lppl_estimate: Estimation of the LPPL model using a nonlinear optimization

In deanfantazzini/bubble: Tests for financial bubbles and explosive behavior

Description

This function estimates a LPPL model using a nonlinear optimization

Usage

lppl_estimate(x)

Arguments

x	is a T x 1 numeric data vector

Details

This function estimates the LPPL model by Johansen, Ledoit, and Sornette (2000) using the original (two-step) nonlinear optimization, see section 4.1 in Geraskin and Fantazzini (2013) for a compact review. The returned parameter vector contained the following parameters:

par_est[1] = beta

par_est[2] = omega

par_est[3] = phi

par_est[4] = tc (i.e. the critical time)

par_est[5] = A

par_est[6] = B

par_est[7] = C

We remark that this estimation method is not recommended, due to the frequent presence of many local minima of the cost function where the minimization algorithm can get trapped. It was included in this package for historical (and teaching) reasons.

Value

par_est is a 7 x 1 vector of estimated parameters

Examples

 ## Not run: 
 tparm=c(0.353689, 9.154368, 2.074608, 7.166421,-0.434324, 0.035405, 0.000071, 530)
 aa=lppl_simulate(500,tparm)

 bb=lppl_estimate(aa); bb;
 
## End(Not run)

deanfantazzini/bubble documentation built on Oct. 22, 2020, 2:43 p.m.