WHen to apply RRXRRY and MLE

[keithtbe]

Hi,

Can anybody help me to understand when to use RRX,RRY or MLE? It is a big different when i use RRX to compare with MLE.

Thanks
Keith

[RS Support]

This is a question that can open a rather big discussion and may have different opinions associated with it. My general recommendation is to use RRX if the data set is small and complete (contains no censored data). For larger data sets, and with data sets with random censoring MLE should be preferred. When using Free Form data in Weibull 6 use RRY.

Hope this helps. Should you have any further questions or require more details please contact our support at Support@reliaSoft.com.

[Herjit Gill]

Hello,

Recently I changed one of my plots from a RRX to a MLE and the reliability increased from 96% to 99.18% (at 730 days). Now I am not sure which figure to use?
The data set was 50000. I had one fixed suspension date.
I attempted to drill down on a particular failure mode, this gave me only 26 failures out of the 50000. Is it possible to do a Weibull on this and if so should I use MLE or RRX?

[RS Support]

MLE results tend to yield higher values for reliability for t>eta and lower values for t<eta when compared to rank regression analysis as noted in your posting. My recommendation given your data set and large amount of censored units is to use MLE.