dbscan (version 1.1-9)

predict: Assigning New Points to a Clustering

Description

Predict the membership of a new point given a clustering.

Usage

# S3 method for dbscan_fast
predict(object, newdata, data, ...)
# S3 method for optics
predict(object, newdata, data, ...)
# S3 method for hdbscan
predict(object, newdata, data, ...)

Arguments

object

clustering object.

newdata

new data points for which the cluster membership should be predicted.

data

the data set used to create the clustering object.

...

further arguments.

Details

For DBSCAN: If a new point is in the eps neighborhood of a point in the clustering, the the label of that point is predicted, otherwise the prediction is 0 (noise).

For OPTICS: Extracts a DBSCAN clustering and use predict for DBSCAN.

For HDBSCAN: Find the k-nearest neighbors with the smallest mutual reachability distance to predict the label

See Also

Other clustering functions: dbscan(), extractFOSC(), hdbscan(), jpclust(), optics(), sNNclust()