Universal Probability and Its Applications in Data Science

4 credits

In many data science problems, there is only limited information on statistical properties of the data. This course develops the concept of universal probability that can be used as a proxy for the unknown distribution of data and provides a unified framework for several data science problems, including compression, portfolio selection, prediction, and classification. Prerequisites: ECE 225A or ECE 250 with a grade of B– or above; graduate standing.