Methods of knowledge discovery in tweets.

Area of Science:

• Social Media Analytics
• Data Mining
• Health Informatics

Background:

• Social media platforms like Twitter generate vast amounts of data relevant to public health.
• Understanding trends and sentiments related to health behaviors requires advanced analytical methods.
• Traditional research methods may not fully capture the dynamic nature of online health discussions.

Purpose of the Study:

• To detail web mining methodologies for knowledge discovery in Tweets.
• To demonstrate the application of these methods using the topic of physical activity.
• To provide a framework for analyzing social media data in health research.

Main Methods:

• Employed structure mining using social network analysis to identify macro-, meso-, and micro-level Tweet network structures.
• Utilized content mining with n-gram based text analysis and sentiment analysis to extract information from Tweet content.
• Highlighted specific web mining tools such as NodeXL, ORA, Pajek, and Weka for each stage of the process.

Main Results:

• Successfully applied novel web mining methods to gain insights into multiple dimensions of physical activity.
• Demonstrated the utility of structure and content mining for understanding complex health-related topics on social media.
• Identified key network structures and content themes within physical activity-related Tweets.

Conclusions:

• The described web mining methods offer a powerful approach for knowledge discovery in social media data.
• These techniques are valuable for researchers seeking to understand health-related behaviors and trends.
• The application to physical activity highlights the potential for broader use in social media health research.