•In 2D, divide the
plane into squares. You can think of
similar tests in higher
dimensions also.
•
• Define: N = number of
trials
•k = number of possible
outcomes of the chance experiment
•f(ZETA_i) = number of
occurrences of ZETA_i in N trials
• E(ZETA_i) = The
expected number of occurrences of ZETA_i
in N trials. E(ZETA_i) = N*Pr(ZETA_i).
• i=k [ f(ZETA_i) - E(ZETA_i) ]**2
•CHISQ = SUM -------------------------------------
• i=1 E(ZETA_i)