COMPLEXITY (STRATIFICATION, ROTATION)
SCALING
SIMPLIFICATION
(PARAMETERIZATION)
HOW VARIABLES ARE SELECTED
(SYSTEM IDENTIFICATION)
ROLE OF UNMEASURED VARIABLES
(NO SIGNALS REACH US…)
WHAT FEATURES ARE CONSTANT
WHAT MATHEMATICAL FORMS ARE
USED (DYNAMICAL)
ERRORS IN INITIAL
CONDITIONS
QUALITY OF DATA,
HETERDGENGITY
ERRORS IN FORECASTS OF
EXOGENOUS VARIABLES
EXTRAPOLATIONS BEYOND
MEASURED RANGES
INTRINSIC VARIABLITY
EXTREME SYSTEM SENSITIVITY
“MANY ®
ONE” PROBLEM
COST (LENGTH, TIME, $$)
WORKING BACK FROM SURVIVORS
(“THE DISAPPEARED")
COMPUTATIONAL
LIMITS
LIMITED REALIZATIONS, SMALL
SAMPLES
TOO MUCH DATA
CULTURE, TABOOS,
INHIBITIONS
CONTRACT KNOWING
---
The GOOD NEWS
1) STATISTICAL
DESCRIPTIONS, PROBABLITIES (in leu of
“DETAILED KNOWLEDGE”)
2) DRIVES (EVOLUTIONARY)
FOR DATA
GATHERING
FOR DATA PROCESSING
--
NEED FOR EFFICIENT FILTERS
\ BUILD ROBOT
MATHEMATICIANS