Skip to main content

DBSCAN

CLUSTER_STATS​

cluster_stats(Bag of Vector b, Number beta) -> Vector of Number
STABLE

Calculates Precision, recall, beta F-measure and Rand-index for the labelled

and classified data d.

  • d Bag of vector where each vector contains at least [label, cluster_id]
  • beta beta for F-measure set to 1 to do F1-measre The following table can be used for reference on how true positives(tp), false negatives (fn), true negatives (tn) and false positives(fp) is defined:
same clusterdifferent clusters
Same classTPFN
Different classFPTN

The precision, recall F-measure and rand index is then calculated using the following formulas:

  • Precision formula: tp/(tp+fp)
  • Recall formula: tp/(tp+fn)
  • F-measure formula: (beta^2 +1)*precision*recall/(beta*precision+recall)
  • Rand index formula: (tp+tn)/(tp+fp+tn+fn)

returns: a vector of number where each index has the following metric:

  1. precision
  2. recall
  3. Fbeta-measure
  4. Rand index

DBSCAN:GENERATE​

dbscan:generate(Charstring name) -> Object
STABLE

No description.

DBSCAN:SAVE_MODEL​

dbscan:save_model(Charstring instance, Charstring model, Charstring file)
STABLE

No description.

dbscan:save_model(Charstring instance, Charstring model)
STABLE

No description.

FALSE_NEGATIVE​

false_negative(Bag of Vector b) -> Number
STABLE

Number of item pairs that are in different cluster and belong to different

classes

FALSE_POSITIVE​

false_positive(Bag of Vector b) -> Number
STABLE

Number of items pairs that are in the same cluster but belong to different

classes

GET_DBSCAN_STRING​

get_dbscan_string(Charstring name) -> Charstring
STABLE

Generate functions for a dbscan instance named name This will create

dbscan functions prefixed with <name>_ Look at the topic dbscan after generating a table.

TRUE_NEGATIVE​

true_negative(Bag of Vector b) -> Number
STABLE

Number of item pair that are in different clusters and blong to different

classes

TRUE_POSITIVE​

true_positive(Bag of Vector b) -> Number
STABLE

Number of item pairs that are in the same cluster and belong to the same

class