Introduction to MachineLearning
Machine Learning is a subset of Artificial Intelligence (AI)
that enables systems to learn from data and improve from
experience without being explicitly programmed.
Types of Machine Learning:
1. Supervised Learning: Learns from labeled data.
(e.g., Classification, Regression)
2. Unsupervised Learning: Learns from unlabeled
data. (e.g., Clustering, Dimensionality
Reduction)
3. Reinforcement Learning: Learns through
rewards and punishments. (e.g., Game playing
agents)
Linear Regression
• Asupervised learning algorithm used for predicting a continuous value based
on input features.
Working:
• Fits a linear relationship between input variables (X) and output (y).
Formula: y = mx + c
• Example:
Predicting house prices based on size and location.
5.
Logistic Regression
• Aclassification algorithm that predicts the probability of a binary outcome.
• Working:
Applies sigmoid function to map predicted values to probabilities.
• Example:
Determining whether an email is spam or not.
6.
Decision Tree
• Atree-like structure where each node represents a feature and each
branch represents a decision.
• Working:
Splits the data based on feature values to reach a decision at leaf nodes.
• Example:
Loan approval based on age, income, and credit score.
7.
Random Forest
• Anensemble of decision trees used to improve prediction accuracy.
• Working:
• Builds multiple decision trees and merges their outputs.
• Example:
• Credit risk assessment using aggregated decision paths.
8.
Support Vector Machine(SVM)
• A classification algorithm that finds the optimal hyperplane to
separate classes.
• Working:
Maximizes the margin between different class boundaries.
• Example:
Classifying images into cats and dogs.
9.
K-Nearest Neighbors (KNN)
•A lazy learning algorithm that classifies based on majority class among k-
nearest points.
• Working:
Computes distance between test data and training samples.
• Example:
Recommending products based on user similarity.
10.
Naive Bayes
• Aprobabilistic classifier based on Bayes' theorem with feature
independence assumption.
• Working:
Calculates posterior probability for each class and selects the highest.
• Example:
Sentiment analysis of customer reviews.
K-Means Clustering
• Anunsupervised learning algorithm that groups data into k clusters.
• Working:
Assigns data points to nearest cluster centroid and updates centroids
iteratively.
• Example:
Customer segmentation for marketing strategies.
13.
Apriori Algorithm
• Anassociation rule learning algorithm used to find frequent itemsets in
transactional data.
• Working: Iteratively expands frequent item sets using a bottom-up approach
based on minimum support threshold.
• Example: “If a customer buys bread and butter, they’re likely to buy jam.”
14.
DBSCAN (Density-Based Spatial
Clusteringof Applications with Noise)
• A clustering algorithm that groups together closely packed points and marks
outliers as noise.
• Working: Starts with an arbitrary point and retrieves all density-reachable
points based on a distance (ε) and minimum points threshold.
• Example: Identifying geographical clusters of seismic activity while excluding
isolated outliers.
15.
Principal Component Analysis(PCA)
• A dimensionality reduction technique that transforms features into principal
components.
• Working:
Finds new orthogonal axes (principal components) maximizing variance.
• Example:
Reducing high-dimensional customer data for visualization.
Q-Learning
• A reinforcementlearning algorithm used to learn optimal actions in a
given state.
• Working:
Updates Q-values using Bellman equation based on reward and future value.
• Example:
Training an agent to navigate a maze.
18.
Conclusion
Machine Learning hasrevolutionized the way we process and analyze data.
Key Takeaways:
• Different algorithms serve different purposes—choose based on data type and problem.
• Supervised learning is ideal for prediction with labeled data.
• Unsupervised learning is useful for exploring unknown patterns.
• Reinforcement learning is powerful in dynamic decision-making scenarios.
• Understanding the working and application of each algorithm helps in building better AI
systems.