n this session, Srinivasan explained the workflow and architecture adopted in a live project for identifying and applying ML algorithms for predicting Energy consumption across multiple machines in a manufacturing unit. He stressed upon checking out for algorithms other than the ones which are popular and widely used, for better accuracy. In this project, Facebook Prophet algorithm was explored in addition to LSTM and ARIMA. Srinivasan concluded the session by responding to multiple questions on transfer learning, examples of domain specific outliers, weightage for over & under forecasting, etc.

Thank you Srinivasan for this insightful session.