Failures prediction

[pvkraju]

Hi
This is PVK Raju and I would like to know the failure prediction
We have field failure data in our organazitation on that one of the model having more failure for that we have data 5 months data.
Here is my question is how to predit the next 6 or 12 months failure
forecast.

Data as follows
Month Sales Failure
Jan 504 1
Feb 2244 44
Mar 1885 39
Apr 2204 56
May 1867 6

The above data collected in the month of May and failure data is wrt Mfg.month

Please help me how to predict the failure next 6 to 12 months

Thanks
Regards,
P V K Raju

[Pantelis]

The quick answer is to create a model that describes your units (such as a Weibull model) and then use that model to predict.

Along those lines then you will find the following links useful:

http://www.weibull.com/LifeDataWeb/warranty_analysis.htm (http://www.weibull.com/LifeDataWeb/warranty_analysis.htm)
http://www.weibull.com/LifeDataWeb/estimation_of_the_weibull_parameter.htm (http://www.weibull.com/LifeDataWeb/estimation_of_the_weibull_parameter.htm)
http://www.weibull.com/lifedatawebcontents.htm (http://www.weibull.com/lifedatawebcontents.htm)

[pvkraju]

Hi
Thank you for immediate response
I have gonthrough the warrenty analysis. on that i have one calrification
what is beta and Eta and how to calculate.
could you please help me

Thanks & Regards,
P V K Raju

[brucelee]

you need to know the service duration for each return.

I guess the return quantities from a particular month corresponding to the shipment in the previous monthes. therefore, you had to be able to sort them out and find out the exact/approximate shipment date so that you can exactract the service duration for each return.

with this info, you can enter the weibull model for the parameters estimation ( start with two parameters estimation to see if you get a reasonable fit ).

jason