lppl.validates: Estimation of the LPPL model using a nonlinear optimization

In mcgyver3/lppl: Log Periodic Power Low(LPPL)

Description

This function estimates a LPPL model using a nonlinear optimization

Usage

lppl.validates(
  tradedate,
  price,
  mb = c(1:3),
  MovingWindows = c(5:10),
  Right_days = 60,
  Right_scale = 0.8,
  Left_days = 262,
  h = 1.5,
  gaRunCnt = 100,
  gaGenerations = 500,
  gaPopulation = 200,
  gaElitism = 0.5,
  gaMutation = 0.1,
  gaCrossover = 0.8
)

Arguments

tradedate	Use a "Date" type vector as the transaction date.
price	Closing price numeric data vector
mb	The LPPL model is estimated using only the data from the critical point to the specified number of days (mb x 20 days). The data mb has a vector type, and the estimated data for each mb specified is returned. The default value is 1 to 3 months.
MovingWindows	This is the size of the window used to detect the peak point in the PeakDetection operation, and input as vector data. The default value is 5 to 10 days.
Right_days	It implements that the peak should drop by 25% during the right_days (60 days) period, and the default value is 60.
Right_scale	set the percentage of drop price from peak price. (Default : 0.8)
Left_days	The peak of critical time is higher than any peak prior to 262 days (Default : 262)
h	Parameter for peak detection. Parameter is a positive coefficient, meaning that the higher the value of h, the tighter the peak is detected. h is generally set at 1 <= h <= 3. According to Palshikar (2009), the default value was set to 1.5.
gaRunCnt	GA Execution Count (Default : 100)
gaGenerations	the maximum number of iterations to run before the GA search is halted. (Default : 500)
gaPopulation	the population size. (Default : 200)
gaElitism	the number of best fitness individuals to survive at each generation.(Default : 0.5)
gaMutation	the probability of mutation in a parent chromosome.Usually mutation occurs with a small probability, and by default is set to 0.1.
gaCrossover	the probability of crossover between pairs of chromosomes. Typically this is a large value and by default is set to 0.8.

Value

LogPrice : Log price of input time series data

id_SRTPT : Starting index of the data used for LPPL estimation as the lowest point after the previous crash time

id_crash : Index of real crash time

id_END : Last index of data used for crash time estimation

DT_SRTPT : Date corresponding to id_SRTPT

DT_crash : Date of real crash time

DT_END : Last date of data used for crash time estimation

Best_order : crash time estimation result (RMSE-based sorting data)

Examples

library(lppl)

tradedate <- lppl::NASDAQ$Date
price <- lppl::NASDAQ$Close

Result <- lppl.validates(tradedate =tradedate
                       ,price = price
                       ,mb=c(1)
                       ,MovingWindows=c(5)
                       ,gaRunCnt=100
                       ,gaGenerations=200
                       ,gaPopulation=50)

mcgyver3/lppl documentation built on Jan. 1, 2021, 9:23 a.m.