PLS.nlopt: PLS estimation by the iterative algorithm via NLOPTR

In zhan-gao/classo: Classifier-Lasso for Panel Data Grouping

Description

PLS estimation by the iterative algorithm via NLOPTR

Usage

PLS.nlopt(N, TT, y, X, K, lambda, beta0 = NULL, R = 500, tol = 1e-04,
  post_est = TRUE, bias_corr = FALSE, algo = "NLOPT_LN_NELDERMEAD")

Arguments

N	The dimension of cross-sectional units in the panel.
TT	The time series dimension in the panel.
y	Dependent variable. (TN * 1). T is the fast index.
X	Independent variable. (TN * P). T is the fast index. P is the number of regressors.
K	The number of groups.
lambda	The tuning parameter.
beta0	N*p matrix. The initial estimator for each i=1,...,N.
R	Maximum number of iteration.
tol	Tolerance level in the convergence criterion.
post_est	A boolean: do post-lasso estimation or not.
bias_corr	A boolean: do bias correction in the post-lasso estimation or not.
algo	choose algorithm for nloptr

Value

A list contains estimated coeffcients and group struncture

b.est	N * p matrix containing estimated slope for each individual
a.out	K * p matrix containing estimated slope for each group
group.est	group_id for each individual
converge	A boolean indicating whether convergence criteria is met

zhan-gao/classo documentation built on April 24, 2020, 11:58 p.m.