S. Baghi - Department of Computer Science, Science and Research Branch, Islamic Azad University, Tehran, Iran
M.H. Ebrahimzadeh - Information Technology Entrepreneurship, Faculty of Entrepreneurship, Tehran University, Tehran, Iran
N. Hedayati - Department of Computer Engineering, South Tehran Branch, Islamic Azad University, Tehran, Iran

Aims: Eating disorders are making a point of challenge for health-related researches. Using big data for this type of researches can effectively help researchers use a beneficial resource of information worldwide in real-time. This study aimed to introduce a more accurate index for analyzing food-related data and making relations between peoplechr('۳۹')s opinions and the prevention treatments for eating disorders. Instrument & Methods: In this data mining study, more than ۲ million eating-related tweets were collected from Twitter in ۲۰۱۷ and analyzed by novel methods for big data research. Three main indicators (Basic-sentiment-rate, Health-rate, and Relation-rate) were used to predict if every user is more likely to have a healthy or unhealthy diet. Finally, these parameters were normalized, clustered, and combined to obtain an overall sentiment rate. Findings: Location and gender were estimated as effective indicators making the relationship between peopleschr('۳۹') opinion and prevention treatments for eating disorders. Some combinations of factors were also considered influencing indicators when applied together, such as gender+age and gender+location. Conclusion: Punishment/reward combination criteria that are predicted with both gender and location data by FSR index is the most effective factor in making the relationship between peopleschr('۳۹') opinion and prevention treatments for eating disorders.

Eating Disorders, Prevention, Data Mining, Big Data, Twitter