Experimental design
Abstract
Meta-analytics does not substitute mathematics for experiments; in fact, for ground truth data, meta-analytics can be used to substitute experiments for the mathematics. Meta-analytic systems incorporate empirical data readily and also avoid some of the overtraining traps of many traditional approaches to the two broad areas of clustering classification and prediction regression. Most importantly, however, meta-analytics allows two or more experimental designs to be compatible within an overall system. The resulting “system of systems” incorporates features such as data normalization; the pruning of aging data; and the use of interfacing, gain, domain normalization, and sensitivity analysis tools to provide ...
Get Meta-Analytics now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
