Traditional business analytics have so far focused mostly on descriptive analyses of historical data using a myriad of sound statistical techniques. The trend is now strongly towards prediction. The field of data fusion has been around since the mid-eighties with a purpose of extracting intelligence from multiple data sources and thereby providing associated decision support. The field has become widespread over the years due to the ubiquity of sensor devices in our everyday lives and an exponential proliferation of both structured and unstructured content.
This seminar draws a close parallel between the fields of analytics and data fusion. It will then show how numerical statistical techniques can be augmented/enriched with those symbolic artificial intelligence, machine learning and control theoretic techniques that are employed for data fusion. The talk will give a high-level overview of architectures, state-of-the-art techniques, practical tools in both fields, and present example demos from the risk and health assessment domain.
The content will be prepared in a highly non-technical manner with a general audience in mind.
- Analytics managers who want an overview of the field without too much technical details
- Analytics and data fusion “want to be” practitioners from both academia and industry
Subrata Das with contributions from other team members.
Lesson 1: Analytics and fusion are defined synergistically
Lesson 2: Market size, big players, tools, and all that
Lesson 3: Architectures and analytics process design
Lesson 4: Good old statistics, tools, and scope for augmentation
Lesson 5: Knowledge acquisition bottleneck in artificial intelligence
Lesson 6: Emergence of machine learning, tools and hybridization
Lesson 7: Unstructured data/text analytics revisited
Lesson 8: Data vs. models matrix with project entries
Lesson 9: Conclusions, gaps, and opportunities .