Stats Silhouette Score

Source header: cuvs/stats/silhouette_score.hpp

Silhouette Score

stats::silhouette_score

main function that returns the average silhouette score for a given set of data and its clusterings

1 float silhouette_score(
2 raft::resources const& handle,
3 raft::device_matrix_view<const float, int64_t, raft::row_major> X_in,
4 raft::device_vector_view<const int, int64_t> labels,
5 std::optional<raft::device_vector_view<float, int64_t>> silhouette_score_per_sample,
6 int64_t n_unique_labels,
7 cuvs::distance::DistanceType metric = cuvs::distance::DistanceType::L2Unexpanded);

Parameters

Name	Direction	Type	Description
`handle`	in	`raft::resources const&`	raft handle for managing expensive resources
`X_in`	in	`raft::device_matrix_view<const float, int64_t, raft::row_major>`	input matrix Data in row-major format (nRows x nCols)
`labels`	in	`raft::device_vector_view<const int, int64_t>`	the pointer to the array containing labels for every data sample (length: nRows)
`silhouette_score_per_sample`	out	`std::optional<raft::device_vector_view<float, int64_t>>`	optional array populated with the silhouette score for every sample (length: nRows)
`n_unique_labels`	in	`int64_t`	number of unique labels in the labels array
`metric`	in	`cuvs::distance::DistanceType`	Distance metric to use. Euclidean (L2) is used by default Default: `cuvs::distance::DistanceType::L2Unexpanded`.

Returns

float

The silhouette score.

stats::silhouette_score_batched

function that returns the average silhouette score for a given set of data and its clusterings

1 float silhouette_score_batched(
2 raft::resources const& handle,
3 raft::device_matrix_view<const float, int64_t, raft::row_major> X,
4 raft::device_vector_view<const int, int64_t> labels,
5 std::optional<raft::device_vector_view<float, int64_t>> silhouette_score_per_sample,
6 int64_t n_unique_labels,
7 int64_t batch_size,
8 cuvs::distance::DistanceType metric = cuvs::distance::DistanceType::L2Unexpanded);

Parameters

Name	Direction	Type	Description
`handle`	in	`raft::resources const&`	raft handle for managing expensive resources
`X`	in	`raft::device_matrix_view<const float, int64_t, raft::row_major>`	input matrix Data in row-major format (nRows x nCols)
`labels`	in	`raft::device_vector_view<const int, int64_t>`	the pointer to the array containing labels for every data sample (length: nRows)
`silhouette_score_per_sample`	out	`std::optional<raft::device_vector_view<float, int64_t>>`	optional array populated with the silhouette score for every sample (length: nRows)
`n_unique_labels`	in	`int64_t`	number of unique labels in the labels array
`batch_size`	in	`int64_t`	number of samples per batch
`metric`	in	`cuvs::distance::DistanceType`	the numerical value that maps to the type of distance metric to be used in the calculations Default: `cuvs::distance::DistanceType::L2Unexpanded`.

Returns

float

The silhouette score.

Additional overload: stats::silhouette_score

main function that returns the average silhouette score for a given set of data and its clusterings

1 double silhouette_score(
2 raft::resources const& handle,
3 raft::device_matrix_view<const double, int64_t, raft::row_major> X_in,
4 raft::device_vector_view<const int, int64_t> labels,
5 std::optional<raft::device_vector_view<double, int64_t>> silhouette_score_per_sample,
6 int64_t n_unique_labels,
7 cuvs::distance::DistanceType metric = cuvs::distance::DistanceType::L2Unexpanded);

Parameters

Name	Direction	Type	Description
`handle`	in	`raft::resources const&`	raft handle for managing expensive resources
`X_in`	in	`raft::device_matrix_view<const double, int64_t, raft::row_major>`	input matrix Data in row-major format (nRows x nCols)
`labels`	in	`raft::device_vector_view<const int, int64_t>`	the pointer to the array containing labels for every data sample (length: nRows)
`silhouette_score_per_sample`	out	`std::optional<raft::device_vector_view<double, int64_t>>`	optional array populated with the silhouette score for every sample (length: nRows)
`n_unique_labels`	in	`int64_t`	number of unique labels in the labels array
`metric`	in	`cuvs::distance::DistanceType`	the numerical value that maps to the type of distance metric to be used in the calculations Default: `cuvs::distance::DistanceType::L2Unexpanded`.

Returns

double

The silhouette score.

Additional overload: stats::silhouette_score_batched

function that returns the average silhouette score for a given set of data and its clusterings

1 double silhouette_score_batched(
2 raft::resources const& handle,
3 raft::device_matrix_view<const double, int64_t, raft::row_major> X,
4 raft::device_vector_view<const int, int64_t> labels,
5 std::optional<raft::device_vector_view<double, int64_t>> silhouette_score_per_sample,
6 int64_t n_unique_labels,
7 int64_t batch_size,
8 cuvs::distance::DistanceType metric = cuvs::distance::DistanceType::L2Unexpanded);

Parameters

Name	Direction	Type	Description
`handle`	in	`raft::resources const&`	raft handle for managing expensive resources
`X`	in	`raft::device_matrix_view<const double, int64_t, raft::row_major>`	input matrix Data in row-major format (nRows x nCols)
`labels`	in	`raft::device_vector_view<const int, int64_t>`	the pointer to the array containing labels for every data sample (length: nRows)
`silhouette_score_per_sample`	out	`std::optional<raft::device_vector_view<double, int64_t>>`	optional array populated with the silhouette score for every sample (length: nRows)
`n_unique_labels`	in	`int64_t`	number of unique labels in the labels array
`batch_size`	in	`int64_t`	number of samples per batch
`metric`	in	`cuvs::distance::DistanceType`	the numerical value that maps to the type of distance metric to be used in the calculations Default: `cuvs::distance::DistanceType::L2Unexpanded`.

Returns

double

The silhouette score.