header
header1
The Modeling Agency

THE MODELING AGENCY NEWSLETTER
data mining / predictive modeling / actionable analytics
guidance and results for the data-rich, yet information-poor

header2
IE Group
TDWI
Predictive Modeling News
destination crm
Predictive Modeling News


In This Issue
FEATURE ARTICLE
TRAINING NEWS
  • Register Early for the Popular Fall Washington-DC Production
    TMA's most popular training production of the year is the fall event at the Crystal Gateway Marriott in Washington, DC. Be sure to prepare your training request now to ensure your space and secure the very early bird discount.
  • Additional TMA Training Resources
    Links to the course series overview, what makes TMA's training unique, introductory webinar, as well as dates and locations for upcoming course productions.
INDUSTRY NEWS & ANNOUNCEMENTS
Feature Article

Alternative Variable Selection Techniques
for Enhanced Classification Models

Using Proportional Variance to Rank
Condition Attributes for Robust Analytics


The Data Warehousing Institute FlashPoint Article
Senior Consultant & Training Director,
Thomas A. "Tony" Rathburn
The Modeling Agency, LLC

Abstract
training_imageOne of the major issues facing an analyst in completing any analytic model is the selection of condition attributes.

In statistics, the coefficient of determination, termed R2 or R-squared, is a measure of how well data points fit a statistical model--the goodness of fit. R2 is a statistic, with a range of 0 to 1, and is commonly reported in a wide variety of statistical modeling techniques. It provides a measure of how well observed outcomes are replicated by the model as the proportion of total variation of outcomes explained by the model.

The computation of proportional variance may offer a markedly different relative ranking of condition attributes, and often results in significantly higher levels of performance when used for condition attribute selection in classification projects.

Because the outcome is categorical and often binary, we are interested in selecting condition attributes that differentiate between the categories. In general, the stronger the proportion variance of a condition attribute relative to the outcome, the more important that condition is to our model...

[ >> Read the Full Article << ]


Training News

Register Early for the Popular Fall Washington-DC Production

The Modeling Agency's most popular training production of the year is the fall event at the Crystal Gateway Marriott in Washington, DC. This September venue has sold out early in the past. Be sure to prepare your training request now to ensure your space and secure the very early bird discount. While your request is being processed, it's a good idea to submit an 'unofficial registration' to hold your space in the interim.

If you're looking for actionable predictive analytics training before the fall, then move quickly to sign up for the Denver, CO series starting on July 14. It's now last call for Denver. Be sure to before Denver registration closes.


[ >> Schedule and Registration << ]



line

TMA's Courses are The Most Effective and
Impactful Training Events Available

TMA's vendor-neutral training is truly like no other in the industry. The primary features of the recently redesigned series from which attendees will benefit are:

  • Staged Progression of Courses as illustrated in the diagram below. Courses may be attended independently, or as a progressive series that builds the tactical and strategic soft skills required to thrive in this practice.
  • Updated Analytic Process Model that modernizes and streamlines the CRoss-Industry Standard Process for Data Mining. The new series now methodically steps through the 6-Phase Modeling Process Methodology in order - building a project from design through results measurement and reporting.
  • Highly Coordinated Work-along Labs allow participants to experience and directly deal with the natural messiness of predictive modeling. The Excel data workbook doesn't just house the lab data, but a full project design environment for documenting, visualizing and reporting results. This template is a highly useful project design, implementation and tracking tool for participants to apply to their own projects.
  • Fully Engaging Delivery motivated by the shortcomings observed in most on-line training delivery that suffer from up to 94% attrition. After establishing a baseline through four hours of asynchronous presentation in the Core Concepts prerequisite, delivery then shifts into highly interactive cycles of short presentation, demonstration and active work-along lab sessions. Gone are the days of extended PowerPoint lectures.
  • Comprehensive Course Materials that provide truly hundreds of pages of concentrated process-guide style material. Visit the course series overview page to learn more about how the content is conveyed in terms of Issues / Insights / Illustrations to provide instruction, application and examples for the most effective skill reinforcement. Corresponding presentation slides accompany the written material by topic. Participants return home with templates, roadmaps and modeling session files.
  • Unrestricted and Unbiased Methods and Tool Exposure with demonstrations leveraging various open-source and commercial software to demonstrate the capabilities and limitations of various methods in difference scenarios.
  • Certification with Formal Accreditation lends credibility to the paper with the NASBA CPE stamp. TMA's courseware has been scrutinized and authorized by this strict training authority.

Additional details on the structure, content and benefits of the progressive course series may be referenced on the following pages:

If you want to attain the reinforced skills to truly stand out in the field of advanced analytics, data mining and predictive modeling, dive straight into The Modeling Agency's Predictive Analytics & Mining course series. The series is designed to get you up to speed faster and more effectively than any other analytic program available. Reserve your space today.


[ >> Schedule and Registration << ]


Courses May Be Delivered At Your Site
Call (281) 667-4200 opt 3 or send an email inquiry to receive
a value-based quotation for training at your site.

Government Buyers
TMA is a SAM and IPP Registered
Veteran-Owned Small Business.



Industry News & Announcements

TDWI World Conference Boston
The Premier Event for Business Intelligence, Data Warehousing, and Analytics
July 20-25, Boston, MA
TMA's Senior Consultant & Training Director
Thomas A. "Tony" Rathburn Will Keynote.


Join TDWI in Boston, July 20-25 and see TMA's Senior Consultant and Training Director Tony Rathburn deliver the conference keynote.

The TDWI Boston World Conference offers education courses for every need along the analytics continuum. From foundational data warehousing approaches, to essential business intelligence principles, all the way to the most advanced business analytics-TDWI provides in-depth training in a conference setting.

The TDWI World Conference in Boston offers a full track-17 full- and half-day courses-that will help you create a solid data architecture, including classic disciplines such as master data management, data integration, and data quality. Join us in Boston to learn how to build a robust BI framework that includes the dashboards, KPIs, and reporting your analysts need.

At the conference you will interact with world-class instructors, network with hundreds of peers, and get product information with minimum hype andhassle. You'll also get access to an exhibit hall full of tools and solutions to meet all your BI, DW, and analytics needs.

Don't float away on a sea of information. Evolve your analytics infrastructure now so you and your organization can take advantage of the best that business intelligence has to offer. Join TDWI in Boston to learn how.

Learn more about the TDWI World Conference in Boston and get registered for courses today.

[ >> Proceed to Conference Details << ]


line

Introduction to R Webinar
By Lion Data Systems, LLC
August 12, 10:00a US Pacific DST


Do you want to learn about R, a free and open-source statistical software package that many people say is even more powerful and flexible than SAS and SPSS? R is a free and open-source statistical software package that many people say is even more powerful and flexible than SAS and SPSS.

This free webinar will provide a general introduction to R and cover aspects such as predictive modeling, data preparation, sub-setting, filtering, visualizations and diagnostics. The session's agenda will consist of a lecture, live demo and question/answer period.

Please note that in addition to attending from your laptop or desktop computer, you can also attend from a Wi-Fi connected iPhone, iPad, Android phone or Android tablet by installing the GoToMeeting App.

This event is brought to you by Lion Data Systems, the world's only R training firm that specializes in one-on-one (private) R training. This live event will consist of a lecture, live demo and Q&A session.

[ >> Sign Up for Webinar << ]


About TMA

The Modeling Agency (TMA) provides training, mentorship and solutions to the data-rich, yet information-poor. More than a hundred highly-recognizable organizations have called upon TMA for strategic direction to establish an internal predictive modeling practice.

Summary

The Modeling Agency newsletter provides training schedule updates, informative non-promotional articles authored by seasoned business practitioners across industries, predictive modeling tricks and tips, as well as current data mining industry announcements. Issues are distributed on a monthly to quarterly basis.

This free newsletter is shared only with those who have activated a subscription and double-opted in through an emailed verification link.

We are grateful for your continued interest. However, if you wish to no longer receive TMA’s newsletter, simply click this link to unsubscribe. You will receive an email verifying your release.

line
phone: (281) 667.4200 | training: (888) 742.2454
One Oxford Centre | 301 Grant St, Suite 4300 | Pittsburgh, PA 15219
header header header