Lecture 02: Machine Learning

                                      Machine Learning 

Lecture Series 02

1.0.0. Introduction to Machine Learning

         Recap on lecture 01:

            We discussed the definition of machine learning, its need and importance, the problems to tackle while practicing Machine Learning. Similarly, we discussed one of the models under machine Learning viz. Linear Regression Model.

          Continuing on lecture 02:

            Now that we have discussed a model under machine learning, let's start formalizing few things                and terms. There are 3 types of learning or ways to train a machine. 

1.0.1. Supervised Learning:

            Supervised Learning is defined as a sub-category of machine learning where the algorithm is trained based on the labeled & classified datasets. To give you intuitive idea, with each data-point, we feed its outcome to train. 

         As input is fed into the model, it adjusts its weights until the model has been fitted appropriately, which occurs as part of the cross-validation process. Linear regression model is an example of supervised learning.

1.0.2. Unsupervised Learning:

        Unsupervised Learning is defined as a sub-category of machine learning where the algorithm is trained based on the classified data. To give you intuitive idea, with each data-point, we expect the model to give an output and keep getting better. 

1.0.3. Reinforcement Learning:

        As the name suggests, in this sub-category, we don't feed any data but we let the model to explore and then with appropriate datasets, exploit.

        In reinforcement learning, its a endless cycle of exploring and exploiting. For any unseen data, model explore and for any known data, it exploits the known pre-information or weight of the data.

1.1.0 Machine learning's flow-chart

To understand different aspect of machine learning, Let's work on a project together. We will work on a project together to build a algorithm that recognizes the 3 different types of iris-flower set based on the image. Before we start, let's discuss few terminologies.

2.1.0. Exploratory Data Analysis (EDA)

        EDA is one of the major step in feature extraction and characterization that helps us to fine tune the given data-sets. EDA is a very different kind of data analysis to understand different key characteristics of data-sets by applying different libraries like (pandaa,NumPy,seaboard in Python).

In EDA, we make different histograms, pdf, cdf to understand the characteristics and choose the correct parameters for training.

2.1.1. Iris Flower project 

    Understand, before analyzing any kind of data-sets, one must has to learn the details of working domain.We don't want you to understand each line of code in the file but want you to learn the flavor of analyzing the data-sets part and a bit of idea about the code you will be writing in future.

Please install jupyter Notebook into your VS code.

You can find a jupyter file attached to this text and go through these discussions.