topic Re: Confidence interval for degradation data analysis in Discussions

Confidence interval for degradation data analysis

billi — Fri, 09 Jun 2023 00:52:44 GMT

Hello,

In degradation analysis, is there any option to save the crossing time with 95% confidence interval. Confidence interval is shown in the plot but I do not see it mentioned anywhere in the report. Is there any option to get those numbers. Below is the plot from degradation data analysis showing the crossing time with 95% CI.

Thank you.

Re: Confidence interval for degradation data analysis

peng_liu — Fri, 22 Jul 2022 02:54:00 GMT

Try to select "Confidence Interval" in Inverse Prediction, then select "Save Crossing Time". See whether that is what you need.

Re: Confidence interval for degradation data analysis

billi — Fri, 22 Jul 2022 12:29:32 GMT

@peng_liu Thank you for your reply. I tried this solution but it just saves the predicted crossing time not the confidence interval.

Re: Confidence interval for degradation data analysis

peng_liu — Fri, 22 Jul 2022 12:37:33 GMT

That's odd. The following data table is what you should see. The two highlighted columns are where confidence interval bounds cross the threshold. If you don't see them. There is something I don't understand. You may either post more details here, or you may contact support@jmp.com if it is inconvenient to discuss here.

Re: Confidence interval for degradation data analysis

billi — Fri, 22 Jul 2022 13:55:50 GMT

@peng_liu I attached the data table and below is snapshot of what I get when I go to Inverse Prediction > Save Crossing Time.

Re: Confidence interval for degradation data analysis

peng_liu — Fri, 22 Jul 2022 14:08:49 GMT

I understand now. You are using "Stability Analysis" part of the platform. Stability analysis is a special application of the general degradation analysis. Its behavior is tailored to suit the special need of stability analysis. The "Prediction" column that you see is already the crossing time of the confidence interval line with the threshold. And it is lower one-sided. The actual crossing times of the degradation lines with the threshold are not reported, because they are not the interest of stability analysis.