Modeling in Revolution R Enterprise(TRA108)
Monday, June 9, 2014 at 8:00 AM - Friday, June 13, 2014 at 11:00 AM (PDT)
San Francisco, California
London, United Kingdom
This training is divided into 3 session and each session is scheduled on a different day. The details of which are given below
06/9/2014 from 8AM - 11AM
06/11/2014 from 8AM - 11AM
06/13/2014 from 8AM - 11AM
This course is designed for Data Scientists who have mastered the basics of R and are interested in learning how to take advantage of the capabilities of Revolution R Enterprise for high performance analytics and modeling. This is a hands-on course filled with real data and examples, case studies, and in-class mini projects.
Introduction to Revolution R Enterprise for Predictive Modeling
One of the many advantages of Revolution R Enterprise is its ability to build predictive models on large enterprise-sized datasets. We will begin with an introduction to the predictive modeling functionality within Revolution R Enterprise.
- Review of key Revolution R Enterprise programming concepts and data preparation
- Algorithm and function overview
- Standard function definitions and parameterization
Linear Regression Modeling and Evaluation
We will first introduce techniques for performing multivariate linear regression modeling available in Revolution R Enterprise.
- Simple and multivariate regression models
- Complex formulas and higher order terms.
- Model review and evaluation including holdout evaluation.
- Model selection using stepwise regression.
- Predictions, model objects, and implementation.
Generalized Linear Models
Revolution R Enterprise includes capabilities to run bigdata GLM models, including logistic regression and tweedie models. We will introduce examples that show off these techniques:
- Logistic regression model building and evaluation.
- Additional forms for bigdata GLMs.
- Predictions and implementation.
Data Mining using Trees and Forests
Revolution R Enterprise allows modelers to use decision trees and decision forests to build predictive models on big data.
- Tree modeling functions and usage
- Model evaluation and graphical review
- Optimization of tree-building parameters.
- Prediction and implementation.
Unsupervised Models and Other Techniques
Revolution R Enterprise includes additional advanced capabilities such as k-means clustering and principal components analysis for big data.
- Introduction to clustering and principal components functions and techniques.
- Review and evaluation of results.
- Running simulations using rxExec