April 2018
Beginner to intermediate
500 pages
11h 26m
English
Estimation of the trend component can occur in two different modes depending on the linear/non-linear characteristic.
If the series trend is linear or linearizable in the parameters through a logarithmic transformation, then these trends can be estimated through the procedures derived from linear regression. We can hypothesize a polynomial trend that can be represented by the following equation:
τ (t) = α0 + α1 t + α2 t2 + ... + αq tq + εt
In this formula, q represents the degree of the polynomial.
Depending on the value assumed by q, the following cases can be represented:
| q | Cases |
| 0 | A constant trend is obtained |
| 1 | We obtain a linear trend |
| 2 | We obtain a parabolic trend |
On the contrary, the presence ...
Read now
Unlock full access