auto.fit {SIMle}

Automated estimation of nonlinear time series regression

Description

This function estimates nonlinear time series regression by sieve methods with chosen bases.

Usage

auto.fit(
  ts,
  c,
  d,
  b_time,
  b_timese,
  mp_type,
  type,
  ops,
  per = 0,
  k = 0,
  fix_num = 0,
  r = 1,
  s = 1,
  upper = 10
)

Arguments

ts

ts is the data set which is a time series data typically

c

the maximum value of number of basis for time input

d

the maximum value of number of basis for variate input

b_time

type of basis for time input

b_timese

type of basis for variate input

mp_type

select type of mapping function, "algeb" indicates algebraic mapping on the real line. "logari" represents logarithmic mapping on the real line

type

select type of estimation."nfix" refers to no fix estimation. "fixt" indicates fix time t estimation. "fixx" represents fix variate estimation

ops

Criteria for choosing the number of bases are provided by the package, offering four options: "AIC," "BIC," "CV," and "Kfold," each corresponding to a specific Criteria

per

the percentage for test set used in cross validation option "CV"

k

the number of fold used in k-fold cross validation "Kfold"

fix_num

fix_num indicates the use of fixed-value nonlinear time series regression. The default value is 0, which is employed for non-fixed estimation. If "fixt" is chosen, it represents a fixed time value. Otherwise, if not selected, it pertains to a fixed variate value

r

indicates number of variate

s

s is a positive scaling factor, the default is 1

upper

upper The upper bound for the variate basis domain. The default value is 10. When "algeb" or "logari" is chosen, the domain is automatically set from -upper to upper

Value

If "nfix" is selected, the function returns a list where each element is a matrix representing the estimation function in two dimensions. Otherwise, if "nfix" is not selected, the function returns a list where each element is a vector representing the estimation function.