thresholdTest {nonlinearTseries}

Threshold nonlinearity test

Description

Computes the likelihood ratio test for threshold nonlinearity with H0 being an AR process and H1 a TAR model.

Usage

thresholdTest(
  time.series,
  p,
  d = 1,
  lower.percent = 0.25,
  upper.percent = 0.75
)

Arguments

time.series

The original time.series.

p

The order of the AR process.

d

Delay used for the threshold value in the TAR process.

lower.percent

The threshold value is searched over an interval defined by lower.percent and upper.percent of the time series values. This defines the lower percent.

upper.percent

The threshold value is searched over an interval defined by lower.percent and upper.percent of the time series values. This defines the upper percent.

Value

A list containing

• p.value: p-value of the test

• test.statistic: Likelihood ratio test statistic.

• percentiles: Since the search for the threshold parameter may occur in a narrower interval than the one specified by the user, the effective lower and upper percents are returned here.

Note

Adapted from the tlrt function of the TSA package.

Author(s)

Kung-Sik Chan

References

Chan, K.S. (1990). Percentage points of likelihood ratio tests for threshold autoregression. Journal of Royal Statistical Society, B 53, 3, 691-696.

See Also

nonlinearityTest, keenanTest, tsayTest, mcleodLiTest