Presentation is loading. Please wait.

Presentation is loading. Please wait.

Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Clustering data in an uncertain environment using an artificial.

Similar presentations


Presentation on theme: "Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Clustering data in an uncertain environment using an artificial."— Presentation transcript:

1 Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Clustering data in an uncertain environment using an artificial immune system Presenter : Cheng-Hui Chen Author : A.J. Graaff *, A.P. Engelbrecht PRL 2011

2 Intelligent Database Systems Lab N.Y.U.S.T. I. M. 2 Outlines Motivation Objectives Methodology Experiments Conclusions Comments

3 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Motivation 3 Lymphocytes network Artificial immune system , AIS Antigenic Antibody

4 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Objectives  An artificial immune models applied to cluster and proposes clustering performance measures in a non-stationary environment.  The investigates different data migration types and proposes a technique to generate artificial non-stationary data. 4

5 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Methodology 5 Lymphocyte Network Antigenic Compare three artificial immune models use to cluster non-stationary data  DWB  SMAIN  LNNAIS Artificial lymphocyte

6 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Methodology  Data migration types ─ Pattern migration A feature vector can migrate from one cluster to another in the data set. ─ Cluster migration Instead of selecting a ratio of feature vectors for migration, a fraction of the number of clusters in the data set is selected for migration. ─ Centroid migration All the clusters in a data set adapt the spatial position of their centroids. 6

7 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Methodology  Initialising B and determining A ─ DWB Bmax ─ LNNAIS Bmax( empty expands to Bmax over time) ─ SMAIN Binit 7  Calculating the affinity and stimulation level ─ DWB Euclidean distance and Initial radius of influence, ϕ min ─ LNNAIS Euclidean distance ─ SMAIN Euclidean distance  Determining the ALC networks ─ DWB Hybrid approach ─ LNNAIS Index based neighbourhood technique ─ SMAIN Network affinity threshold, NAT Proximity matrix  Adapting the ALCs in H ─ DWB If membership function value > threshold  Age is mature Mutation rate, ς ─ LNNAIS Mutation clone of the ALC in H ─ SMAIN If antigen affinity >NAT  Antigen is initialised as a cloned ALC  Determining the number of clusters ─ DWB K-means ─ LNNAI Index based neighbourhood technique ─ SMAIN Hierarchical agglomerative A: B:

8 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Methodology  The average compactness of the clusters.  The minimum separation between clusters is determined by inter min.  The validity index is based on the ratio of intra- clustering distance to the inter clustering distance. 8

9 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Methodology  The cluster quality, at time t  The mean value of measured clustering quality for a specific run, r  The collective mean quality 9

10 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Methodology  Generating artificial non-stationary data 10

11 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Experiments 11

12 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Pattern migration 12

13 Intelligent Database Systems Lab N.Y.U.S.T. I. M. 13

14 Intelligent Database Systems Lab N.Y.U.S.T. I. M. 14 Cluster migration

15 Intelligent Database Systems Lab N.Y.U.S.T. I. M. 15

16 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Centroid migration 16

17 Intelligent Database Systems Lab N.Y.U.S.T. I. M. 17

18 Intelligent Database Systems Lab N.Y.U.S.T. I. M. 18  A non-parametric Mann– Whitney U test with a 0.95 confidence interval. ─ H0: There is no difference in Q. ─ H1: There is a difference in Q.

19 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Conclusions  The proposed clustering performance measures were used for comparison between three network based AIS models for clustering of the generated artificial non-stationary data sets. 19 DWBSMAINLNNAI Cluster of qualityLowHighMid CostHighMidLow

20 Intelligent Database Systems Lab N.Y.U.S.T. I. M. Comments  Advantages ─ Context detailed  Drawback ─ …  Applications ─ Clustering in non-stationary environment. 20


Download ppt "Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Clustering data in an uncertain environment using an artificial."

Similar presentations


Ads by Google