Epidemic outbreak detection is a key challenge in the healthcare domain for governing bodies. Problem:  Customer needed a solution for predicting the probability and magnitude of disease outbreaks before they occur. Methodology: We created an innovative predictive analytics solutions with the below methodology:

1. Pathogens are mapped to Etiology
2. Etiology progression modelled
   
3. Hierarchical Clustering of the states into outbreak regions
4. Inter-cluster and intra-cluster Bayesian outbreak probabilities calculated
   

Results: ROC True vs False: 0.86666667 Key concepts used: Hierarchical Clustering, Bayesian networks, SIRS, GAM Model, Decision Trees, Time-series modeling so on. Domain: Healthcare Analytics