Naive Bayes
- From an open JMP® table, select Analyze > Predictive Modeling > Naive Bayes.
- Select a nominal or ordinal response variable from Select Columns and click Y, Response.
- Select candidate predictor variables and click X, Factor.
- If desired, enter the Validation Portion or select a validation column and click Validation as was used in this illustration.
- Click OK. JMP displays:
- Total misclassification counts and rates.
- Confusion Matrix detailing the classification performance.
- ROC curves and AUC values (only results for the training data shown here).
Note: Click on the top red triangle and select Profiler to display an interactive tool that shows the predicted probability of each class as a function of the levels of the predictor variables. Other options, such as Assess Variable Importance can be accessed from the red triangle next to Prediction Profiler. The profiler for the three predictor variables HDL, BMI, and LTG shown on bottom right.
Resuts of the Naive Bayes classifier to predict the level (Low/High) from the 442 diabetes patients:
• There are 309 observations in the Training Data. Of these, 66 (21%) where misclassified. 41/(185+41) = 18% of the Low observations were misclassified as High. 25/(25+58) = 30% of the High observations were misclassified as Low.
• There are 133 observations in the Validation Data. Of these, 32 (24%) were misclassified. 25/(70+25) = 26% of the Low
observations were misclassified as High. 7/(7+31) = 18% of the High observations were misclassified as Low.
Notes: Additional options, such as Lift Curves, Save Predicteds, Save Prediction Formula, Save Probability Formula, as well as Publish Probability Formulas are accessible from the top red triangle.
Visit Predictive and Specialized Models > Naive Bayes in JMP Help to learn more.