ESS210B
Prof. Jin-Yi Yu
Independent Predictors
qBased on the previous analysis, we wish to use predictors that are independent of each other
q     è r12 = 0
q     è minimum useful correlation = 0.
qThe worst predictors are r12 = 1.0
qThe desire for independent predictors is part of the motivation for Empirical Orthogonal Function (EOF) analysis. qEOF attempts to find a relatively small number of  independent quantities which convey as much of the original information as possible without redundancy.