JMP 12 Online Documentation (English)
Discovering JMP
Using JMP
Basic Analysis
Essential Graphing
Profilers
Design of Experiments Guide
Fitting Linear Models
Specialized Models
Multivariate Methods
Quality and Process Methods
Reliability and Survival Methods
Consumer Research
Scripting Guide
JSL Syntax Reference
JMP iPad Help
JMP Interactive HTML
Capabilities Index
JMP 13.2 Online Documentation
Reliability and Survival Methods
•
Destructive Degradation
• The Destructive Degradation Report
Previous
•
Next
The Destructive Degradation Report
The Destructive Degradation report contains an outline for each model that you fit. When you fit a model, the Model List is updated with a row for that model.
Note:
All models are fit using the maximum likelihood method.
Model List and Model Outlines
Model List
The first three columns in the Model List reflect the choices that you made in the plot options. The -Loglikelihood, AICc, and BIC statistics are information-based measures that can be used for model comparisons. For descriptions of these measures, see
Fitting Linear Models
.
The three information-based measures in the Model List are comparable across models as long as the models being compared have the same Number of Actual Observations. If this is not the case, exercise caution because different models might use different subsets. The Number of Actual Observations might be reduced due to the choice of distribution or the choice of transformation. Choosing a log-location-scale distribution excludes all non-positive Y values. Also, the Log10 and Sqrt transformations exclude all non-positive values.
Each row of the Model List table has a red triangle menu with the following options:
Scroll To
Scrolls the report window to the corresponding model outline.
Remove
Removes the model from the Model List and removes the corresponding model outline from the report.
Model Outlines
The outline for each model contains a red triangle menu with the following option:
Remove
Removes the model outline from the report and removes the model from the Model List.
The outline for each model contains the following reports:
Formula Picture
Shows the equation for the location parameter.
Estimate
Shows parameter estimates and their standard deviations.
Distribution, Quantile, and Inverse Prediction
For a description of these profilers, see
Profilers
.
Reports within a Model Outline
Profilers
Three profilers appear in the model report.
Degradation Profiler
The Degradation Profiler shows a profiler view of the Degradation Data Analysis plot for the given model. The response is the degradation response Y. The profiler includes a plot against the Time variable and a plot against the optional explanatory variable X (if you have specified one). The plot against Time shows the median of the fitted distribution of Y as a solid curve. The dashed curves show the 0.025 and 0.975 percentiles of the fitted distribution of Y.
Crossing Probability Profiler
Use the Crossing Probability Profiler to determine the probability that the degradation measurement falls below a given threshold at some point in time.
The profiler plots the estimated cumulative distribution function of the response Y as a function of Time, the optional X variable, and Y. The plots for Time and Y show Wald confidence intervals.
Crossing Time Profiler
Use the Crossing Time Profiler to determine the time at which a specified proportion of measurements falls below a given threshold value.
The profiler plots the estimated Time as a function of the optional X variable, quantile values for Y (Probability), and Y. The plots for Probability and Y show Wald confidence intervals.