Brief Overview on Supervised Machine Learning | Types of Machine Learning

Supervised Machine Learning is a type of Machine Learning in which machines are trained based on labeled data and based on data provided machine learning model is capable to predict future outputs. The labeled data means some amount of correct output data is provided to a machine learning model.

Table of Contents

1. Introduction to Supervised Learning
2. How supervised Learning algorithms Works?
3. Steps Involved performing any Supervised Machine learning tasks
4. Types of Supervised Learning
• Regression
• Classification
5. Advantages and Disadvantages of Supervised Machine Learning
6. End Notes

In supervised learning, machines work under supervision, and here supervisor is training data that teaches the machine to predict the output correctly. This is the same as a student learns under the supervision of a teacher in class. 

This is a straightforward concept, only focus on the things and visualize the situation and you will be capable to understand each and everything. So, supervised learning is a technique in which input data and correct output data are provided to a model and based on which it learns and can predict the output on unseen data. 

How Supervised Machine Learning Works?

In supervised machine learning, the model is trained based on training data, and then to know the performance we test it on unseen data known as testing data. The first step in any type of machine learning is to prepare and understand your data. After that in supervised Machine learning if the data is too huge then it is divided into two parts training and testing and then the modeling process begins.

The working of supervised machine learning can be understood with the help of the example discusses below.

Working of Supervised Machine Learning

Suppose we have a dataset with different images of animals which include dog, cow, cat, goat, frog. Now the first step is to train the model on each label data.
The model will try to depict the label based on the various characteristics like color, palettes, brightness, shape, and many more to understand that, on these features, And try to find the hidden patterns and learn the characteristics of each label.

Now, after training, we test our model, and the task of a model to identify a correct animal.

Steps in Supervised Learning

we have understood how supervised learning works. let's take note of step-by-step procedures to follow while working with supervised machine learning.

• First, understand the type of dataset.
• collect the labeled training dataset.
• split the dataset into training, test, and validation dataset. splitting should be done in such a manner that training data should contain the proper distribution of data.
• Determine the suitable algorithm for the model to be trained on.
• Execute the algorithm on a training dataset. sometimes we need validation datasets as control parameters, which are a subset of the training dataset.
• Evaluate the performance of a model on the test set. If the model gives the correct output and performs better on new data then, it means our model is accurate.

Types of Supervised Learning Algorithm

Supervised Machine learning can be further classified into 2 categories:

1) Regression                                2) Classification

1) Regression: 

Regression is used to find the relationship between the input variable(independent variable) and the output variable(dependent variable). Regression is used for the prediction of continuous values, such as stock prices, sales prediction, Fees prediction, weather forecasting.

Below are some of the popular supervised machine learning algorithms used to perform the Regression task.

• Linear Regression
• Polynomial Regression
• Regression Trees
• Support Vector Regression
• Bayesian Linear Regression

1) Classification: 

Classification algorithms are used when the output variable is categorical. It is used when you want to classify the outputs in two or more two classes such as Spam Detection, Loan approval, Disease detection(Person is having a disease or not). 
when output is having two classes then it is known as Binary classification such as yes/no, Male/Female, True/False, Spam/Ham.

When multiple classes are there, it is known as a Multiclass classification problem such as the name of an animal, color name or you can say when you have a nominal or ordinal variable in output.

The most popular supervised algorithms used for classification are:

• Logistic Regression
• Decision Trees
• Random Forest
• Support Vector Machines
• K-Nearest Neighbour
• Naive Bayes

Advantages of Supervised Learning

• A model can predict the output better based on prior experience.
• We have an exact idea about the classes or objects.
• Supervised learning helps to solve many real-world use cases such as fraud detection, Trend detection, etc.

Disadvantages of Supervised Learning

• Supervised learning models are not capable to handle complex tasks.
• models cannot predict the output correctly if training and test data distributions are different.
• The training phase requires lots of computation time.

Summary

• In supervised machine learning, you train machines based on labeled data known as training data.
• You have to train a model based on some data which will help you to drive future relations and predictions.
• Regression and Classification are two types of Supervised Machine learning.
• The biggest challenge in supervised machine learning is a different distribution of datasets(irrelevant inputs) which produces inaccurate results.
• The biggest advantage of supervised learning is you are aware of outputs and it allows you to train and produce output based on prior experience.
• As a best practice, you must decide what kind of data be used as training data.

I hope that the concept and intuition of supervised learning are clear. we will be delving deep into each algorithm and understanding complete mathematical intuition with code.

keep learning, keep smiling😊.