Fujitsu announces new machine-learning technology

by

12 September 2015

Fujitsu Laboratories says it has developed a technology that estimates machine-learning results from a small set of sample data and the accuracy of past predictive models, extracts the learning algorithm and configuration combination that produce the most accurate result, and applies it to the larger dataset. This will result in highly accurate predictive models from datasets of 50 million records in a few hours, the company said.

According to Fujitsu, the predictive models produced by this technology can work to quickly make improvements, such as minimising membership cancellations on e-commerce websites and enhancing response times to equipment failures.

The company said it has built up a database of combinations of previously used algorithms and configurations, along with the accuracy of the predictive model they produced, and uses this to estimate the predictive accuracy of new combinations. This, it says, makes it possible to make an assessment based on the smallest amount of data possible without sacrificing predictive accuracy.

The company ran internal tests using a dataset of 50 million records and eight servers with 12 processor cores each. Existing techniques, according to Fujitsu, would take roughly one week to develop a predictive model with 96 per cent accuracy but this new technique reached that level in slightly more than two hours.

Additionally, it is demonstrated that this technology would make the practical application of machine-learning possible when used for access-log analysis with 30 million lines of web access logs, the company said.

Fujitsu Laboratories will release details of this technology at the meeting of the Information-Based Induction Sciences and Machine Learning (ISIMBL), which opens 14 September at Ehime University in Japan.