Design Text Mining for Anxiety Detection using Machine Learning based-on Social Media Data during COVID-19 pandemic
Yuli Fauziah, Shoffan Saifullah, Agus Sasmito Aribowo
Abstract
The COVID-19 pandemic has a profound impact on all groups, including governments, agencies, and individuals. It can make anxiety have a bad effect. So it is necessary to detect the existence of anxiety from the government to suppress and improve the community's psychology. This research aims to design text mining to detect anxiety during a pandemic by applying machine learning technology. Two methods of machine learning are designed, namely, random forest and xgboost. This design uses a sample of data from You Tube comments with a total of 4862 consisting of 3211 for negative data and 1651 for positive data. Negative data identify anxiety, while positive data identifies hope (not worry). The design of the application of this method was carried out by preliminary testing with three calculations, namely accuracy, precision, and recall. The accuracy of the Random Forest and XGB OOST methods is 83% and 73%. Meanwhile, precision and recall have an inversely proportional value. Random Forest has a precision value greater than 45% compared to xgboost. Whereas Recall, XGBOOST is bigger than ten compared to Random Forest. Random Forest can reference machine learning methods to detect someone's anxiety based on data from social media.