Anomaly Detection

  1. Machine Learning-Based Anomaly Detection
    1. Unsupervised Learning Approaches
      1. Clustering-Based Methods
        1. K-Means Clustering
          1. Cluster Assignment Strategies
            1. Distance to Centroids
              1. Silhouette Analysis
                1. Limitations and Assumptions
                2. Hierarchical Clustering
                  1. Agglomerative Methods
                    1. Divisive Methods
                      1. Dendrogram Analysis
                        1. Outlier Identification Strategies
                        2. Gaussian Mixture Models
                          1. Expectation-Maximization Algorithm
                            1. Component Assignment
                              1. Likelihood-Based Scoring
                              2. Spectral Clustering
                                1. Graph Laplacian
                                  1. Eigenvalue Analysis
                                    1. Anomaly Detection Applications
                                  2. Isolation-Based Methods
                                    1. Isolation Forest
                                      1. Random Tree Construction
                                        1. Path Length Calculation
                                          1. Anomaly Score Interpretation
                                            1. Parameter Tuning
                                            2. Extended Isolation Forest
                                              1. Isolation-Based Ensemble Methods
                                              2. Reconstruction-Based Methods
                                                1. Principal Component Analysis
                                                  1. Reconstruction Error Calculation
                                                    1. Principal Component Selection
                                                      1. Subspace Methods
                                                      2. Autoencoders
                                                        1. Standard Autoencoders
                                                          1. Architecture Design
                                                            1. Training Procedures
                                                              1. Reconstruction Error Analysis
                                                              2. Variational Autoencoders
                                                                1. Latent Space Modeling
                                                                  1. KL Divergence
                                                                    1. Anomaly Score Calculation
                                                                    2. Denoising Autoencoders
                                                                      1. Sparse Autoencoders
                                                                      2. Matrix Factorization Methods
                                                                        1. Non-negative Matrix Factorization
                                                                          1. Singular Value Decomposition
                                                                            1. Robust PCA
                                                                        2. Semi-Supervised Learning Approaches
                                                                          1. One-Class Classification
                                                                            1. One-Class Support Vector Machine
                                                                              1. Kernel Functions
                                                                                1. Linear Kernel
                                                                                  1. RBF Kernel
                                                                                    1. Polynomial Kernel
                                                                                    2. Hyperparameter Optimization
                                                                                      1. Decision Boundary Analysis
                                                                                      2. Support Vector Data Description
                                                                                        1. Spherical Boundary Construction
                                                                                          1. Kernel Trick Application
                                                                                            1. Parameter Selection
                                                                                          2. Positive-Unlabeled Learning
                                                                                            1. Two-Step Methods
                                                                                              1. Biased Learning Methods
                                                                                                1. Ensemble-Based Approaches
                                                                                              2. Supervised Learning Approaches
                                                                                                1. Traditional Classification Methods
                                                                                                  1. Logistic Regression
                                                                                                    1. Feature Engineering
                                                                                                      1. Regularization Techniques
                                                                                                        1. Probability Calibration
                                                                                                        2. Support Vector Machines
                                                                                                          1. Kernel Selection
                                                                                                            1. Cost-Sensitive Learning
                                                                                                              1. Multi-class Extensions
                                                                                                              2. Decision Trees
                                                                                                                1. Tree Construction
                                                                                                                  1. Pruning Strategies
                                                                                                                    1. Feature Importance
                                                                                                                    2. Ensemble Methods
                                                                                                                      1. Random Forest
                                                                                                                        1. Gradient Boosting
                                                                                                                          1. AdaBoost
                                                                                                                            1. XGBoost
                                                                                                                          2. Deep Learning Methods
                                                                                                                            1. Feedforward Neural Networks
                                                                                                                              1. Architecture Design
                                                                                                                                1. Activation Functions
                                                                                                                                  1. Regularization Techniques
                                                                                                                                  2. Convolutional Neural Networks
                                                                                                                                    1. Image-Based Anomaly Detection
                                                                                                                                      1. Feature Map Analysis
                                                                                                                                      2. Recurrent Neural Networks
                                                                                                                                        1. LSTM Networks
                                                                                                                                          1. GRU Networks
                                                                                                                                            1. Sequence Modeling
                                                                                                                                          2. Handling Class Imbalance
                                                                                                                                            1. Sampling Techniques
                                                                                                                                              1. Undersampling Methods
                                                                                                                                                1. Oversampling Methods
                                                                                                                                                  1. Hybrid Sampling
                                                                                                                                                  2. Cost-Sensitive Learning
                                                                                                                                                    1. Class Weight Adjustment
                                                                                                                                                      1. Cost Matrix Design
                                                                                                                                                      2. Threshold Optimization
                                                                                                                                                        1. ROC-Based Selection
                                                                                                                                                          1. Precision-Recall Based Selection
                                                                                                                                                          2. Ensemble Approaches
                                                                                                                                                            1. Balanced Bagging
                                                                                                                                                              1. Balanced Random Forest
                                                                                                                                                                1. EasyEnsemble

                                                                                                                                                          Previous

                                                                                                                                                          4. Proximity-Based Anomaly Detection Methods

                                                                                                                                                          Go to top

                                                                                                                                                          Next

                                                                                                                                                          6. Time-Series Anomaly Detection