eventscribe

The eventScribe Educational Program Planner system gives you access to information on sessions, special events, and the conference venue. Take a look at hotel maps to familiarize yourself with the venue, read biographies of our plenary speakers, and download handouts and resources for your sessions.

close this panel
‹‹ Go Back

Faraz Niyaghi

Department of Statistics, Oregon State University



‹‹ Go Back

Sharmodeep Bhattacharyya

Department of Statistics, Oregon State University



‹‹ Go Back

Sarah C. Emerson

Department of Statistics, Oregon State University



‹‹ Go Back

Please enter your access key

The asset you are trying to access is locked for premium users. Please enter your access key to unlock.


Email This Presentation:

From:

To:

Subject:

Body:

←Back IconGems-Print

131 – Predictive Modeling in Data Science

Variable Selection Using Intersection and Average of Random Forests

Sponsor: Section on Statistical Learning and Data Science
Keywords: Variable Selection, Random Forest, Intersection of Forests, Stability

Faraz Niyaghi

Department of Statistics, Oregon State University

Sharmodeep Bhattacharyya

Department of Statistics, Oregon State University

Sarah C. Emerson

Department of Statistics, Oregon State University

Random forest (RF) has demonstrated the ability to select important variables and model complex data. However, due to the random sampling of data points and variables within RF algorithm, rankings of the selected variables can alter among fitted models to the same data set. This can result in selecting a noise variable over a main variable. This research investigates intersection and average methods to stabilize RF's variable selection. First, multiple RF models are fitted to the data, and ranking of variables and their relative importance are evaluated for each model. Average method ranks the variables based on their mean relative importance. Intersection method iteratively selects variables that are in common among top-ranked variables of these models. These methods also showed potential in detecting main effects in interaction terms.

"eventScribe", the eventScribe logo, "CadmiumCD", and the CadmiumCD logo are trademarks of CadmiumCD LLC, and may not be copied, imitated or used, in whole or in part, without prior written permission from CadmiumCD. The appearance of these proceedings, customized graphics that are unique to these proceedings, and customized scripts are the service mark, trademark and/or trade dress of CadmiumCD and may not be copied, imitated or used, in whole or in part, without prior written notification. All other trademarks, slogans, company names or logos are the property of their respective owners. Reference to any products, services, processes or other information, by trade name, trademark, manufacturer, owner, or otherwise does not constitute or imply endorsement, sponsorship, or recommendation thereof by CadmiumCD.

As a user you may provide CadmiumCD with feedback. Any ideas or suggestions you provide through any feedback mechanisms on these proceedings may be used by CadmiumCD, at our sole discretion, including future modifications to the eventScribe product. You hereby grant to CadmiumCD and our assigns a perpetual, worldwide, fully transferable, sublicensable, irrevocable, royalty free license to use, reproduce, modify, create derivative works from, distribute, and display the feedback in any manner and for any purpose.

© 2017 CadmiumCD