“Machine learning”: Microsoft creates platform for analysing data in the cloud

 
Sarah Spickernell
Follow Sarah
Users will need access to a web browser (Source: Getty)
Microsoft has created a platform, located in the cloud, to allow people to analyse large quantities of data.
Called Azure, it uses something called "machine learning", which has been around for over 20 years but not available to the masses. It involves using algorithms to look at past behaviour and make business predictions about revenues, spend and customer habits.
Joseph Sirosh, vice president of machine learning at Microsoft, says it could be used by businesses to make predictions that would increase revenues and decrease costs. “It can help businesses bring out every bit of value from the data,” he told City A.M.
“Machine learning has been in existence for a while, but has been limited to large enterprises since deploying the software can take years,” said Sirosh. “It has typically been very hard to build the systems, and only a few large organisations have managed it.”
But as the availability of data is increasing, there is more opportunity to make the most out of large data sets. This new machine-learning “cloud” aims to “democratise” the tool by giving a wider range of people the ability to quickly analyse their own data, as long as they have access to a web browser.
So where can machine learning be put into practice? And what types of business can it benefit? Something already adopted by a number of organisations is website recommendations – using past activity on a site to make predictions about other things a person will be interested in looking at: “A website looking at your previous actions can help you solve your purchasing problems.”
Another area for which Sirosh believes machine learning has potential is medicine: “Machine learning can be used to predict the probability of a patient being admitted to hospital again,” he explained.
“The service is entirely resident in the cloud, so all businesses can connect to it and start publishing their own machine-learning applications there. Anyone with a background in statistics or engineering can start using it in day-to-day life.”

Related articles