Deep Learning & LSTMs for Industrial IoT and 'Zero Unplanned Downtime'
Apply Deep Learning (Long Short Term Memory Networks) to a new type of problem (Industrial Internet of Things) to predict Machine Stops 15 minutes before they occur. Key academic concepts: 1. Handling Imbalanced Data in LSTMs; 2. Balancing Look Ahead time and Look back time; 3. Re-sampling Time Series data; 4. Avoiding over-fit and Time-Series Cross-Validation.
Naresh R. Shah is a Data Engineer at Procter & Gamble Co., where is responsible for setting up Data Science team for utilizing IoT systems in order to improve productivity and drive efficiency of manufacturing systems in Beauty Product segment. He has a special interest in novel use cases of Natural Language Processing, Social Network Analytics and their application to Marketing. He’s a beta tester at the forefront of using and advocating new data science tools at organizations. Naresh holds a Masters in Business Analytics and Big Data degree from IE Business School, Madrid. At IE Business School, he participated, led teams and won several events like IE Data Dive, Hotelbeds Big Data Challenge, and Procter and Gamble’s IT Business Challenge. He was also part of the winning team at the first Forbes Hackathon, in New York.