DBSCAN

1. Package name: org.apache.spark.ml.clustering
2. Class name: DBSCAN
3. Method name: fitPredict
4. Input: training sample data (Dataset[_]). The following are mandatory fields.

   Parameter	Type	Default Value	Description
   featuresCol	Vector	features	Feature vector

5. Algorithm parameters

   def setMinPoints(value: Int): DBSCAN.this.type
   def setEpsilon(value: Double): DBSCAN.this.type
   def setSampleRate(value: Double): DBSCAN.this.type

   1. epsilon indicates the maximum distance two neighbors can be from one another while still belonging to the same cluster. Its value must be greater than 0.0.
   2. minPoints indicates the minimum number of neighbors of a given point. Its value must be greater than 1.
   3. sampleRate indicates the sampling rate of the input data. It is used to divide the space of the full input data based on the sampled data. The value range is (0.0, 1.0]. The default value is 1.0, indicating that the full input data is used by default.
      

      Code interface example:

      1 2 3 4

      val model = new DBSCAN() .setEpsilon(params.epsilon) .setMinPoints(params.minPoints) .setSampleRate(params.sampleRate)

6. Output: clustering result. The fields are as follows:

   Parameter	Type	Default Value	Description
   predictionCol	Int	prediction	Category of the sample. -1: noise sample 0: core sample 1: border sample
   labelCol	Int	label	Cluster ID of the sample. For noise samples, the cluster ID is -1 by default. For core/border samples, the cluster ID is greater than or equal to 0.

val dbscan = new DBSCAN()
      .setEpsilon(0.2)
      .setMinPoints(3)
      .setSampleRate(1.0)
val result = dbscan.fitPredict(trainData)