by Angela Guess
According to a recent press release, “Feature Labs, Inc., the industry’s first software solution that automates feature engineering for machine learning and AI applications, today launched… Feature Labs, Inc. was founded out of the Computer Science and AI Lab at MIT – CSAIL by Max Kanter, Kalyan Veeramachaneni and Ben Schreck. The company’s software products combine open source and proprietary algorithms that enable data scientists to perform automated and precise feature engineering on structured data so that machine learning and predictive modeling can be leveraged at scale and allows customers to rapidly iterate on prediction problems to achieve business goals. Organizations including Kohl’s, NASA, BBVA, DARPA, Monsanto and MIT have been using Feature Labs software to automate their feature engineering processes.”
The release goes on, “Today the process of feature engineering is manual, time consuming, error-prone and frequently does not work. Current methods do not allow for newcomer data scientists & domain experts to make an impact. The best way for machine learning and predictive models to enable business value is to automate this critical step. According to Gartner analyst Dr. Carlton Sapp, ‘Machine Learning enterprise applications will require a strong automated feature engineering process in order to deploy effective predictive models.’ Feature Labs uses ‘Deep Feature Synthesis’ to automatically create features from raw relational and transactional datasets such as user behavior logs or credit card transactions. The software creates complex features similar to humans by stacking primitive feature functions using the relationships between rows and tables in a dataset. As it explores a dataset, it updates the recommendations of the top features to use for specific prediction problems.”
Read more at Business Wire.
Photo credit: Feature Labs