Leveraging machine learning approaches for predicting potential Lyme disease cases and incidence rates in the United States using Twitter

Table 2 Performance comparison of NLP classification models on the test dataset. The highest score values are shown in bold

NLP Classification Model	Accuracy	F1 Score	Precision	Recall
Keyword-based labelling	0.84	0.86	0.97	0.77
Word2vec and XGBoost	0.76	0.75	0.78	0.73
\(\text {TF-IDF}\) and Logistic Regression	0.88	0.87	0.94	0.81
BERT	0.90	0.89	0.96	0.83
BERTweet	0.90	0.90	0.95	0.85

ISSN: 1472-6947