*
* Graphs from section 11.3, pp 374-376
*
all 50
*
* These generate random walks with N(0,1) increments. The first option pegs the
* first value to 0. (If you don't use this or something like it, the series will
* be all NA's, since there won't be a lag to compute the first data point).
*
* Note that these won't reproduce the series shown in the book, since those are
* based upon another set of random numbers. In fact, as written, you won't be
* able to reproduce the results graphed here if you rerun the program. If you
* want to control the "random" numbers chosen (actually, they're more
* appropriately known as pseudo-random), use the SEED instruction somewhere near
* the top of the program (for instance, SEED 95039).
*
set(first=0.0) rw1 = rw1{1}+%ran(1.0)
set(first=0.0) rw2 = rw2{1}+%ran(1.0)
graph(footer="Figure 11.1 Two Realizations of a Random Walk") 2
# rw1
# rw2
*
* Drifting random walk with N(0,9) increments. Note that %ran takes as its
* argument the standard deviation, thus 3 in this case.
*
set(first=0.0) rwd = 2.0+rwd{1}+%ran(3.0)
set trend = 2.0*(t-1)
graph(footer="Figure 11.3 Realization of a Random Walk with Drift") 2
# rwd
# trend