Article Per Year (5 Year)
p-Index From 2019 - 2024
0.23
P-Index
This Author published in this journals
All Journal BAREKENG: Jurnal Ilmu Matematika dan Terapan
Agus Mohamad Soleh
Department of Statistics, Faculty of Mathematics and Natural Sciences, IPB University
Computer Science & IT Control & Systems Engineering Economics, Econometrics & Finance Energy Engineering Mathematics Mechanical Engineering Physics Transportation

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
TEXT CLUSTERING ONLINE LEARNING OPINION DURING COVID-19 PANDEMIC IN INDONESIA USING TWEETS Maulida Fajrining Tyas; Anang Kurnia; Agus Mohamad Soleh
BAREKENG: Jurnal Ilmu Matematika dan Terapan Vol 16 No 3 (2022): BAREKENG: Jurnal Ilmu Matematika dan Terapan
Publisher : PATTIMURA UNIVERSITY

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1033.901 KB) | DOI: 10.30598/barekengvol16iss3pp939-948

Abstract

To prevent the spread of corona virus, restriction of social activities are implemented including school activities which reaps the pros and cons in community. Opinions about online learning are widely conveyed mainly on Twitter. Tweets obtained can be used to extract information using text clustering to group topics about online learning during pandemic in Indonesia. K-Means is often used and has good performance in text clustering area. However, the problem of high dimensionality in textual data can result in difficult computations so that a sampling method is proposed. This paper aims to examine whether a sampling method to cluster tweets can result to an efficient clustering than using the whole dataset. After pre-processing, five sample sizes are selected from 28300 tweets which are 250, 500, 2500, 10000 and 20000 to conduct K-Means clustering. Results showed that from 10 iterations, three main cluster topics appeared 90%-100% in sample size of 2500, 10000 and 20000. Meanwhile sample size of 250 and 500 tend to produced 20%-60% appearance of the three main cluster topics. This means that around 8% to 35% of tweets used can yield representative clusters and efficient computation which is four times faster than using entire dataset.
Co-Authors Anang Kurnia Maulida Fajrining Tyas