LPPL model for bubble burst forcasting

Try this:

lines(exp(fitted(lm(log(pr) ~ ti))), col = "purple")


On Fri, Jul 17, 2009 at 4:11 AM, Wind<windspeedo99 at gmail.com> wrote:

Sorry that I forgot including the first line of the code:
library(quantmod)

So the replicable codes as following:

library(quantmod)
hsi<-read.csv("http://32xiang.appspot.com/static/hsi-1970.csv",header=TRUE,stringsAsFactors=FALSE)
pr<-hsi$close

plot(pr,type="l",log="y")
grid()

ti<-index(pr)
ti2<-index(pr)^2

lines(lm(pr~ti+ti2)$fit,col="red")
lines(lm(pr~ti)$fit,col="blue")
lines(lm(pr~ti2)$fit,col="pink")




On Fri, Jul 17, 2009 at 3:12 PM, Wind<windspeedo99 at gmail.com> wrote:

There is a chart of Heng Seng index ?in page 24 of Prof. Sornette's paper:
## Financial Bubbles, Real Estate bubbles, Derivative Bubbles, and the
Financial and Economic Crisis
## http://arxiv.org/abs/0905.0220
The picture has also been attached as .hong kong.jpg

The y axis of the chart is log-axis. ?And there is a straight line in
the chart. ?"This is indeed the long-term behavior of this market, as
shown by the best linear fit represented by the solid straight line,
corresponding to an average constant growth rate of 13.8% per year."
The following codes could not plot the same straight line. ?I wonder
how could plot the straight best fit line ?in the log plot.


## Financial Bubbles, Real Estate bubbles, Derivative Bubbles, and the
Financial and Economic Crisis
## http://arxiv.org/abs/0905.0220
## chart in Page 24

hsi<-read.csv("http://32xiang.appspot.com/static/hsi-1970.csv",header=TRUE,stringsAsFactors=FALSE)
pr<-hsi$close

plot(pr,type="l",log="y")
grid()

ti<-index(pr)
ti2<-index(pr)^2

lines(lm(pr~ti+ti2)$fit,col="red")
lines(lm(pr~ti)$fit,col="blue")
lines(lm(pr~ti2)$fit,col="pink")

Thanks in advance.

wind