In the past, I did many Business Intelligence projects in the field of healthcare. In those kind of projects it is very uncommon to use techniques of Machine Learning and Artificial Intelligence, because business stakeholders have their own methods and procedures to deal with data. In a population based cancer registry for instance the active monitoring for cancer cases in regional spread is a subject of discussion. Moreover in some other fields you find existing law regulations.
But in our daily life these techniques are omnipresent. We can see it on social media or online shopping: algorithms decide what we see on the news and set prices we have to pay for a particular product.
Therefore to me it feels natural to dive into this subject. Currently I am in the phase of learning and understanding.
To acquire knowledge, next to using textbooks I take free online university courses, also called Moocs. It is a great concept learning new things regardless of time and place. In addition it is fantastic to have the opportunity to take courses at some of the best universities in the world like Harvard, Berkeley or MIT. I currently enjoy edx.org to learn.
To get an idea of the capabilities of Machine Learning I attend courses about R-Project, Apache Spark and Azure Learning. What I’ve found so far is that for ML you need two things: First, know your data and have the capabilities to visualize the data to find pattern and influences and second, algorithms to model these findings. In ML these algorithms are mostly linear systems of equitation.
Next to the procedures interesting to see is the spread of these technologies. Recently, cloud provider like Amazon and Microsoft provide services to easily create ML services on your own. So in the near future we will find many services that will provide AI on a click.
I am currently using these services to create my own little ML sandbox. I hope I’ll soon be able to present some of my findings here.