Automated Linear Regression for Really, Really Big Data

What do you think about Inora, a company that advertises itself as the New Linear Regression Approach - Scalable to Big Data. Our data scientist also developed automated regression for big data, offering source code, even an Excel implementation, and comparing his results with traditional regression techniques. Click here for details.

This type of product raises another issue: is linear regression a technique worth using on very large data sets? This article disagrees on this. This one also cites concerns about regression in general.

Having said that, I would love to hear a different opinion. What other companies offer reliable, totally automated regression? How do they compare to Inora? At least Inora offers free source code, but limited to 1,000 observations They seem to think that more than 1,000 observations qualifies as big data, and thus should be charged a license fee. Here are some of their selling points:

The InoraRAE (Robust Analytical Evaluation) is an ADI (Artificial Deterministic Intelligence) program, in this case applied to linear regression. The ADI software creates value by automatically detecting unexpected deviations and finds the true behavior of the data set, with no human intervention.

The InoraRAE does not use any human decisions, rules or inputs when analyzing data. This provides unbiased data analysis and instant access to the true behavior of the data.

This software version of the InoraRAE Linear Regression is for personal and/or educational use only. It is limited to a maximum of 1,000 data points and two-class functionality (finding expected vs. unexpected only

DSC Resources

Career: Training | Books | Cheat Sheet | Apprenticeship | Certification | Salary Surveys | Jobs
Knowledge: Research | Competitions | Webinars | Our Book | Members Only | Search DSC
Buzz: Business News | Announcements | Events | RSS Feeds
Misc: Top Links | Code Snippets | External Resources | Best Blogs | Subscribe | For Bloggers

Additional Reading

1 member likes this

Comment

You need to be a member of Data Science Central to add comments!

Join Data Science Central

Comment by Jared Armatis on February 17, 2016 at 4:51am: Inora is not claiming that the free version of the RAE Linear Regression software can solve complex data challenges.

The linear regression is just an application of the generalized core Math Engine to that specific task (y=mx+b). Of course, the free version is limited to a 2-class task (finding what is line and what is not line).

The Inora Math Engine is also capable of detecting and analyzing multiple patterns in any size data set. So, even though the free linear regression software is limited, it will demonstrate how the Math Engine core is different from traditional statistics, random sample and Least Squares approaches. It is the first tool of its kind to provide a purely numerical combinatory procedure which determines the mechanical components in data autonomously; hence it is the first Hyper-Mechanics data analysis tool available.

The Math Engine at the core provides the all decisive mechanical components for the automatic decision making process to reveal hidden or latent constraints, noise, outliers - simply all Unexpected Deviations (CD) in general! Based on those result parameters any functional model analysis becomes truly deterministic and true knowledge finding. Especially when it comes to Big Complex Data!

Linear Regression is by no means the solution to complex data, but the core Math Engine is the ultimate solution to big, complex data.