In the previous post i gave an introduction to what Machine Learning actually is and also told you about the types of machine learning (Take a look at it here if u haven’t already). This post will cover the 2nd type of machine learning, i.e. Unsupervised learning.

Unsupervised learning is a type of machine earning in which no parameters are given to the algorithm at all. This means that unlike Supervised learning, where the algorithm is given a correct database and incorrect database, there is no such differentiation done with the input data given to algorithm. In Unsupervised learning, the data is just dumped into the algorithm. It is then the algorithms work to find out a kind of relation in the data that is given to it.

Consider for example that data of certain selected patients is given to the algorithm. The data consists of age, BMI, disease they suffer from, etc. The algorithm will go through this data and find out relations between them and segregate the data into groups based on the parameters that it observes.

Thus the main difference between Supervised and Unsupervised learning is that the algorithm is already told how to segregate the data in Supervised and is supposed to find its own relation for segregation in Unsupervised learning.

Another important way of splitting machine learning learning into types is on the basis of whether the algorithm is iterative or not. This means whether the algorithm keeps evolving (learning) or whether the machine learning is done only once and the data is continuously compared with this model. (Dont worry if you didn’t understand what this means ;))

The last part of the machine learning series will tell you about the types of machine learning actually used in the industry ( On SVM’s and Neural Nets).

Hope this post helped!! Stay tuned for the last post!