Introduction To Machine Learning World

!pip install pycaret &> /dev/null print("pycaret installed sucessully!!")

This piece of code was basically my first footstep in the world of machine learning during my first semester at the university, I remember our professor gave us a file containing some code ( In python programming language ) and told us to copy that code into our system. We did the same and I remember after following what our professor told us to do I was able to predict the price of the car based on the mileage of the car, engine power, number of seats, Kilometers driven etc.

This program which I made even by copy-pasting rooted a deep seed of curiosity in me towards the world of Machine learning . So from there onwards I dived deep into this domain and today I will be sharing with you the most simple and interesting , Introduction To Machine Learning

First of all, before knowing about Machine learning itself, let me introduce the family of Machine learning, Machine learning is the only child of Artificial Intelligence, and machine learning also have its own child known as Deep learning , herecis the simple illustration to understand this hierarchy better 👇

Now since you are aware about the definition of machine learning , now equip yourself with the knowledge about the types of machine learning So there are basically 3 types of Machine learning :

Supervised , Unsupervised and Reinforced are the main 3 categories of Machine learning even though there are other also like semi supervised learning but these 3 are the ones to be taken into consideration . But now the problem arrises that how to know which type of machine learning has to be implemented in which case 🤔

The answer for this is simple " Just Observe your data " , because your observation or analysis of data will give you clear signals about which type of machine learning will best suite your data , by the way doing analysis of your data is not at all difficult thing to do , just remember this 👇

But wait a minute, what about the third type of machine learning, when do we use it?

For now, it is first important to completely understand supervised and unsupervised learning because reinforcement learning is completely different from these 2 conventional types and you will not be able to grasp that without knowing completely about supervised and unsupervised machine learning.

So first, let us start with supervised machine learning. Supervised Machine learning is used in the case of labeled data where the labeled data means that for every input we are providing to our machine learning algorithm there is an output related to it and that output is basically a label and when we our whole data have such labels for the features we provide as input then we use supervised machine learning

For example: We have to make a machine learning model that will predict what would be the best price to sell a car based on some features like : mileage, number of tyres, engine power , transmission of the car

So here for making this machine learning model we are provided with some data which is shown below , now according to you what type of machine learning will best suite this type of data ?

If your answer was supervised machine learning , then congratulations you guessed it right , and if not then don't worry let me explain that why is it supervised machine learning problem and not the unsupervised one

Basically here in this case our data is well labeled means for the features like ( year , km_driven , fuel , seller_type and transmission ) , we have label ( selling price ) , so here we will use the supervised machine learning

Also If you want to know how the data for unsupervised machine learning looks like then here it is

In this data, we are only provided with the features ( i.e. only input ) and our machine learning model will have to work more in order to find some patterns for making predictions in the absence of the labeled output

Now since you are aware about where we use which type of machine learning , so in the next blog post we will take a closer look at how reinforcement learning is different from these 2 conventional types of machine learning