(Korespondensi) Sentiment Analysis of Madura Tourism in New Normal Era using Text Blob and KNN with Hyperparameter Tuning

Tourism during the Covid-19 pandemic has
paralysis, even though tourism is a source of regional income. In
the new normal period, tourism began to rise again. Madura
Tourism Sentiment Analysis is needed for regional parties and
tourism developers to find a public opinion about tourism places
in Madura that have been vacuumed for a long time. The dataset
used is opinion data on Twitter for nature, culinary and
religious tourism in Madura. Data was taken during the New
Normal period between April 2020 to August 2021. This
research compared Manual Lexicon Based and TextBlob for
labeling data. TF-IDF for term weighting. SVM, Naïve Bayes,
and KNN methods with Tuning Parameters are compared for
classification methods in sentiment analysis. Based on this
research, the best Accuracy value is 94% for SVM Method or
KNN Method using Manhattan measure and K-Value = 1. The
most positive labels are obtained for three tourism categories:
nature, culinary, and religious.