Identifies rare or unexpected items in datasets, often used in fraud, system health monitoring, or security breaches.
Techniques: Isolation Forests, One-Class SVM, Autoencoders, Z-score, DBSCAN
Anomaly Detection
Computer Vision
Extracting, analyzing, and generating meaningful information from digital images or video streams.
Techniques: CNNs, YOLO, U-Net, Haar Cascades, GANs
Data Reduction & Transformation
Dimensionality Reduction
Reduces feature space to make data easier to visualize or process while preserving structure.
Techniques: PCA, t-SNE, UMAP, LDA
Interpolation and Curve Fitting
Fills missing data or smooths values across a range using known reference points.
Techniques: Linear, Spline, Polynomial, Bezier
Learning Problems
Classification
Sorting input into labeled categories (e.g., spam vs. non-spam).
Techniques: Logistic Regression, SVM, Decision Trees, CNNs, k-NN
Regression
Predicting continuous values based on one or more input features.
Techniques: Linear, Polynomial, Ridge, Lasso, GPR, Neural Nets
Clustering
Grouping data by similarity without predefined categories.
Techniques: k-Means, DBSCAN, GMM, SOM
Reinforcement Learning
Learning optimal actions via environment feedback—used in robotics, games, and autonomous systems.
Techniques: Q-Learning, DQN, Actor-Critic, MCTS
Natural Language Processing
Analyzing and generating human language for applications like sentiment analysis, translation, or chatbots.
Techniques: Transformers, TF-IDF, Word2Vec, RNNs, Topic Modeling
Optimization & Simulation
Optimization
Finding the best solution from many, often under constraints (e.g., cost, speed, efficiency).
Techniques: Linear/Integer Programming, Gradient Descent, Genetic Algorithms, Constraint Programming
Simulation and Modeling
Creating digital models to explore system behavior, often used in physics, economics, logistics.
Techniques: Monte Carlo, Agent-Based Modeling, Markov Chains, FEA
Search & Recommendation
Recommendation Systems
Suggesting relevant items based on behavior or preferences (e.g., Netflix, Amazon).
Techniques: Collaborative Filtering, Matrix Factorization, Hybrid Models
Search and Ranking
Retrieving and ranking information (e.g., web search).
Techniques: TF-IDF, BM25, Inverted Indices, Neural Search
Graph Theory
Analyzing connected data like social networks, routing, knowledge graphs.
Techniques: Dijkstra's, BFS, DFS, GNNs, PageRank
Time-Based Analysis
Predicting future values using historical trends (e.g., stock forecasting, weather prediction).
Techniques: ARIMA, RNNs, Prophet, Holt-Winters, State Space Models