weibulltools (version 1.0.1)

mcs_delay_report: Adjustment of Operating Times by Delays in Report using a Monte Carlo Approach

Description

The delay in report describes the time between the occurence of a damage and the registration in the warranty database. For a given date where the analysis is made there could be units which had a failure but are not registered in the database and therefore treated as censored units. To overcome this problem this function uses a Monte Carlo approach for simulating the operating times of (multiple) right censored observations, taking account of reporting delays. The simulation is based on the distribution of operating times that were calculated from complete data, i.e. failed items (see dist_delay_report).

Usage

mcs_delay_report(date_repair, date_report, x, event,
  distribution = "lognormal", details = FALSE, seed = NULL)

Arguments

date_repair

a vector of class "character" or "Date", in the format "yyyy-mm-dd", indicating the date of repair of a failed unit. If no date is available use NA.

date_report

a vector of class "character" or "Date", in the format "yyyy-mm-dd", indicating the date of report of a failed unit. If no date is available use NA.

x

a numeric vector of operating times.

event

a vector of binary data (0 or 1) indicating whether unit i is a right censored observation (= 0) or a failure (= 1).

distribution

supposed distribution of the random variable. The default value is "lognormal". So far no other distribution is implemented.

details

a logical variable, where the default value is FALSE. If FALSE the output consists of a vector with corrected operating times for the censored units and the input operating times for the failed units. If TRUE the output consists of a detailed list, i.e the same vector as described before, simulated random numbers, estimated distribution parameters and a seed for reproducibility.

seed

if seed = NULL a random seed is used. Otherwise the user can specify an integer for the seed.

Value

A numeric vector of corrected operating times for the censored units and the input operating times for the failed units if details = FALSE. If details = TRUE the output is a list which consists of the following elements:

• time : Numeric vector of corrected operating times for the censored observations and input operating times for failed units.

• x_sim : Simulated random numbers of specified distribution with estimated parameters. The length of x_sim is equal to the number of censored observations.

• coefficients : Estimated coefficients of supposed distribution.

• int_seed : Integer seed number for reproducibility.

Examples

# NOT RUN {
date_of_repair <- c(NA, "2014-09-15", "2015-07-04", "2015-04-10", NA,
                   NA, "2015-04-24", NA, "2015-04-25", "2015-04-24",
                    "2015-06-12", NA, "2015-05-04", NA, NA,
                    "2015-05-22", NA, "2015-09-17", NA, "2015-08-15",
                    "2015-11-26", NA, NA)

date_of_report <- c(NA, "2014-10-09", "2015-08-28", "2015-04-15", NA,
                    NA, "2015-05-16", NA, "2015-05-28", "2015-05-15",
                    "2015-07-11", NA, "2015-08-14", NA, NA,
                    "2015-06-05", NA, "2015-10-17", NA, "2015-08-21",
                    "2015-12-02", NA, NA)

op_time <- rep(1000, length(date_of_repair))
state <- c(0, 1, 1, 1, 0, 0, 1, 0, 1, 1, 1, 0, 1, 0, 0, 1, 0, 1, 0, 1, 1, 0, 0)

# Example 1 - Simplified vector output:
x_corrected <- mcs_delay_report(date_repair = date_of_repair,
                                date_report = date_of_report,
                                x = op_time,
                                event = state,
                                distribution = "lognormal",
                                seed = NULL,
                                details = FALSE)

# Example 2 - Detailed list output:
list_detail <- mcs_delay_report(date_repair = date_of_repair,
                                date_report = date_of_report,
                                x = op_time,
                                event = state,
                                distribution = "lognormal",
                                seed = NULL,
                                details = TRUE)
# }