Duke University
\[\begin{equation} \begin{aligned} \dot{x} &= \sigma(y-x) \\ \dot{y} &= \rho x - y - xz \\ \dot{z} &= -\beta z + xy \end{aligned} \end{equation}\]
#| fig-height: 8
#| fig-align: center
#| fig-cap: A random walk with noise
library(ggplot2)
theme_set(theme_classic(base_size = 25))
library(latex2exp)
library(patchwork)
set.seed(42)
Y0 <- 10
wt <- rnorm(100, sd = 1)
vt <- rnorm(100, sd = 3)
dat <- data.frame(
t = 1:100,
Y = Y0 + cumsum(wt) + vt,
vt = vt,
wt = wt
)
p1 <- dat |>
ggplot(aes(t,Y)) +
geom_line() +
geom_line(aes(y=Y0+cumsum(wt)), lty=1, color='darkgreen') +
labs(y=TeX('$Y_t$')) +
theme(axis.title.x.bottom = element_blank())
p2 <- dat |>
ggplot(aes(t,vt)) +
geom_point(color='red', size=.5) +
geom_linerange(aes(ymin=0,ymax=vt), color='red') +
geom_hline(yintercept = 0, alpha=.5) +
labs(y=TeX('$v_t$')) +
theme(axis.title.x.bottom = element_blank())
p3 <- dat |>
ggplot(aes(t,wt)) +
geom_segment(aes(x=t,xend=t,y=0,yend=wt), color='darkgreen',
arrow = arrow(length = unit(0.2, "cm"))) +
geom_hline(yintercept = 0, alpha=.5) +
labs(y=TeX('$w_t$'))
p1/p2/p3
Here is a reference paper (Gardiner et al. 2015), and another (Liedekerke et al. 2010)