Options
All
  • Public
  • Public/Protected
  • All
Menu

Module cuml/src

Index

Namespaces

Enumerations

Classes

Type aliases

Functions

Type aliases

Numeric

Numeric: Integral | Float32

outputType

outputType: "dataframe" | "series" | "devicebuffer"

Functions

trustworthiness

  • trustworthiness<T>(features: (undefined | null | T)[], embedded: (undefined | null | T)[], nFeatures: number, nComponents?: number, nNeighbors?: number, batch_size?: number): number

  • Expresses to what extent the local structure is retained in embedding. The score is defined in the range [0, 1].

    Type parameters

    • T: number | bigint

    Parameters

    • features: (undefined | null | T)[]

      original high dimensional dataset

    • embedded: (undefined | null | T)[]

      low dimesional embedding

    • nFeatures: number

      Number of features in features

    • nComponents: number = 2

      Number of features in embedded

    • nNeighbors: number = 5

      Number of neighbors considered

    • batch_size: number = 512

      It sets the number of samples that will be included in each batch

    Returns number

    Trustworthiness of the low-dimensional embedding

trustworthinessDataFrame

  • trustworthinessDataFrame<T, R, K>(features: DataFrame<{ [ P in string]: T }>, embedded: DataFrame<{}>, nNeighbors?: number, batch_size?: number): number

  • Expresses to what extent the local structure is retained in embedding. The score is defined in the range [0, 1].

    Type parameters

    Parameters

    • features: DataFrame<{ [ P in string]: T }>

      original high dimensional dataset

    • embedded: DataFrame<{}>

      low dimesional embedding

    • nNeighbors: number = 5

      Number of neighbors considered

    • batch_size: number = 512

      It sets the number of samples that will be included in each batch

    Returns number

    Trustworthiness of the low-dimensional embedding

trustworthinessSeries

  • trustworthinessSeries<T, R>(features: Series<T>, embedded: Series<R>, nFeatures: number, nComponents?: number, nNeighbors?: number, batch_size?: number): number

  • Expresses to what extent the local structure is retained in embedding. The score is defined in the range [0, 1].

    Type parameters

    Parameters

    • features: Series<T>

      original high dimensional dataset

    • embedded: Series<R>

      low dimesional embedding

    • nFeatures: number

      Number of features in features

    • nComponents: number = 2

      Number of features in embedded

    • nNeighbors: number = 5

      Number of neighbors considered

    • batch_size: number = 512

      It sets the number of samples that will be included in each batch

    Returns number

    Trustworthiness of the low-dimensional embedding