Scribd
Upload
20 views6 pages

Learning Types in Machine Learning

The document outlines three main types of machine learning: Supervised Learning, Unsupervised Learning, and Reinforcement Learning. Supervised Learning uses labeled data for predictions, while Unsupervised Learning operates on unlabeled data to find patterns, and Reinforcement Learning involves agents learning through feedback from their actions. Each type has specific applications and subcategories, such as regression and classification in supervised learning, and clustering and association in unsupervised learning.

Uploaded by

chilukasoonu
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
20 views6 pages

Learning Types in Machine Learning

The document outlines three main types of machine learning: Supervised Learning, Unsupervised Learning, and Reinforcement Learning. Supervised Learning uses labeled data for predictions, while Unsupervised Learning operates on unlabeled data to find patterns, and Reinforcement Learning involves agents learning through feedback from their actions. Each type has specific applications and subcategories, such as regression and classification in supervised learning, and clustering and association in unsupervised learning.

Uploaded by

chilukasoonu
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd

TYPES OF LEARNING

Types of Machine Learning


At a broad level, machine learning can be classified into three types:

1. Supervised Learning
2. Unsupervised Learning
3. Reinforcement Learning
SUPERVISED LEARNING
In this type of machine learning algorithm,
 Supervised Learning is the most popular paradigm for performing ML operations. It
is widely used for data where there is a precise mapping between Input-Output data.
 The training data set is a labelled data set. i.e., the algorithm identifies the features
explicitly and carries out predictions or classification accordingly.
 In other words, the training data set contains the input value (X) and target value
(Y).
 The learning algorithm generates a model.
 Then, new data set (testing data) consisting of only the input value is fed.
 The model then generates the target value based on its learning.
 Diagram:

 Example:
Consider a sample database consisting of two columns where-
 The first column specifies mails.
 The second column specifies whether those emails are spam or not.
 In this training data set, emails categorized as spam or not are done by a
supervisor’s knowledge.
So, it is supervised learning algorithm.

Mails (X) IsSpam (Y)

Mail-1 Yes

Mail-2 No

Mail-3 No

Mail-4 No

 Real-Life Examples
 Spam Filtering
 House Price Prediction
 Credit Scoring (high risk or a low-risk customer while lending loans by the
banks)
 Face Recognition etc
 Types of Supervised Machine Learning
There are two types of supervised learning algorithm-
1. Regression
2. Classification

 Regression
 The target variable (Y) has continuous value or real value. (like rupees,
height….)
 Example- house price prediction, Risk assessment, Score prediction etc….
 Classification
 The target variable (Y) has discrete values or category or a group such as Yes or
No, 0 or 1, black and white, spam or no spam and many more.
 Example- Credit Scoring, Email Spam Filtering, Image classification, Fraud
detection, Diagnostics etc….
UNSUPERVISED LEARNING
 Unsupervised learning is a learning method in which a machine learns without any
supervision.
 The training is provided to the machine with the set of data that has not been
labelled, classified, or categorized, and the algorithm needs to act on that data
without any supervision.
 In this type of machine learning algorithm,
The training data set is an unlabelled data set.
In other words, the training data set contains only the input value (X) and not the
target value (Y).
Based on the similarity between data, it tries to draw inference from the data such as
finding patterns or clusters.
 Diagram

 Diagram

 Real-Life Applications
 Document Clustering
 Finding fraudulent transactions
 DNA patterns to analyse evolutionary biology.
 Customer segmentation, or understanding different customer groups around
which to build marketing or other business strategies.
 Types of Unsupervised Machine Learning
Unsupervised Learning can be further classified into two types, which are given
below:
 1. Clustering
 2. Association

 Clustering
 Clustering where you want to discover the inherent groupings in the data.
 Ex: grouping customers based on their purchasing behaviour
 Association
 Association where you want to discover rules that describes large portions of
your data.
 Ex: people that buy X also tend to buy Y
REINFORCEMENT LEARNING
 Reinforcement learning is a feedback-based learning method, in which a learning
agent gets a reward for each right action and gets a penalty for each wrong action.
 The agent learns automatically with these feedbacks and improves its performance.
 In reinforcement learning, the agent interacts with the environment and explores it.
 The goal of an agent is to get the most reward points, and hence, it improves its
performance.
 The agent acts in an environment in order to maximize the rewards and minimize the
penalty.
 Follow trial and error method to solve the problem.
 Unlike supervised learning, no data is provided to the agent.
 The agent itself takes action or sequence of actions whether right or wrong to
perform a task and learn from the experience.
 Popular algorithms are Q-learning and SARSA.
 Diagram

 Example
You can see a dog and a master. Let’s imagine you are training your dog to get the
stick. Each time the dog gets a stick successfully, you offered him a feast (a bone
let’s say). Eventually, the dog understands the pattern, that whenever the master
throws a stick, it should get it as early as it can to gain a reward (a bone) from a
master in a lesser time.

 Real-life Applications
 Game Playing
 Robot Navigation
 Self-driving cars.

Common questions

Powered by AI

Unsupervised learning is preferred for tasks like customer segmentation because it can analyze unlabelled data to identify inherent groupings and patterns. This ability to cluster data based on similarities without predefined categories allows businesses to glean insights about customer behavior and preferences, which helps in building targeted marketing strategies.

The absence of labeled data in unsupervised learning necessitates reliance on algorithms that can automatically discover patterns or structures in data. This impacts its application because it can be more challenging to assess the quality of results without pre-defined labels, requiring the use of indirect methods such as evaluating the cohesion and separation of discovered clusters to determine effectiveness.

The concept of reward and penalty in reinforcement learning can be compared to training a dog. When a dog successfully fetches a stick, it receives a reward in the form of a treat, reinforcing the desired behavior. Over time, the dog learns that fetching the stick is associated with a reward and adjusts its actions to increase positive outcomes. This mirrors how an agent in reinforcement learning uses feedback to improve its actions.

Reinforcement learning is commonly applied in areas such as game playing, robot navigation, and self-driving cars due to its feedback-driven learning process. This type is suited for these applications as it relies on trial and error to find optimal strategies, allowing the agent to adapt to dynamic environments and make decisions that maximize cumulative rewards.

Regression and classification are both approaches within supervised learning. Regression is used for predicting continuous values, such as house prices or risk assessments, while classification deals with categorizing discrete outcomes, like email spam detection or image classification. Regression applications often involve numerical predictions, whereas classification focuses on categorizing data into groups.

Association in unsupervised learning involves discovering rules that can describe large portions of data, such as identifying items frequently bought together. This can benefit retail marketing strategies by enabling businesses to create effective product bundles or recommend complementary products, thus improving sales and customer satisfaction.

Reinforcement learning is effective in developing self-driving cars because it allows the vehicle to autonomously learn from its environment through trial and error. As it receives feedback on its driving decisions, it can improve its performance by maximizing rewards, such as safety and efficiency, while minimizing penalties like accidents or inefficiencies. This adaptability is crucial for navigating the complex, dynamic nature of real-world driving environments.

Supervised learning requires labeled datasets, meaning the data consists of input-output pairs, which helps the algorithm identify features explicitly to perform predictions or classifications. Examples include spam filtering and house price prediction. In contrast, unsupervised learning operates on unlabeled data with only input values, and it aims to draw inferences like patterns or clusters without external guidance. It is used for tasks such as customer segmentation and document clustering.

Clustering in unsupervised learning can uncover hidden structures or relationships within data sets by identifying groups based on similarities between data points. This process can reveal patterns that were not apparent, such as customer groups with similar purchasing behaviors, potentially leading to new insights and strategies for targeting different segments.

Reinforcement learning differs from supervised learning as it does not require a labeled dataset for training. Instead, it relies on feedback through rewards and penalties to learn. The learning agent interacts with its environment, making decisions to maximize rewards, unlike supervised learning where the model is trained with known input-output pairs.

You might also like