Machine Learning algorithms used to enable computers Algorithms:
techniques are methods or of task, the nature of the data, and the learning process.
Techniques:
1. Supervised Learning
The model is trained Machine learning (ML) techniques output are known.
Algorithms:
o (e.g., predicting house prices)
o ression (e. Linear Regression g., binary classification for spam detection)
o Machines (SVM)
o Neural Networks (for structured data tasks like classification or regression)
Applications: Fraud detection, sales forecasting, sentiment analysis.
o 2. Unsupervised Learning Decision Trees aLogistic Regnd Random Forests
Support Vector
The model is trained on unlabeled data to identify underlying patterns or structures.
Algorithms:
o Clustering (e.g., K-Means, DBSCAN)
o , PCA, t-S Dimensionality Reduction (e.g.NE)
o Association Rules (e.g., market basket analysis)
Applications: Customer segmentation, anomaly detection, recommendation systems.
3. Semi-Supervised Learning
A to improve combination of labeled and unlabeled data is used model accuracy.
Techniques: Self- odels.
As: Text classificatrpplicationaining, graph-baslimited labeled data, mediced mtion with al
diagnosis systems.
4. Reinforcement Learning
orming An agent ecisi ons by perf learns to make dactions in an environment to maximize
rewards.
, actions, rewards.
Key Elements: Algorithms:
o ing
o Deep Q-Networks (DQN)
o Policy
Applications: Robotics GStates Q-Learnradient Methods, gaming (e.g., AlphaGo), autonomous
vehicles.
5. Neural Networks and Deep Learning
Specialized models inspired by the human brain, used for complex data and large datasets.
s:
o Convolutional Neural NetwoTypes of Networkrks (CNNs): For image and video analysis.
o works (RNNs): For sequRecurrent Neural Netential data like time series or natural
language.
o Transformers: For NLP tasks like translation and sentiment analysis (e.g., BERT, GPT).
Ans: Image recognition, natural language processing, deep reinforcement learning.
Techniques: pplicatio
a new Leveraging a pre-trained task.
Fed Learning: devices while Traneratedsfer Learning: model for Decentralized ML across
preserving data privacy.
s fornGenerativ image gee Models: GANeration Creating new data (e.g.,).
