correlate

Format: correlate( rand, index )

Arguments: (real) rand Call to one of the random number generators

(int) index The random input's location in the correlation matrix

Returns: (real) Random number with the distribution of rand but correlated with other random inputs

Description: To use the correlate primitive, your model must contain a Data Table sheet named Cormat (character case is important). This table contains a row and column for each random input that will be correlated with other inputs. The intersection of each row and column must contain the correlation coefficient for the corresponding pair of inputs. For example, in the table below the correlation between Bonds and Large Cap is 0.60.

The argument rand is any statement you would use to generate a random number without correlation and index indicates the row/column that contains the specific random input's correlation data.

Examples: Definitions for uncorrelated random inputs could look like this:

Cash:=lrand(1.01,0.01)
Bonds:=lrand(1.02,0.04)
Large Cap:=lrand(1.05,0.08)
Small Cap:=lrand(1.06,0.11)
International:=lrand(1.08,0.17)

With correlation, the same nodes look like this:

Cash=correlate(lrand(1.01,0.01),1)
Bonds:=correlate(lrand(1.02,0.04),2)
Large Cap:=correlate(lrand(1.05,0.08),3)
Small Cap:=correlate(lrand(1.06,0.11),4)
International:=correlate(lrand(1.08,0.17),5)

See Also: rand, drand, trand, nrand, lrand, brand, prand, grand, betarand, crand, setrand