ManhattanDistance<T>
Models & Types · AiDotNet.Clustering.DistanceMetrics
Computes Manhattan (L1) distance between vectors.
Inheritance: DistanceMetricBase<T> → ManhattanDistance<T>
For Beginners
Manhattan distance is like walking along city blocks. You can only move horizontally or vertically, not diagonally.
Example: The distance between points (0, 0) and (3, 4) is 7 (|3-0| + |4-0| = 3 + 4 = 7)
Good for:
- High-dimensional data where Euclidean distance loses meaning
- When features have different scales
- Sparse data
How It Works
Manhattan distance, also called taxicab or city-block distance, is the sum of absolute differences along each dimension. It’s named after the grid-like street layout of Manhattan.
Formula: d(a, b) = sum(|a[i] - b[i]|)
Constructors
| Constructor | Summary |
|---|---|
ManhattanDistance() |
Properties
| Property | Summary |
|---|---|
string Name { get; } |
Methods
| Method | Summary |
|---|---|
T Compute(Vector<T> a, Vector<T> b) |