Translate into a different language

Friday, May 01, 2015

From Sci-fi to reality: Machine Learning by Joey Withers

Joey Withers, DBA & Senior Database Developer at Rockpool Digital, discusses the development and business applications of Machine Learning.

It almost seems like the beginning of a sci-fi movie when people talk about machine learning. It refers to systems that learn and adapt to data, rather than follow pre-programmed instructions.

While we won’t be facing Skynet’s Terminators anytime soon, companies are now starting to understand the advantages of intelligent systems.

Data has become one of the most coveted assets in the world of business. It can be used to provide a better product or service by helping companies to understand what customers want.

The systems that analyse this data are powered by complex algorithms. Historically, these algorithms were pretty set in stone and if a company wanted a different output from its analyses, the algorithm needed to be changed.

Now, the systems can adapt its algorithm depending on the data that it receives.

Take Facebook, for instance. Algorithms used to automatically tag people in a photograph have a higher success rate, as it ‘learns’ if it was correct based on interactions from its users. The more times a person uses the feature, the more opportunities there are for the algorithm to learn and the better it becomes.

Machine learning is improving, but due in part to the specialist skills required to utilise the technology, its wider usage is still limited to programmers. That could soon be changing.

Microsoft now has Machine Learning functionality on its Azure cloud computing platform. The platform allows users to build predictive models via a graphical drag and drop interface, without the need to write any code.

With easier access to this technology, businesses can identify its most influential clients, ensure that inventory is in the right place at the right time and predict results, therefore allowing the appropriate action to be taken.

Azure ML is designed specifically to make machine learning accessible to business users and more generalised developers.

Classically, in order to benefit from Machine Learning, you would require an in depth understanding of statistics and mathematics and be skilled in writing code in R, the functional programming language used among statisticians and data miners to utilise machine learning.

Source: The Information Daily

If you enjoyed this post, make sure you subscribe to my Email Updates!