What is Supervised Machine Learning? || LN-TECHINFO

on

Supervised Machine Learning

Supervised Machine Learning
As the name suggests, Supervised Learning has a Supervisor who we can also call a teacher. 
Basically Supervised Learning is a learning in which we train the machine with Supervised Data or Labeled Data, then after this training, the machine is able to estimate the correct output for any other Input.

Let's understand this with an example, so let's say you have a bucket which has four kinds of fruits like apple, banana, grapes and oranges. Now, because we have already told the Supervised Model about the fruits in which color and which shape is its shape, it will easily separate all of them. So we call it Supervised Learning because we had already trained the model with training data.

So in this way we can say that it is learning in which we train the model with Labeled or Training Data which is actually Training Example ; Consist by keeping, then this model becomes capable that it can easily classify any kind of problem or Situation.

Steps of Supervised Machine Learning

 In order to solve any supervised learning problem, we have to follow these steps.
  1. Determine the type of training examples: First of all, find out what type of data will be used in the training set.
  2. Gather a training set: Then preparing their training set , in this along with the Input Objects, their Output Objects are also included.
  3. Determine the input feature representation of the learned function: Then determine the Input Feature Representation of the function to learn means that the Input Function should have enough information to get the correct output from it. There should be neither more nor less information in it because if this happens then the model will either not be able to learn properly or will learn too much so that there is a possibility of error.
  4. Determine the structure of the learned function and corresponding learning algorithm: Then to find out the structure of this function means that it will use Support Vector Machine (SVM) or Decision Trees.
  5. Complete the design: Then complete this complete design.
  6. Run the learning algorithm on the gathered training set: So now we have to run this algorithm for our training set .
  7. Evaluate the accuracy of the learned function : To measure the performance of the Resultant Function after learning.

Algorithms

Some of the most used algorithm names -
  • Support Vector Machines (SVM)
  • Linear Regression
  • Logistic Regression
  • Naive Bayes
  • Linear Discriminant Analysis
  • Decision Trees
  • K-Nearest Neighbor Algorithm(KNN)
  • Neural Networks (Multilayer perceptron)
  • Similarity

              A variety of algorithms by engineers; are used which are divided into two Groups. These algorithms are called supervised learning methods. These methods are - Regression and  Classification. The main goal of both is to determine the relationship or structure in the input data which can get us the correct output data and both have Goal to find the value of the dependent variable from the attribute variable. The only difference between the two is that the dependent attribute for the regression; is numerical and the dependent variable for classification; It is categorical.

              So here comes a question, when should we use which method?

              Classification: Output Variables in this are called Labels or Category because it has output; In the form of categorical means it means that the output is divided into two or more categories such as an email can be categorized in two ways Spam and Not Spam, categorize an age group into three children, young and older. do. 

              So when we have to solve an AC problem where we have to create categories or labels. 

              So we use Classification Algorithms 

              Discrete or Real-Valued in Classification; Input can be Variable.

              Classification problem can be described in two ways-
              1. Single Class Classification
              2. Multi Class Classification
              Regression: Output Variables; There are continuous values ​​such as Amount, Volume or Size. For example, if the price of a house can be between 5 lakh to 10 lakh, then Regression; Task is to predict the constant quantity. 

              The Regression Problem can also be described in two ways -
              1. Linear Regression
              2. Non-Linear Regression

              Comments

              Post a Comment