With the growth of machine learning and artificial intelligence, you may think we have more and better analytic insight than ever before. We do, but there’s a catch. The models used by data mining and deep learning use the language of probability and inference. We may be used to sales forecasts and budget predictions, but in the future our data will be more like weather forecasts and sporting predictions – open to change and chance.
In this session, we explore what business and IT need to know about these new analytics. We will consider the advantages and pitfalls of building decision support in an uncertain world, looking at issues with data quality, data visualization and regulatory compliance as relevant issues. The session will cover:
- Why machine learning is different to BI
- Sources of uncertainty in machine learning
- Probability and Data Literacy
- Communicating probabilities
- Uncertainty and visualization