In order to build the interval variation series, it is first necessary to select the optimal number of intervals and set the length of each of them. In this case, note that the length of the interval should be constant, as in the analysis of variational series compare frequencies from different groups. The optimal number of groups should be selected to reflect the variety of signs together, however, their natural distribution, and also to eliminate distortion together with random fluctuations of frequencies. Note that if the groups is too small, will not be visible distribution pattern, and conversely, if too many random leaps of population units will distort the number distribution.
To determine the number of groups in the variational row, use the formula Sterides:
h = 1 + 3,322 x ln(n), where
h – the number of groups in the variational row;
n is the population size.
If the resulting value will be fractional, then the value of the step value of the interval, take the nearest integer.
Then, determine the length of the interval:
i = (Hmag – Xmin)/h, where
HMO – maximum characteristic value in the aggregate;
Xmin – the minimum value of the characteristic in the aggregate.
Then, complete the boundaries of the interval. They can be specified in different ways: the upper bound of the previous interval and may repeat the lower boundary of the next (5-10, 10-15, 15-20) or not to repeat (5-10, 10,1-15, 15,1-20). For the beginning of the first interval A0 takes the following value:
A0 = GMP – i/2, where
i – length of the interval.
At the end of the j-th interval is assumed A that represents the upper boundary of the j-th interval and the beginning of the (j+1)-th interval:
Aj = A(j-1) + i.
The scale of intervals continues as long as the value of A satisfies the ratio Aj< Hmah + i/2.