Skip to main content

INTRODUCTION TO MACHINE LEARNING AND ITS TYPES


The basic definition of Machine Learning can be:

“Give computers the ability to learn without explicitly programmed”.

OR

“The process where the machines learn from data and improve from previous experiences without human intervention”.

And this data can be labeled or unlabeled.

We have seen the internet definition of Machine Learning above, but what actually is this and why Machine Learning?

In earlier days, if we want to decide upon any decision, we used “if” and “else” to process the given input and adjusted to get the desired output. Yes, it is easy to apply on the applications which are easier at understanding, but there are some disadvantages,

  • The logic which is written is only for a single task, if we want to rewrite the logic, we need to rewrite the whole system as well.
  • The problem and logic may be bigger where the human experts cannot handle.

But Machine Learning Algorithms have no such type of problems. The most successful kinds of machine learning algorithms are those that automate decision-making processes by generalizing from known examples and now we can say that Machine Learning is about extracting knowledge from data.

Machine Learning Algorithms are divided into three types,

  • Supervised Learning
  • Unsupervised Learning
  • Reinforced Learning

Supervised Learning:

The foremost important point is “Labeled data” is used to train the model.

Simple Definition can be “Machine Learning Algorithms learns from input-output pairs which are given by human and predict the output for data it has never seen before without any help from human”.

Technical Definition can be “Supervised learning algorithms try to model relationships and dependencies between the target prediction output and the input feature such that we can predict the output values for new data based on those relationships which it learned from the previous data sets”.

Examples of Supervised Machine Learning:

Determining whether a tumor is benign or not

Image is the input and we need to classify the output whether the tumor is benign or not. To build this model, we need a dataset of medical images and a medical expert who can review the images and decides which tumors are benign and which are not.

Detecting fraudulent activity in credit card transactions

Input is the record of credit card transactions and output is likely to classify whether the transaction is fraudulent or not. To build this model, we need a dataset that stores all transactions of users and records where the user reports the fraudulent transaction.

Identifying the zip code from handwritten digits on an envelope

Input is scanning the handwritten digits and the desired output is actual digits in zip code. To build this model, we need a dataset of many envelopes where we can read the zip codes and store the digits as the desired outcomes.

Unsupervised Learning:

The foremost important point is “Unlabeled data” is used to train the model.

Simple Definition can be “Group the data in order to find the similarities between the features or attributes in the dataset”.

Technical Definition can be “Unsupervised learning algorithms looks for previously undetected patterns in a data set with no pre-existing labels and with a minimum of human supervision [Wiki]”.

Examples of Unsupervised Machine Learning:

Segmenting customers into groups with similar preferences

We want to identify which customers are similar and if there are any customers with similar preferences, for example, “books”,” music”,” games”. We don’t in advance what type of customers they are, hence there will be no known outputs.

Identifying topics in a set of website posts

Here, we want to summarize the findings on a large data collection where we previously do not know about such topics. There are no known outputs.

Reinforced Learning:

It can be called a “Goal Oriented Algorithm”.

This Algorithm aims to learn sequences of actions that will lead an agent to achieve its goal or maximize its function objective. It works on the concept of rewarding the right step and penalizing the wrong step.

And we won't be much discussing this algorithm anywhere when compared to the above two algorithms.

I hope you enjoy reading this and I hope you understand what is the simple definition of Machine Learning and its types.

Comments

Post a Comment

Popular posts from this blog

SUPERVISED MACHINE LEARNING ALGORITHM

Supervised Learning is one of the successful types of Machine Learning. Remember that supervised learning is predicting the outcome from given input where we have examples of input-output pairs. The Goal is to “ predict the outcome for unseen data ”. One important thing about supervised Machine Learning is it requires human effort to build a training set ( input-output pairs), but later automates and often speeds up the task. There are two types of Supervised Machine Learning Algorithm Classification   Regression Classification: The goal of Classification is to predict a class label (which can be termed as outcome sometimes). Classification can sometimes be distinguishing between exactly two classes which are called  binary classification and classification between more than two classes are called multiclass classification . Binary classification can be referred to as answering a YES/NO question. For example, Is this email spam? The answer is either yes or no. Exa...
Post a Comment
Read more