Rice Analytics

Automated Reduced Error Predictive Analytics

We invented Reduced Error Logistic Regression (see Press Release for our US patent 8,032,473), and this portfolio also now includes two pending patent applications.  Reduced Error Logistic Regression (RELR) is a completely automated and very general machine learning method that avoids traditional problems related to dimensionality and multicollinearity.  RELR is the subject of a book by Daniel M. Rice titled Calculus of Thought: Neuromorphic Logistic Regression in Cognitive Machines published by Elsevier, Academic Press in November of 2013. RELR is currently used daily to help profile likely customers and thus target media buying in a very large number of US consumer brands using a SAS macro application that we developed a few years ago.

Key News Events
February 5, 2014.  We are making progress in developing our own simple Python and C-based code to implement the RELR methods in a more widespread venue than our current SAS implementation. This Python/C implementation could be utilized in a stand alone massively parallel API or could be implemented in existing logistic regression software like in-memory processing systems or other massively parallel processing. This Python/C implementation is based upon RELR being largely embarrassingly parallel. For example, even much of the pre-processing in RELR can be made suitable for parallel processing by breaking the data into sub-samples of observations or features where the computation of interaction features, nonlinear features and means and standard deviations for t values can run in parallel on each sub-sample of observations or features.  The same is true of most other aspects including the logistic regression as detailed in Calculus of Thought.

January 28, 2014.  Calculus of Thought is now fully released in both print and electronic versions and the companion website to the book is now also available at this link to the Elsevier website: http://booksite.elsevier.com/9780124104075/   This companion website contains a pdf document that lists a couple of small typing errata in the book and links to Excel Workbooks that run easy to understand toy model examples from the book that anyone is able to download even without purchasing the book. 

November 15, 2013.  Calculus of Thought is available as of today for its first day of sale out on Amazon.com and the Store.Elsevier.com websites and will be available at Barnes and Noble and other booksellers in the next few weeks.  The companion website to the book which contains a few small errata and links to Excel Workbooks that run toy model examples in the book is still under construction, but should be complete within the next week or so. 

November 1, 2013. Update: You can now "Look Inside" Calculus of Thought out on Amazon.com and read much of the preface and first chapter.  Dan Rice will be presenting at Society for Neuroscience next week in San Diego and likely will have an advance copy of a printed book at that time.

January 11, 2013.  Dan Rice is under contract with Elsevier for a book on RELR to be titled Calculus of Thought likely to be published later this year in the fall.  The title is based upon Leibniz's idea that the goal of calculus is to develop a Calculus Ratiocinator or a computational machine that mimics human cognition but without the subjective biases of humans.

Oct 4, 2011.  We were issued a patent today for our RELR technology by the US Patent Office. A full description of the significance of this patent is available at this link to our Patent Press Release Article

August 4, 2011.  A new case study is available from a completely independent researcher with no connections to Rice Analytics that is a comparison of RELR with Random Forests Logistic Regression, LASSO, LARS, Stepwise Regression, and Bayesian Networks and shows that RELR outperforms these other algorithms in classification accuracy by an average of 2-4%.  Here is the link RELRCaseStudyAugust2011 to the page on our website where it can be viewed.

September 9, 2009.  St. Louis, MO (USA) - Our new executive white paper written by Dan Rice and entitled "Breiman's Quiet Scandal: Stepwise Logistic Regression and RELR" was in the Publications section of the online industry newletter KDnuggets.com on August 27, 2009 (issue 09:n16).  This item had the Most Clicks by Subscribers and was the 2nd Most Viewed item overall of 41 items that were published that week.   This article written in "plain business English" for executives reviews the major difficulties with Stepwise Logistic Regression that were pointed out by the late statistician Leo Breiman.  This article also reviews evidence that our RELR method may be a solution to these problems. The complete white paper can be downloaded by clicking this link to the Executive White Paper  page of this website. 

June 15, 2009.  St. Louis, MO (USA) - Dan Rice gave an invited address last week at the 2009 Classification Society Annual Conference from June 11-13 at the Washington University Medical School.  This conference brought together roughly a hundred experts from major universities and businesses in the areas of machine learning, choice modeling, and classification research. This conference was truly international in scope and had attendees from many industrialized countries. However, the relatively small size of this conference compared to JSM allowed for an extended discussion between attendees over the course of several days.  The title of this talk was "Reduced Error Logistic Regression".  This talk can be downloaded from the Papers and Presentations page of this website.         

August 6, 2008. Denver, CO (USA) - Dan Rice spoke today at the Data Mining and Machine Learning Session of the 2008 Joint Statistical Meetings in Denver, Colorado. This session was chaired by Bill Heavlin of Google Inc. and had good speakers from the United States Army, Medical University of China, University of Alabama, Bell Labs, and the University of California at Berkeley.   This session was extremely well attended with a standing-room-only crowd.  This standing-room-only crowd and the lively discussions prompted Bill Heavlin to say that this session "was the best session at the conference".  The title of Rice's talk was "Generalized Reduced Error Logistic Regression Machine".  In this presentation, Rice provided evidence that Reduced Error Logistic Regression is able to reduce error significantly compared to Penalized Logistic Regression, Step-Wise Logistic Regression and four other standard methods.  A full article coinciding with this talk and published in JSM 2008 Proceedings can be downloaded from the Papers and Presentations page of this website.   The Joint Statistical Meetings is one of the largest gatherings of statisticians in the world.  Approximately 5000 people attended this conference in Denver this summer.


Copyright, 2014 Rice Analytics, All rights reserved.
Rice AnalyticsTM  is a trademark of Rice Analytics, St. Louis, MO. SAS is a registered trademark of SAS Institute, Cary, NC.

Machine Learning  Segmentation  Consumer Surveys  Predictive Modeling  Risk Management