2015年4月8日

Fundamentals of Modern Statistical Methods


日期:2015/04/07

The book Fundamentals of Modern Statistical Methods  by Rand Wilcox describes many situations where slight departures from normality create large distortions in the usual methods of statistical analysis. It recommends more robust procedures such as using the median, the trimmed median, bootstrap simulation, absolute deviations, running correlations, likelihoods, and something I hadn't seen before, M-estimators, to overcome what seem like trivial departures from normality like mixed normal distributions rather than single such.

The book is self contained and doesn't require a high level of previous mathematical or statistical background. It contains summaries of each chapter in five easy steps. It's a good primer and spark for improved methods of looking at data.

沒有留言: