Using social networks to predict changes in health: Extended abstract

Abstract

Social networking sites not only have billions of users but detailed content about each individual's daily life. This detailed information about a person's life could be exploited to allow individuals to learn more about themselves. In this paper, we introduce the concept of using social networks to foresee changes in an individual's health. We develop a new model that can predict if a person has recently undergone weight loss by analyzing the text from the person's tweets. Sentiment analysis, parts-of-speech (POS) tagging, and categorization are used in this model. The model is tested on Twitter users and a good statistical accuracy is observed. The success of this model suggests that this idea could be further explored to identify other patterns and create new models for a variety of health changes and health problems, particularly those that are of huge interest to individuals and businesses.