Class RKOF

constructor

• new RKOF(
      k: number,
      h: number,
      alpha: number,
      kernel?:
          | "gaussian"
          | "epanechnikov"
          | "volcano"
          | { name: "gaussian" }
          | { name: "epanechnikov" }
          | { beta?: number; name: "volcano" }
          | ((arg0: number[]) => number),
  ): RKOF

  Parameters

  • k: number

    Number of neighborhoods
  • h: number

    Smoothing parameter
  • alpha: number

    Sensitivity parameter
  • Optionalkernel:
        | "gaussian"
        | "epanechnikov"
        | "volcano"
        | { name: "gaussian" }
        | { name: "epanechnikov" }
        | { beta?: number; name: "volcano" }
        | ((arg0: number[]) => number)

    Kernel name

  Returns RKOF

_alpha

_h

_k

_kernel

_s

_distance

• _distance(a: any, b: any): number

  Parameters

  • a: any
  • b: any

  Returns number

predict

• predict(datas: number[][]): number[]

  Returns anomaly degrees.

  Parameters

  • datas: number[][]

    Training data

  Returns number[]

  Predicted values