SOM Predictor TODO
Labels / Tags
- Predictor
- Numerical vector
Principle
$K$MeansPredictor inherit from ClosestPrototypePredictor and follow the same logic in a more specific way. Here both queries and prototypes have share the type numerical vector. Result of a query will be the ClusterId of the closest $K$MeansModel prototype.
Scalability
When dissimilarity measure complexity is considered negligeable the computational complexity of a KMeansPredictor query is in O(K)
Where :
- K is the number of prototype of the $K$MeansModel.
If there are n queries, the complexity will be O(K.n). Considering than in most of case K is neglieable compared to n, a collection of query will be executed in linear time.
Be sure to know the computational complexity of used dissimilarity measure.
Input
Single query
A numerical vector value.
Multiple queries, i.e collection of data observations
A collection of numerical vector value.
Parameters
1 : KMeansModel
KMeansModel contains the list of prototypes returned by KMeans algorithm.
Output
Single query
Returns the ClusterId of the closest $K$MeansModel prototype.
Multiple queries, i.e. a collection of queries of numerical vectors
Returns the HardClustering associates to input data.
Associate visualizations
- HardClustering