I have been assigned two brand-new Ph.D. students from our Financial Services Analytics program to work under my guidance for their first two semesters here at the University of Delaware. This is my list of tips, tools, and resources for… Continue Reading
Dr. Tufte thinks so clearly about both the beauty and pitfalls of data analysis. This is a must watch:
There are two types of uncertainty in this world that make predictions difficult. One type is called epistemic, this is the potentially reducible uncertainty which motivates the gathering of more data (e.g. demand for Apple’s latest innovation that could flop or soar).… Continue Reading
“There is, in fact, no way to represent any causal information in contingency tables (such as Tables 1.1 and 1.2), on which statistical inference is often based. There are, however, extra-statistical methods that can be used to express and interpret causal assumptions. These… Continue Reading
In a post by Kim Larsen, the interpretability advantage of Bayesian models is discussed with examples provided for using the bsts package in R: Sorry ARIMA, but I’m Going Bayesian