Time series

[kenhung123]

Is there any 'tricks' to identify/distinguish between trend, seasonal, cyclical and random variations?
I thought, where there is a pattern every year it is seasonal, if there is a general rise and drop it is a trend and cyclical is when there is a similar pattern over a set time period. It doesn't seem to work for the question below

Also, just to clarify, when we use the 3 and 5 smoothing mean, do we just replace the middle values and forget the others i.e. 2 in 3 moving mean and 3 in 5 moving mean

[tolga]

I didn't really get the question i think its not clear enough and i dont see how its a seasonal variation i believe its a cyclical.

[v3n0m]

Yes you do replace the middle value. However, I have not done this question myself but my friend is an expert at these.

Hope it helped =]

[stonecold]

i would probably go:
a-random
b-cyclical
c-increasing/secular trend

i stink at these though.

[akira88]

i would probably go:
a-random
b-cyclical
c-increasing/secular trend

i stink at these though.

I would agree with stonecold with this!
I don't understand your question about the smoothing means?

[Gloamglozer]

i would probably go:
a-random
b-cyclical
c-increasing/secular trend

i stink at these though.

I would agree with stonecold with this!
I don't understand your question about the smoothing means?

Smoothing is a technique that is used to make the pattern more distinguishable.  There are two types of smoothing - the mean or median.  Usually you'd either use the 3-point or 5-point mean/median smoothing method because it's easier.  However, the question will tell you which one to use.  For more info, check out Chapter 7 in Essentials.