24724

2

Function fitting(eye view)

Test the significance of time variables in your model Statistics -> multiple regression We're trying the following model:

Y = b₀+bi*t+b₂*t²+b₃*t³

One of the time variables is statistically significant - all p-values for them are bigger than significance level (0.05). It means that the order of my polynomial is too high. On the other hand, according to backward stepwise-regression rule we should eliminate that variable, which has the highest p-value. Both criteria say that t^ł should be eliminated from the model.

Now wetry:

Y = bo+bi^łt+b₂*t²

All variables are statistically significant. It means we have found the proper model for our trend.