Open AccessSentiment Analysis of Perpustakaan Nasional Republik Indonesia Through Social Media Twitter
Author(s) -
Fakhris Khusnu Reza Mahfud,
Nita Siti Mudawamah,
Wahyu Hariyanto
Publication year - 2020
Publication title -
matics
Language(s) - English
Resource type - Journals
eISSN - 2477-2550
pISSN - 1978-161X
DOI - 10.18860/mat.v12i1.8973
Subject(s) - social media , weighting , naive bayes classifier , computer science , precision and recall , preprocessor , information retrieval , world wide web , artificial intelligence , medicine , support vector machine , radiology
The library is a gate of science and a heart of civilization. Indonesia already has a Perpustakaan Nasional consisted of 27 floors and is equipped with facilities that are adequate for user needs. Apart from that, we need to see opinions from the community as users. Public opinion about the library is critical for library managers to evaluate services and facilities from the library. One way to find out the views of the community is by using social media twitter. Twitter social media is often used in channelling opinions or expressing opinions about specific topics; besides social media, twitter is commonly used for digital campaign movements. Submission of views and even digital campaigns on Twitter social media greatly influence the opinions and even behaviour of society in various ways. This study analyzes tweets about national libraries by classifying, positive opinions, negative opinions and neutral opinions. In this study, twitter data will go through the preprocessing, weighting, and classification stages. TF-IDF and TF binary are used in weighting in this study. The classification used in this study is Naive Bayes and KNN. Accuracy, precision, and recall values were also used in this study to evaluate classification performance. The highest classification performance using KNN classification with TF-IDF weighting resulted in the value of accuracy, precision, and recall of 83.33%, 79.2%, and 83.3% respectively.