Aspect Level Sentiment Analysis on Zoom Cloud Meetings App
Review Using LDA

Aspect Level Sentiment Analysis on Zoom Cloud Meetings App
Review Using LDA

LDA, SVM, review, aspect

During the Covid-19 pandemic, almost all community activities are conducted from home. Therefore, video conference
technology is needed for people to carry out their normal activities from home. One of the video conference applications is
ZOOM Cloud Meetings. Applications certainly have been reviewed given by their users as a reference for new users and
companies of the application to know the application’s performance. However, in reviews, some constraints are the number of
reviews as well as irregular. Therefore, a solution is needed with sentiment analysis that aims to classify the reviews of the
application to be organized by categorizing positive or negative sentiment. In this study, aspect-based sentiment analysis was
conducted on ZOOM Cloud Meetings app reviews from Google Play Store. The analysis’s result of the review data obtained
three aspects, namely aspects of usability, system, and appearance. The modeling topic used is the Latent Dirichlet Allocation
(LDA) method and classification using the Support Vector Machine (SVM). This research resulted in the best performance with
the best parameters resulting in the performance accuracy of usability aspect is 88.83%, system aspect with 91.2%, appearance
aspect with 94.78%, and performance accuracy of all aspects 91.61%.

Janu Akrama Wardhana1
, Yuliant Sibaroni2

Telkom University

Telkom University

Fajar bagus W

PDF

Indonesia

Text