Subscribe Bookmark RSS Feed
tom_donnelly

Staff

Joined:

Jul 7, 2014

8 December Webcast - Building Better Models - Data Mining Methods in JMP and JMP PRO

Return to JMP Federal Government User Resources Page


By working with the attached data sets, you'll learn to build better and more useful models with predictive modeling techniques such as regression, neural networks and decision trees. The focus will be on getting the most out of JMP with a few examples where JMP PRO is able to deliver additional value.  You'll learn to partition your data into training, validation (tuning) and test sets to prevent over fitting. And you'll see how to use comparison techniques - especially plots of actual vs. predicted values by validation group - to find the best predictive model. This webcast is for JMP users interested in learning how predictive modeling can help them use the data they have to make better predictions. 


Response data will include binary responses (pass/fail, accept/reject, hit/miss, etc.), continuous responses (including proportions between 0 and 1), and categorical responses (different types of cyber attacks).  Independent variables will include continuous and/or categorical types.

Related webcasts from the Mastering JMP Series include:

- Predictive Modeling by Malcolm Moore

- Advanced Modeling Using JMP Pro by Sam Gardner

- Data Mining and Predictive Modeling by Sam Gardner

- Predictive Modeling Using JMP 9 Neural Nets by Sam Gardner


Good reference books in this area include:

Fundamentals of Predictive Analytics with JMP, by Klimberg and McCullough - all examples worked with JMP

The Elements of Statistical Learning by Hastie, Tibshirani, and Friedman - authoritative, mathematically deeper

The links above will lead one to the data sets associated with the books.

Below is actual versus predicted scatterplot matrix of a binary response vs. binary prediction for six models.

7614_Scatterplot confusion matrix.jpg

Below is a Prediction Profiler for the same response for three different types of models with 6 explanatory factors.

7616_Profiler 3 models.jpg

Below is Actual vs. Predicted plot in Graph Builder by Validation Group for a Partition model.

7615_Act vs Pred by Val Group.jpg

Article Labels
Contributors