Sentiment Analysis Concerning the Thailand National Health Security Office by Using the Sentiment Analysis Model Based on Twitter

Sentiment Analysis concerning the Thailand National Health Security Office by using the sentiment analysis model based on Twitter. The research objectives were to: (1) hear information from people who express their opinions on Twitter about the National Health Security Office; (2) know public opinions about the National Health Security Office on Twitter; (3) assess the effectiveness of the opinion classification learning model. Data were collected from Twitter by specifying 4 keywords for extracting a total of 4314 comments, categorized as 3608 positive comments and 706 negative comments. Then input data into the model with a total of 3 algorithms: Logistic Regression, Naïve Bayes, and Random Forest with two feature extraction transformations: feature extraction with count vectorizer combined with good and bad word count in the vocabulary library and feature extraction with TF-IDF combined with good and bad word count in the vocabulary. It was found that the use of the Logistic Regression algorithm with feature extraction with TF-IDF combined with good and bad word count in the vocabulary. It is the most efficient model with the accuracy at 0.897.