SENTIMENT ANALYSIS OF SOCIAL MEDIA TWITTER USING NAIVE BAYES ALGORITHM CASE: COVID-19 IN INDONESIA

MELANI SINAGA

Informasi Umum

Kode

20.04.4084

Klasifikasi

000 - General Works

Jenis

Karya Ilmiah - Skripsi (S1) - Reference

Subjek

Algorithms

Dilihat

47 kali

Informasi Lainnya

Abstraksi

Severe acute respiratory syndrome coronavirus 2 or called COVID-19 is a new type of Virus found in Wuhan, Hubei Province, China in December 2019, this new type of disease is encountered in humans, which is transmitted between animals and humans and then spread to humans others quickly, primarily through the sparks of respiratory symptoms that often occur namely coughing, hot body temperature, coughing shortness of breath and difficulty breathing, until now there has been no vaccine and drugs found to treat patients who are positive the COVID-19 Virus can only focus on reducing symptoms, transmission from human to human is so fast that COVID-19 has spread throughout the world, to prevent it by only keeping clean, washing hands, keeping distance from others, closing smoothly when coughing and using masks everywhere, the death rate of COVID-19 very high, therefore many people are concerned about the spread of the COVID-19 Virus so see a trend community response to COVID-19 will be sentiment analysis using social media Twitter know how the reaction of the Indonesian people in dealing with the COVID-19 pandemic, arch on COVID-19 sentiment analysis in Indonesia and uses three weighting comparisons, therefore research will conduct research on COVID-19 sentiment analysis in Indonesia, this study uses the Naïve Bayes algorithm and uses three weights namely TF-IDF dan TF as well as doing four times the ratio of the ratio of 65:35, 70:30, 75:25 and 80:20 to produce accuracy, this study also uses Python machine learning and the implementation of this algorithm will then be tested using the Confusion Matrix to get precise accuracy results. So that this research has produced an accuracy of the 80:20 ratio with the best weighting using TF-IDF. With an accuracy of 81.91% obtained from the testing data which resulted in d 670 comments showing positive reactions, 568 getting neutral reactions and 100 comments showing negative so that the tendency of Twitter users comments to be more positive towards COVID-19 Virus in Indonesia this study is still very minimal due to the bound journals and data COVID-19 because COVID-19 is still ongoing and even under research.

Key Words: COVID-19, Media Social, NaiveBayes, Sentiment Analysis, Confusion matrix, TF-IDF, TF.