MFT (version 2.0)

MFT.peaks: MFT.peaks

Description

The multiple filter test for peak detection in time series or sequences of random variables

Usage

MFT.peaks(x, autoset.H = TRUE, S = NULL, E = NULL, H = NULL,
  alpha = 0.05, method = "asymptotic", sim = 10000, Q = NA,
  blocksize = NA, two.sided = FALSE, perform.CPD = TRUE,
  print.output = TRUE)

Arguments

x

numeric vector, input sequence of random variables

autoset.H

logical, automatic choice of window size H

S

numeric, start of time interval, default: NULL, if NULL then 1 is chosen

E

numeric, end of time interval, default: NULL, if NULL then length(X) is chosen, needs E > S

H

vector, window set H, the smallest element must >= 3 be and the largest =< (T/2). H is automatically set if autoset.H = TRUE

alpha

numeric, in (0,1), significance level

method

either "asymptotic", "bootstrap" or "fixed", defines how threshold Q is derived, default: "asymptotic", If "asymptotic": Q is derived by simulation of limit process L (Gaussian process); possible set number of simulations (sim), If "bootstrap": Q is derived by (Block)-Bootstrapping; possibly set number of simulations (sim) and blocksize (blocksize), If "fixed": Q may be set manually (Q)

sim

integer, > 0, No of simulations of limit process (for approximation of Q), default = 10000

Q

numeric, rejection threshold, default: Q is simulated according to sim and alpha

blocksize

NA or integer >= 1, if method == 'bootstrap', blocksize determines the size of blocks (number of life times) for bootstrapping

two.sided

logical, if TRUE a two sided test is performed and also negative peaks are considered in peak detection

perform.CPD

logical, if TRUE change point detection algorithm is performed

print.output

logical, if TRUE results are printed to the console

Value

invisible

M

test statistic

Q

rejection threshold

method

how threshold Q was derived, see 'Arguments' for detailed description

sim

number of simulations of the limit process (approximation of Q)

blocksize

size of blocks (number of life times) for bootstrapping (approximation of Q)

CP

set of change points estmated by the multiple filter algorithm, increasingly ordered in time

S

start of time interval

E

end of time interval

Tt

length of time interval

H

window set

alpha

significance level

two.sided

logigal, if TRUE also negative peaks are considered

perform.CPD

logical, if TRUE change point detection algorithm was performed

tech.var

list of technical variables with processes x and D_ht

type

type of MFT which was performed: "peaks"

References

Michael Messer, Hendrik Backhaus, Albrecht Stroh and Gaby Schneider (2019+). Peak detection in times series

See Also

MFT.filterdata, plot.MFT, summary.MFT, MFT.mean, MFT.rate, MFT.variance

Examples

# NOT RUN {
# Normal distributed sequence with 2 peaks
set.seed(12)
m <- c(rep(0,30),seq(0,3,length.out = 100),seq(3,0,length.out = 80),rep(0,10),
       seq(0,6,length.out = 50),seq(6,0,length.out = 50),rep(0,30))
x <- rnorm(length(m),m)
mft <- MFT.peaks(x)
plot(mft)
# Set additional parameters (window set)
mft <- MFT.peaks(x,autoset.H = FALSE, H =c(30,60,90))
plot(mft)

# }