Statistical analysis of simulation results

Simulation of random vectors

Suppose you want to generate a sample N values ​​of the random vector with predetermined characteristics:

expectation and covariance matrix .

Express the vector with independent, identically distributed components with independent, identically distributed components

For example, for r = 2, we write:

(12)

viously, Y ₁ and Y ₂ have the desired mat ematicheskie expectations. I f = 0, then they are functionally (fully) dependent, if = 0, then they are not dependent

We choose o that Y ₁ and Y ₂ and had the desired standard deviation and correlation coefficient.

Once we have taken care of properly distributed randomness and random external influences and / or parameters of the model, our model bude t "live" almost real life and fix a variety of e f values ​​from run to run - it's like watching a real system repeatedly. This means that these data can be recorded and shall apply the whole arsenal of mathematical statistics to estimate the nature ISTIC quantities of interest, their dependence on it (correlation), and so on. D.

Дата добавления: 2014-10-23