Optimization on machine learning based approaches for sentiment analysis on HPV vaccines related tweets

Area of Science:

• Computational linguistics
• Public health informatics
• Machine learning

Background:

• Low human papillomavirus (HPV) vaccine coverage is a public health concern.
• Understanding public sentiment on social media is crucial for developing effective vaccination strategies.
• Machine learning approaches can analyze large volumes of social media data to gauge public opinion.

Purpose of the Study:

• To develop and evaluate a machine learning system for comprehensive sentiment analysis of public opinion on HPV vaccines on Twitter.
• To improve the performance of sentiment analysis models for HPV vaccine-related social media data.

Main Methods:

• A dataset of 6,000 HPV vaccine-related tweets was manually annotated to create a gold standard corpus.
• A Support Vector Machine (SVM) model was employed with a hierarchical classification approach.
• Feature sets and model parameters were optimized to enhance classification performance.

Main Results:

• A 10-category hierarchical classification scheme was established for comprehensive public opinion analysis.
• The annotated corpus achieved a Kappa agreement of 0.851.
• The optimized hierarchical classification model improved F-scores from 0.6732 (micro) and 0.3967 (macro) to 0.7442 (micro) and 0.5883 (macro).

Conclusions:

• The developed system offers a systematic method to enhance machine learning model performance on imbalanced social media datasets concerning HPV vaccines.
• The approach can be scaled to analyze large-scale public opinion on HPV vaccines from extensive Twitter data.