Sign In

Communications of the ACM

ACM TechNews

Crowdsourcing Big-Data Analysis


A Web-based system automatically evaluates proposals.

Researchers at the Massachusetts Institute of Technology have developed a collaboration tool intended to make feature identification more efficient and effective.

Credit: MIT News

Researchers at the Massachusetts Institute of Technology (MIT) have developed FeatureHub, a collaboration tool to make feature identification more efficient and effective.

FeatureHub lets data scientists log onto a central site, review a problem, and propose features, and then tests various feature combinations against target data to determine which are the most useful for a given predictive task.

During testing, the researchers recruited 32 analysts with data science experience, who each spent five hours working with the system and used it to propose candidate features for each of two data-science problems. FeatureHub generated predictive models that were tested against those submitted to a data-science competition, and the models were within three and five points of the winning entries for the two problems.

"The concept of massive and open data science can be really leveraged for areas where there's a strong social impact but not necessarily a single profit-making or government organization that is coordinating responses," says MIT's Micah Smith.

From MIT News
View Full Article

 

Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA


 

No entries found