ANALISIS SENTIMEN PEMBERLAKUAN PEMBATASAN KEGIATAN MASYARAKAT (PPKM) DI TWITTER MENGGUNAKAN METODE LONG SHORT-TERM MEMORY
Total View This Week0

Institusion

Universitas Sriwijaya

Author

KRISYANI, LARASHATI
Abdiansah, Abdiansah
Utami, Alvi Syahrini

Subject

Q325.5 Machine learning 

Datestamp

2023-05-11 08:04:19 

Abstract :
CARE is a policy made by the government to deal with the Covid-19 pandemic that occurred in Indonesia. Twitter as one of the social networking sites is also not spared from the discussion on the CARE topic. Various public opinions regarding CARE on Twitter contain pros and cons because CARE has made a sudden change in people's lives. This study aims to analyze public sentiment towards CARE policies that have been made. Sentiment analysis uses the Long Short-Term Memory algorithm to classify sentiment into negative, neutral, or positive classes. The study used 12.959 data obtained from the results of scraping. The result of this study is a system that can classify sentiment. The level of accuracy obtained is 82% with an average macro precision value of 0,78, an average macro recall value of 0,76 and an average macro F1-score value of 0,77; and the weighted average value of precision, the weighted average value of recall, and the weighted average value of F1-score each is 0,82 on the test data. 

File :
RAMA_55201_09021181823021.pdf
RAMA_55201_09021181823021_TURNITIN.pdf
RAMA_55201_09021181823021_0001108401_0022127804_01_front_ref.pdf
RAMA_55201_09021181823021_0001108401_0022127804_02.pdf
RAMA_55201_09021181823021_0001108401_0022127804_03.pdf
RAMA_55201_09021181823021_0001108401_0022127804_04.pdf
RAMA_55201_09021181823021_0001108401_0022127804_05.pdf
RAMA_55201_09021181823021_0001108401_0022127804_06.pdf
RAMA_55201_09021181823021_0001108401_0022127804_06_ref.pdf
RAMA_55201_09021181823021_0001108401_0022127804_07_lamp.pdf