Profiler

Explore Cross Sections of Responses across Each Factor

The Prediction Profiler gives you a wealth of information about your model. Use the Prediction Profiler to do the following:

• See how your prediction model changes as you change settings of individual factors.

• Set desirability goals for your response or responses, and find optimal settings for your factors.

• Gauge your model’s sensitivity to changes in the factors, where sensitivity is based on your predictive model.

• Assess the importance of your factors relative to model predictions, in a way that is independent of the model.

• Simulate your response distribution based on specified distributions for both factors and responses, and control various aspects of the appearance of the profiler.

Figure 3.1 Prediction Profiler for Four Responses with Simulator and Importance Coloring 

Contents

Example of the Prediction Profiler

Launch the Prediction Profiler Platform

Prediction Profiler Options

Maximization Options Window

Desirability Profiling and Optimization

Construction of Desirability Functions

How to Use the Desirability Function

The Desirability Profile

Customized Desirability Functions

Assess Variable Importance

Bagging

Additional Examples of the Prediction Profiler

Example of Desirability Profiling for Multiple Responses

Example of a Noise Factor in the Prediction Profiler

Example of Variable Importance for One Response

Example of Variable Importance for Multiple Responses

Example of Bagging to Improve Prediction

Example of Bagging to Indicate the Accuracy of Predictions

Example of Confidence Intervals Using Bagging

Example of Confidence Intervals Using StdErr Pred Formula Column

Statistical Details for the Prediction Profiler

Assess Variable Importance

Propagation of Error Bars

Extrapolation Control Metrics

Want more information? Have questions? Get answers in the JMP User Community (community.jmp.com).