING NAÏVE BAYES

MARETHA FITRIE PURUHITA

Informasi Umum

Kode

23.04.5132

Klasifikasi

006.312 - Data mining

Jenis

Karya Ilmiah - Skripsi (S1) - Reference

Subjek

Data Mining

Dilihat

422 kali

Informasi Lainnya

Abstraksi

<p>During the COVID-19 pandemic that interferes normal life around the world, people have an obligation to stay at home and quarantine themselves. This has led to an increase in the consumption of entertainment, especially online gaming which is known to be less harmful than other stress and aversive emotions. And Genshin Impact is one of the online games that won Google Play's the Best Game of 2020 award when pandemic happening. Released in September 2020 by China video game developer, miHoYo. Co., Ltd, Genshin Impact has been a hot trend on the microblogging platform, Twitter. The purpose of this research is to provide information regarding people's opinion emotion in their tweets toward Genshin Impact and this information will be a helpful resource for game improvement and can be used as reference of future research. By using sentiment analysis to help analyze the emotion contained in the text, the result will be categorized into three categories: positive, negative, or neutral sentiment. The data is gained through text mining then will be processed as text classified using Naive Bayes algorithm. Thus, the model will be going through evaluation of model's performance to measure how accuracy it is. The result of it stated that the best ratio between training and test set is 60:40 with 71.80% test accuracy, yet the accuracy between 3 others ratio is not much difference. That’s why using hyperparameter tuning can find the optimal result. After finding the optimal result, the highest result it can get is 72.14%. Besides that, people on Twitter mostly perceive the game in neutral sentiment.</p>