Introduction to Machine Learning

Hi! Yesterday my book "Hands-on Machine Learning" arrived, so I wanted to give you a little introduction to this topic, based on the book. What is Machine Learning(ML)? "Machine Learning is the field of study that gives computers the ability to learn without being explicily programmed." (Arthur Samuel)

A good example is your E-Mail Spam Filter. The Spam filter learned to identify spam because it has been trained by examples of spam emails, the so called training set. When you mark an Email as spam this training instance will become part of the training set. With Machine Learning you avoid to program each spam case by yourself.

Not only can machines learn from humans, its also possible the other way around. When the spam filter has been trained for a while, we can inspect it by letting it reveal the list of words that it believes are the best predictors of spam. When you apply ML on large amounts of data(="data mining") you can discover patterns.

ML systems can be trained with or without human supervision (supervised/unsupervised). In supervised learning the training data includes the desired solutions, which are called labels. This means that you give the ML algorithm examples of Emails which are already marked as "not Spam" and examples marked as "Spam".

In unsupervised learning we dont use labels. You simply run different ML algorithms on the data. You could for example run a clustering algorithm on your Blogs visitors to detect groups of similar visitors. A possible result of the algorithm could be that 30% of the visitors are female and visit your blog on average at 8 p.m.

What interesting things do you know about Machine Learning? Have a nice day.