only_mod_plot.R

# Script to generate deterministic LPPLS model. 
# You can also add by yourself some kind of noise in the time series to simulate syntetic random LPPLS price paths

library(cmaes)
library(quantmod)
library(zoo)
library(lubridate)
library(dplyr)
library(parallel)
library(doParallel)
library(progress)
library(ggplot2)
library(nloptr)
library(optimx)
library(latticeExtra)


lppl_est <- function(data, tc, m, w, a, b, c1, c2){
  
  dif_time = abs(tc-data$t)
  
  est = a + dif_time^m*(b + ((c1 * cos(w * log(dif_time))) 
                             + (c2 * sin(w * log(dif_time)))))
  
  return(est)
}

data <- data.frame(t=seq(1,20,0.01))


#######

tc=19.4

m=0.8
w=7
a=15
b=-6
c1=0.6
c2=0.5
-b*m-( abs(sqrt(c1^2+c2^2))*sqrt(m^2+w^2) )
data$fit1 <- lppl_est(data,tc,m,w,a,b,c1,c2)



m=0.6
w=7
a=15
b=-6
c1=0.6
c2=0.5
data$fit2 <- lppl_est(data,tc,m,w,a,b,c1,c2)

#Fit1
ggplot(data, aes(x=t, y=fit1)) + 
  geom_line(size=0.7) +
  geom_vline(xintercept=19.4, linetype="dotted", color="red", size=1) +
  xlab("Time") + ylab("log-price") + theme_bw() + 
  theme(axis.text.y = element_blank())
#Fit2
ggplot(data, aes(x=t, y=fit2)) + theme_bw() +
  geom_line(size=0.7) +
  geom_vline(xintercept=19.4, linetype="dotted", color="red", size=1) +
  xlab("Time") + ylab("log-price") + 
  theme(axis.text.y = element_blank())