Batch Processing

Algorithm Parameters

The DBSCAN clustering algorithm has the following parameters:

Parameter

Default Valude

Description

algorithmFPType

float

The floating-point type that the algorithm uses for intermediate computations. Can be float or double.

method

defaultDense

Available methods for computation of DBSCAN algorithm:

  • defaultDense – uses brute-force for neighborhood computation

epsilon

Not applicable

The maximum distance between observations lying in the same neighborhood.

minObservations

Not applicable

The number of observations in a neighborhood for an observation to be considered as a core one.

memorySavingMode

false

If flag is set to false, all neighborhoods will be computed and stored prior to clustering. It will require up to \(O(|\text{sum of sizes of all observations' neighborhoods}|)\) of additional memory, which in worst case can be \(O(|\text{number of observations}|^2)\). However, in general, performance may be better.

Note

On GPU, the memorySavingMode flag can only be set to true. You will get an error if the flag is set to false.

resultsToCompute

\(0\)

The 64-bit integer flag that specifies which extra characteristics of the DBSCAN algorithm to compute.

Provide one of the following values to request a single characteristic or use bitwise OR to request a combination of the characteristics:

  • computeCoreIndices for indices of core observations

  • computeCoreObservations for core observations

Algorithm Input

The DBSCAN algorithm accepts the input described below. Pass the Input ID as a parameter to the methods that provide input for your algorithm. For more details, see Algorithms.

Input ID

Input

data

Pointer to the \(n \times p\) numeric table with the data to be clustered.

Note

The input can be an object of any class derived from NumericTable.

weights

Optional input. Pointer to the \(n \times 1\) numeric table with weights of observations.

Note

The input can be an object of any class derived from NumericTable except PackedTriangularMatrix, PackedSymmetricMatrix.

By default all weights are equal to \(1\).

Note

This parameter is ignored on GPU.

Algorithm Output

The DBSCAN algorithms calculates the results described below. Pass the Result ID as a parameter to the methods that access the result of your algorithm. For more details, see Algorithms.

Result ID

Result

assignments

Pointer to the \(n \times 1\) numeric table with assignments of cluster indices to observations in the input data.

Noise observations have the assignment equal to \(-1\).

nClusters

Pointer to the \(1 \times 1\) numeric table with the total number of clusters found by the algorithm.

coreIndices

Pointer to the numeric table with \(1\) column and arbitrary number of rows, containing indices of core observations.

coreObservations

Pointer to the numeric table with \(p\) columns and arbitrary number of rows, containing core observations.

Note

By default, this result is an object of the HomogenNumericTable class, but you can define the result as an object of any class derived from NumericTable except PackedTriangularMatrix, PackedSymmetricMatrix, and CSRNumericTable.