Supervised vs Unsupervised Machine Learning
The algorithm of machine learning works in two ways in particular:
- Supervised Learning
- Unsupervised Learning
It is the data that the machine will be processing is based on some historical background, i.e., it will be checking some past events related to the data and will further proceed with defining the data. In other words, it deals with the data when provided with ‘labeled’ inputs earlier.
For example, let us assume that we have trained the machine by providing two labels, red and blue by programming it with the following two pictures:
Now there enters a new figure and the machine has to put it in a category of either red or blue.
So, what it will do is that as we have trained it by providing labels, it will check that with which group the features of the new one matches. After verifying the features, it will put the new one accordingly.
This shows that the machine has learned from the previously provided data and now it is applying that to the new information provided.
There are two categories of Supervised Learning:
It defines that the machine model is not trained. It is not provided with any sort of previous data or labels through which it can learn. So, here the model group the data provided on the basis of similarities it finds in the data. Machine Model makes its own structure and learns by itself.
For example, you provided the machine with some figures as shown below:
The machine does not know what each of them represents or what is their labels. So what it will do is that it will pick the things which have similarity on their physical features and will group them together like below:
There are two categories of Unsupervised Learning: