According to Fujitsu its newly-developed technology estimates machine-learning results from a small set of sample data and the accuracy of past predictive models, “extracting the learning algorithm and configuration combination that produces the most accurate result, and applies it to the larger dataset."
“This results in highly accurate predictive models from datasets of 50 million records in a few hours,” Fujitsu says in today’s announcement.
Conversely, according to Fujitsu, current techniques for generating highly accurate predictive models need to examine every combination of learning algorithm and configuration, taking more than one week to learn from a dataset containing 50 million records.
Fujitsu will release the details of the technology in presentations at the meeting of the Information-Based Induction Sciences and Machine Learning (ISIMBL) opening Monday, 14 September at Ehime University in Japan.
Backgrounding the development of its machine-learning technology, Futjitsu says the popularity of smartphones and other advances make it possible to gather massive quantities of sensor data, and “machine learning and other advanced analytic techniques are being used extensively to extract valuable information from that data.”
And, according to Fijitsu, using the access logs of e-commerce websites, for example, it is possible to discover when people are most likely to cancel memberships on a given website, to identify those people quickly, and to take measures to discourage cancellation.
In addition, the company says using detailed daily power-consumption data, it is possible to discover patterns of increased or decreased usage and to predict periods and times when power usage will increase - leading to a reduction in power costs by applying more precise controls over power generation, transmission, and storage.