Table 1.
The curated data from the Twitter Dataset.
Fig 1.
Conventional (Traditional) Machine Learning model.
Fig 2.
Proposed model architecture Augment Machine learning model.
Fig 3.
Augmented Ensemble Model (AEM).
Table 2.
Health Term Text Data Augmentation (HTTDA) transformation for reduction to n-Classes.
Fig 4.
Health Terms Text Data Augmentation with the augmentation technique of “word substitution”.
Fig 5.
The PCA Health Terms Scatter plot of the devised model.
Fig 6.
The Disease Text class distribution before applying the HTTDA word vector.
Fig 7.
The disease text class distribution after the application of text data augmentation.
Table 3.
Accuracy of the classifiers using machine learning pipeline.
Table 4.
The F1 macro, weighted, and micro scores for various classifiers and a voting ensemble.
Fig 8.
A comparative plot of the results of the performance of the classifiers for respective F1 scores (macro, weighted, and micro).
Fig 9.
The plot of location-wise unique health terms per tweet over time.
Table 5.
Comparison with existing models.
Fig 10.
Country-wise aggregation of health terms from the tweet corpus.
Fig 11.
Comparative study of Accuracy as performance parameter.
Fig 12.
Illustrative Group Class to Class mapping for the Group Class “flu”.
Fig 13.
Illustrative Group Class to Class mapping for the Group Class “cancer”.