In this project, we are designing novel methods in time-series analysis, information theoretic- feature extraction and optimization methods as well as game theoretic-based approaches. We are also combine these with an integrative clustering algorithm to predict parameters indicative of success and failure in weight loss and retention of loss in obese subjects. In addition, these tools can be applied to address other weight loss-related effects that have clinical relevance such as increase in physical activity and improvement in insulin sensitivity. We believe that such an approach can address the shortcomings in existing approaches and thereby produce a more accurate and reliable predictive model. We plan to design, test, and validate this system using the the unique infrastructure and databases of the Investigational Weight Management Clinic (IWMC), a demonstration unit of the NIH-funded Michigan Nutrition Obesity Research Center (MNORC), directed by Co-PI, Dr. Burant. The longitudinal databases with diverse time-series data create the ideal foundations to train, validate, and test the proposed computational methods.