ABSTRACT: This talk presents aspects of the work of a statistical consultant. As a crash course, the focus is on topics that are not usually formally taught in university. These include data collection, data cleaning, visualizations, formulating questions and how to create a narrative from data analysis. Illustrative examples from government, industry and academic problems will be discussed.
Sous-titre:
Statistical consulting or the art of data storytelling
We will discuss measures to describe financial risk: credit risk, operational risk, insurance risk among others. In particular, we will introduce Value-At-Risk and Expected Shortfall. We will calculate risk measures for some simple probabilistic models. Next, we will address the issue of estimation of risk measures: parametric, semi-parametric and non-parametric methods will be introduced.
Sous-titre:
Mathematical Foundations of Financial Risk Management
Abstract: Judiciously coding the symbols in a file can lead to compression without losing information. We will discuss Huffman's solution to the optimal prefix-free coding problem.
We will also discuss the relationship between Shannon's entropy and the optimal coding problem.