The irregular variations are very much erratic in nature. They remain so mixed up with the cyclical variations that are very difficult to separate them from the cyclical variation in a meaningful manner.

However, the following methods may be suggested for identifying and isolating them in a time series, some how or other.

**1. Additive Model:**

Under this model, the irregular variations are identified by subtracting the sum of the other three components of a time series viz : trend, seasonal and cyclical, from its observed value. Symbolically this given by:

I = Y – (T + S + C)

Where I = irregular variation:

Y = observed value i.e. Y_{c}

T = trend value ,i.e. Y_{c}

S = Seasonal value and

C = Cyclical variation.

**2. Multiplicative Method.**

Under this model, the irregular variations are measured by dividing the observed values in a time series by the product of its other three components viz : T, S and C. Symbolically, this is given by

Under this method, the irregular variations can, also, be identified in any of the following two forms:

- As ratio of the index of cycle & irregular to the index of the cyclical variation and multiplied by 100 i.e. I =
- As a measure of cyclical normalcy (percentage deviation) by deducting 100 from the index of irregular variation.

Thus, Cyclical normalcy or

Percentage deviation = 1 – 100