WELCOME TO THE HOOD OF MACHINE LEARNING

SYNOPSIS

1. INTRODUCTION
2. WHAT IS MACHINE LEARNING
3. WHY MACHINE LEARNING
4. MACHINE LEARNING TYPES
5. END TO END MACHINE LEARNING

INTRODUCTION

When most people hear “Machine Learning,” they picture a robot: a deadly Terminator, But it is not you think it is.

Machine learning is a subfield of artificial intelligence (AI). Machine learning aims generally to understand the structure of data and fit that data into models that can be understood and utilized by people. Sound confusing right?

Machine learning is simply giving the machines a thinking ability, making them think like a human for making a decision. Making the machines learn by themselves is called machine learning.

So the next question arises, How does the machine learn? to answer this question ask yourself, right from childhood till now how did you learn the circumstances and the situation around you? EXPERIENCE is the answer. For humans, we use Experience similarly the machine uses the data where data is the experience for the machine, and the machine is revised through the data(Experience) to gain the knowledge and finally, it is used to make a decision.

eg. if there is a person who wants to predict whether the stock market price will rise/ falls that person will go and ask an experienced person who is skilled at the stock market, similarly by using the data we also can predict the stock market price with much precision faster and efficient way.

WHAT IS MACHINE LEARNING

1. Machine Learning is the science (and art) of programming computers so they can learn from data.

Here is a slightly more general definition:

2. Machine Learning is the field of study that gives computers the ability to learn without being explicitly programmed.

— Arthur Samuel, 1959

And a more engineering-oriented one:

3. A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E.

— Tom Mitchell, 1997

So the whole machine learning is like a newborn child to you, where you teach it at some time and make themself to teach on their own.

WHY MACHINE LEARNING

Machine learning models learn, identify patterns, and make decisions with minimal intervention from humans. Ideally, it increases accuracy and efficiency and removes (or greatly reduces) the possibility of human error.

There is still confusion between Machine Learning and Programming, where programming is also used for developing machine learning models, but the main difference is illustrated below,

In programming what we usually do is we make the computer to find the output:

a=5 , b= 10

c=a+b

print(c)

this is what usually happens while we work with a program where we make the computers to find the output and not how it finds it simply makes an addition based on the user given input, it simply performs what we say on the other hand machine learning is simply a higher grade math function where teachers make the student find the value of x so we know how the function work this applies same to machine learning

10 + x =25 find x?

then here goes what we do is put the 10 to the other side of the assignment operator and change the sign and It becomes 15 which obviously assigns the value of x

10 + x =25

x=25–10

x=15

wherein general programming the machines do what we give as instruction wherein machine learning it tries to learn and implies it to the current situation it has given to work with. From the above example, it is learned to find x by putting the value next to it putting it to the other side of the assignment operator, and then changing the sign and do the operation, even if given a similar situation like this instead of 10 the situation is given as (-20) it works ad doesn't throw an error, since it learned how to do (generalization)and not what to do?

From this, we can understand that machine learning models can learn from the data and then generalize their knowledge from the data, and then it will give us a prediction.

So the whole machine learning is like a newborn child to you, where you teach it at some time and make themself to teach on their own.

MACHINE LEARNING TYPES

There are many criteria where its type is classified, here I'm going to simply speak only on Human Supervision only

1. Supervised learning
2. Unsupervised learning
3. Semisupervised learning
4. Reinforcement Learning

Another type is whether it can learn on-fly or not

1. online versus batch learning

Another type is whether it can find patterns or not

1. instance-based versus model-based learning

SUPERVISED LEARNING

In supervised learning, the training set you feed to the algorithm includes the desired solutions, called Labels.

Supervised learning is simply learning where it is told to identify the objects, that is dissecting supervised learning, consider yourself inside a classroom where your teacher(Experienced person)will teach you about everything that you are learning, or as a small child, your parent(Experienced person) will teach or help you to identify dogs vs cat as they provide a detailed difference between it (Labelling).

To be precise, similarly here we teach the machine with a labeled dataset.

Supervised learning is classified as REGRESSION and CLASSIFICATION based on their Data to be predicted

Regression — where it predicts the continuous data(Present in all timesteps)

Classification — where it predicts the discrete data(Present in particular timesteps only)

ALGORITHMS COMES UNDER SUPERVISED LEARNING :

Algorithms are simply the Teacher

REGRESSION :

1. Linear Regression
2. Multiple Linear Regression
3. Polynomial Regression
4. Decision Tree Regression
5. Random Forest Regression
6. Support Vector Regression
7. KNN Regression
8. Ridge Regression
9. Lasso Regression
10. Elastic Net Regression
11. Artificial Neural Networks(Deep learning)

CLASSIFICATION :

1. Logistic regression
2. Decision Tree classifier
3. Random Forest Classifier
4. Support Vector Classifier
5. KNN Classifier
6. Artificial Neural Network(Deep Learning)

UNSUPERVISED LEARNING

In unsupervised learning, as you might guess, the training data is unlabeled. Same as a Situation where You will be learning things from your own experience without a teacher teaching you(Unlabeled), that is you will be finding the nuke and corner of your own that's how unsupervised learning works.

Here are some of the most important unsupervised learning algorithms

@ Clustering

1.K-Means

2.DBSCAN

3.Hierarchical Cluster Analysis (HCA)

@ Anomaly detection and novelty detection

1.One-class SVM

2.Isolation Forest

@ Visualization and dimensionality reduction

1.Principal Component Analysis (PCA)

2.Kernel PCA

3.Locally Linear Embedding (LLE)

4.t-Distributed Stochastic Neighbor Embedding (t-SNE)

@ Association rule learning

1.Apriori

2.Eclat

SEMI-SUPERVISED LEARNING

Since labeling data is usually time-consuming and costly, you will often have plenty of unlabeled instances and few labeled instances. Some algorithms can deal with data that’s partially labeled. This is called semisupervised learning.

eg. Google Photos is the best example where based on the one photo you say it as yours(Supervised)and the model automatically finds all the other photos of yours(face)and will make a new album of it(Unsupervised).

REINFORCEMENT LEARNING

Reinforcement Learning is a very different beast. The learning system, called an agent in this context, can observe the environment, select and perform actions, and get rewards in return (or penalties in the form of negative rewards). It must then learn by itself what is the best strategy, called a policy, to get the most reward over time. A policy defines what action the agent should choose when it is in a given situation. This is a very good example of how typically a human learns.

END TO END MACHINE LEARNING

1. Look at the big picture.

Analyzing what the problem is.

2. Get the data.

Collecting the real-world data

3. Discover and visualize the data to gain insights.

Gathering insights from collected Data

4. Prepare the data for Machine Learning algorithms.

Data cleaning

5. Select a model and train it.

Model training

6. Fine-tune your model.

Hyperparameter tuning

7. Present your solution.

Displaying the big picture

8. Launch, monitor, and maintain your system.

Deploying our working model to a real-world solution

So the whole machine learning is like a newborn child to you, where you teach it at some time and make themself to teach on their own.

IN MY UPCOMING BLOGS I WILL ELABORATELY DISCUSS EACH OF THE ALGORITHMS COMPLETELY.

Check out My previous post on Artificial Intelligence: https://karthikvegeta.medium.com/a-childs-introduction-to-artificial-intelligence-c95cb7fb695d