bar_unit_runs: Construct unit runs bars

In larryleihua/fmlr: Financial Machine Learning using R

Description

Construct unit runs bars

Usage

bar_unit_runs(
  dat,
  u_0 = 2000,
  w0 = 10,
  de = 1,
  bkw_T = 5,
  bkw_Pb1 = 5,
  bkw_U = 5,
  filter = FALSE
)

Arguments

dat	dat input with at least the following column: Price, Size
u_0	average unit (volume*price) for each trade, and it is used to create the first bar
w0	the expected time window length of the first bar
de	a positive value for adjusting the expected window size, that is, de*E0T; default: 1
bkw_T	backward window length when using exponentially weighted average T
bkw_Pb1	backward window length when using exponentially weighted average P[b_t=1]
bkw_U	backward window length for exponentially weighted average volumes
filter	whether used as a filter; default FALSE. If TRUE, then only i_feabar, the ending time index of feature bars, is returned

Value

If filter==FALSE, a list of vectors for tStamp (if returned), and HLOCV of volume runs bars. Note that the remaining data after the latest ending time point detected will be formed as a bar. If filter==TRUE, i_feabar a vector of integers for the time index.

Author(s)

Larry Lei Hua

Examples

set.seed(1)
dat <- data.frame(Price = c(rep(0.5, 4), runif(50)), Size = floor(runif(54)*100))
bar_unit_runs(dat, u_0=mean(dat$Price)*mean(dat$Size))
bar_unit_runs(dat, u_0=mean(dat$Price)*mean(dat$Size), filter=TRUE)

larryleihua/fmlr documentation built on Jan. 31, 2020, 8:55 a.m.