Why machine learning works

During our illustration of training the ML model, we talked about generating a form of representation that is used to build our ML system. One thing to note is that this information or the data used to generate the form of representation is the data representation of our future source of information. Why do we need the data representation of our future source of information? In this sub-section, we will look into this and see how it helps in creating ML models.

Let's assume we are told to carry out research about product interest in a particular community. Imagine this community consists of a large number of people and we are only able to reach a few people – let's say 50% of the community population.

The idea is that, from the information obtained from 50% of the population, we should be able to generalize for the remaining 50% of the population. We assume this because a set of people from the same community or population are assumed...