lxy009 / weibullpp Goto Github PK

• reliability for given time
• prob of failure for given time
• mean life
• failure rate

need way to get p-value from D statistics from modified KS-test.
might need to find tables and interpolate

make it easier for those to add on

align types to

• probability plot aka Weibull paper
• reliability vs time aka 1-CDF, normal scale
• unreliability vs time aka CDF, normal scale
• pdf
• failure rate vs time aka hazard function

is this a stacked bar graph? i.e., should suspension bars be stacked on top of failure bars?

create 'recurrentdata' class
list of 3 items, 1 for serial id, 1 for time, 1 for censor status
only right censoring
attribute for time units

• prob of failure over time
• reliabilty vs time
• pdf plot
• failure/suspension histogram

probably won't work if all data is suspension as well

• update readme markdown
• get object help file started
• description should have some link to more thorough documentation

similar to Surv class in survival package
list with 2 items: one for lifetimes, one for censor status
for now, only right censoring
attribute for time units

the sorting doesn't seem to be done prior to rank calculation

function take life data and outputs list of class 'fitted_life_data'
other input is distribution choice
MLE method only
2 dist available only: exponential and weibull
no confidence bounds
optimization use optim default
for initial guess use 1 for shape and mean for scale
for exponential, no fitting is necessary

use log transformations to tame

solve for MR

where N is the sample size and i the order number.

• time until reliability reaches certain value
• time until prob of failure reaches a certain value

don't know how exactly Weibull++ gets the results seen here
http://www.weibull.com/hotwire/issue71/relbasics71.htm

depends on both how they calculate bins + what their probability statement actually means

Since the MLE of the scale is the sample mean, can use this distribution to derive standard errors instead of approximations

type 1 refresh model,
weibull only
mle only
input is recurrent data type

investigate