24712

po - intercept (the place in which trend function crosses OY axis (vertical axis), or this is the average level of Y in t=0 (time unit number byO)

Pi - slope - the average change of Y in 1 time unit (in our case the average yearly change)

We estaimate trend parameter with OLS (Ordinary Least Sguares) method.

Linear trend sample (our time series) in the population: y_f = b₀ + b_lt + s_t y_t - observed values of variable y t - time variable

b₀ - estimator of p₀: sample intercept

bx - estimator of p_x: sample slope (sometimes called trend coeficient)

e_t - residuals (reszty) or error term, observed values of random component ę (xi) £-(y hat) - theoretical value of a trend function (value calculated from the model)

There are two generał types of time series:

o Additive - Y = trend + seasonality + error o Multiplicative - Y = trend*seasonality*error

Generally we have 2 seasonai models:

1. Additive - Y=trend+seasonallty+random ness - lt's wehen diviations from the trend are of the same level no matter where we r on our time series.