LocalOutlierFactor<T>
Models & Types · AiDotNet.AnomalyDetection.DistanceBased
Implements the Local Outlier Factor (LOF) algorithm for density-based anomaly detection.
Inheritance: ModelBase<T, Matrix<T>, Vector<T>> → AnomalyDetectorBase<T> → LocalOutlierFactor<T>
Implements: IGeneratedParameterRegistrar<T>
For Beginners
LOF detects anomalies by comparing the local density of a point to the local densities of its neighbors. Points in low-density regions (relative to their neighbors) are considered anomalies.
How It Works
The algorithm works by:
- Finding the k-nearest neighbors for each point
- Computing the local reachability density (LRD) for each point
- Comparing each point’s LRD to its neighbors’ LRDs to get the LOF score
- Points with LOF score significantly greater than 1 are anomalies
When to use: LOF is particularly effective for:
- Detecting local anomalies (points that are anomalies relative to their neighborhood)
- Data with varying densities across different regions
- When you want interpretable density-based detection
Industry Standard Defaults:
- Number of neighbors (k): 20 (good balance between local and global detection)
- Contamination: 0.1 (10%)
Reference: Breunig, M. M., Kriegel, H. P., Ng, R. T., and Sander, J. (2000). “LOF: Identifying Density-Based Local Outliers.” ACM SIGMOD Record.
Constructors
| Constructor | Summary |
|---|---|
LocalOutlierFactor(int numNeighbors, double contamination, int randomSeed) |
Creates a new Local Outlier Factor detector. |
Properties
| Property | Summary |
|---|---|
int NumNeighbors { get; } |
Gets the number of neighbors used for LOF computation. |
Methods
| Method | Summary |
|---|---|
Vector<T> ScoreAnomalies(Matrix<T> X) |
|
void Fit(Matrix<T> X) |