Data Science: Similarity Least Squares (SLSTM) + physics + Statistical Design of Experiment (DOE)

            A New Paradigm for Analysis and Management of Complexity

horizontal rule

Statistical inference by object similarity

Consider theoretically a set of variables {V} that completely determines a target phenomenon's state.
Partition {V} into sets {x}, {y}, and {n}. SLSTM estimates the expected value of {y} from its distribution as induced by {n} and conditioned (constrained) by given values of {x}.
In this sense, {x} are predictors of the response(s) {y} with error that is induced by the set {n}, treated as a collection of stochastic noise factors.
The usual "law of large numbers" does not include combinatorial noise effects, while SLSTMdoes includes these important factors. Hence, one has a "generalized law of large numbers". 

Note inherent in this brief summary are many assumptions involving topology, function, probability, and number theory as well as assertions concerning physics concepts such as determinism, complementary uncertainty, and state space.


horizontal rule

Copyright of James M Minor, July 4, 2004.
For problems or questions regarding this web contact email
Last updated: June 20, 2013.