Skip to main content

Notice: this Wiki will be going read only early in 2024 and edits will no longer be possible. Please see: https://gitlab.eclipse.org/eclipsefdn/helpdesk/-/wikis/Wiki-shutdown-plan for the plan.

Jump to: navigation, search

Difference between revisions of "Anomaly Detection Tool Test Cases"

(Test Case 2)
(Test Case 3 LOF Test case)
Line 14: Line 14:
 
* No. of data points: 1000
 
* No. of data points: 1000
 
* Data points are easily separable into 2 clusters, however anomalies are not easy to find) as depicted in Fig.3. The point marked with an arrow can be a potential anomaly, as we consider its local density.
 
* Data points are easily separable into 2 clusters, however anomalies are not easy to find) as depicted in Fig.3. The point marked with an arrow can be a potential anomaly, as we consider its local density.
[[File:lof_j_p_g.pdf]]
+
[[File:lof.jpg]]
  
 
== Test Case 4 100 data points ==
 
== Test Case 4 100 data points ==

Revision as of 19:06, 26 February 2015

Test Case 1

  • No. of clusters: 2
  • No. of Data points: 1000
  • Data points are easily separable into 2 clusters as depicted in Fig.1. Kmeans clustering algorithm can easily identify two distinct clusters in this case.

Test Case 2

  • No. of clusters: 2
  • No. of data points: 1000
  • Data points are not easily separable into 2 clusters as depicted in Fig.2.

Single cluster.jpg

Test Case 3 LOF Test case

  • No. of clusters: 3
  • No. of data points: 1000
  • Data points are easily separable into 2 clusters, however anomalies are not easy to find) as depicted in Fig.3. The point marked with an arrow can be a potential anomaly, as we consider its local density.

Lof.jpg

Test Case 4 100 data points

  • No. of clusters: 2
  • No. of data points: 100
  • Data points are easily separable into 2 clusters as depicted in Fig.4.

Test Case 5 100 data points

  • No. of clusters: 2
  • No. of data points: 100
  • Data points are not easily separable into 2 clusters as depicted in Fig.5.

Test Case 6 LOF Test case: 100 data points

  • No. of clusters: 3
  • No. of data points: 100
  • Data points are easily separable into 2 clusters, however anomalies are not easy to find) as depicted in Fig.6. The point marked with an arrow can be a potential anomaly, as we consider its local density.

Back to the top