detect_outliers {evprof}

Detect outliers

Description

Detect outliers

Usage

detect_outliers(
  sessions,
  MinPts = NULL,
  eps = NULL,
  noise_th = 2,
  log = FALSE,
  start = getOption("evprof.start.hour")
)

Arguments

sessions

tibble, sessions data set in evprof standard format.

MinPts

MinPts parameter for DBSCAN clustering

eps

eps parameter for DBSCAN clustering

noise_th

noise threshold

log

logical, whether to transform ConnectionStartDateTime and ConnectionHours variables to natural logarithmic scale (base = exp(1)).

start

integer, start hour in the x axis of the plot.

Value

sessions tibble with extra boolean column Outlier

Examples

library(dplyr)
sessions_outliers <- california_ev_sessions %>%
  sample_frac(0.05) %>%
  detect_outliers(start = 3, noise_th = 5, eps = 2.5)