CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with - - PowerPoint PPT Presentation

▶
cb dbscan a novel clustering algorithm for adjacent
SMART_READER_LITE
LIVE PREVIEW

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with - - PowerPoint PPT Presentation

Sep 10, 2023 117 likes •457 views

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities Gashin Ghazizadeh Outline Introduction Multi-density Clustering CB-DBSCAN Results CB-DBSCAN: A Novel Clustering Algorithm for Adjacent

slide-1
SLIDE 1

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities

Gashin Ghazizadeh

slide-2
SLIDE 2

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Outline

  • Introduction
  • Multi-density Clustering
  • CB-DBSCAN
  • Results

2020-05-04 2

slide-3
SLIDE 3

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Introduction

  • Clustering
  • Machine Learning algorithms to find groups of similar data points in data
  • Density-based and center-based algorithms
  • DBSCAN is a clustering algorithm that tries to find clusters based on the

density of different regions of the data.

2020-05-04 3

slide-4
SLIDE 4

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

slide-5
SLIDE 5

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

slide-6
SLIDE 6

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

𝑫𝟐 𝑫𝟑

slide-7
SLIDE 7

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

𝑫𝟐 𝑫𝟑

slide-8
SLIDE 8

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

𝑫𝟐 𝑫𝟑

slide-9
SLIDE 9

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

Core point 𝑫𝟐 𝑫𝟑

slide-10
SLIDE 10

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

Core point 𝑫𝟐 𝑫𝟑

slide-11
SLIDE 11

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

Core point Border Point 𝑫𝟐 𝑫𝟑

slide-12
SLIDE 12

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

Core point Border Point 𝑫𝟐 𝑫𝟑

slide-13
SLIDE 13

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

Core point Noise Border Point 𝑫𝟐 𝑫𝟑

slide-14
SLIDE 14

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

Core point Noise Border Point 𝑫𝟐 𝑫𝟑

slide-15
SLIDE 15

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Multi-density Clustering

2020-05-04 4

Core point Noise Border Point Boundary Point 𝑫𝟐 𝑫𝟑

slide-16
SLIDE 16

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

CB-DBSCAN

  • Clustering with DBSCAN
  • Merging
  • Noise points
  • Parameter (Cdis and Ddis)

2020-05-04 5

slide-17
SLIDE 17

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Mini-Clusters

How to choose correct parameters?

2020-05-04 6

slide-18
SLIDE 18

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Mini-Clusters

How to choose correct parameters?

2020-05-04 6

slide-19
SLIDE 19

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Mini-Clusters

How to choose correct parameters?

2020-05-04 6

slide-20
SLIDE 20

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Mini-Clusters

How to choose correct parameters?

2020-05-04 6

slide-21
SLIDE 21

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Mini-Clusters

How to choose correct parameters?

2020-05-04 6

slide-22
SLIDE 22

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Mini-Clusters

How to choose correct parameters?

2020-05-04 6

slide-23
SLIDE 23

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Merging

2020-05-04 7

slide-24
SLIDE 24

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Merging

2020-05-04 7

slide-25
SLIDE 25

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Merging

2020-05-04 7

slide-26
SLIDE 26

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Merging

2020-05-04 7

slide-27
SLIDE 27

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Merging

2020-05-04 7

slide-28
SLIDE 28

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Merging (cont.)

2020-05-04 8

▪ No additional parameters (3*Ddis and Cdis/2) ▪ To combine boundary points and center points of clusters

slide-29
SLIDE 29

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Results

2020-05-04 9

DBSCAN CB-DBSCAN PACA-DBSCAN kAA-DBSCAN OPTICS Flame 0.89 0.99 0.98 0.98 0.91 Seeds 0.56 0.90 0.72 0.85 0.8 Path based 0.66 0.91 0.92 0.99 0.69 Breast 0.61 0.97 0.75 0.73 0.66 Wine 0.67 0.94 0.72 0.72 0.71

F-measure Comparison of Clustering Algorithms

slide-30
SLIDE 30

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

Results (cont.)

2020-05-04 10

Accuracy F-Measure Silhouette Coefficient DBSCAN 0.91 0.47 0.38 CB-DBSCAN 0.98 0.65 0.51

CB-DBSCAN and DBSCAN on Twitter data

slide-31
SLIDE 31

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al. 2020-05-04 11

T H A K Y O U N !

slide-32
SLIDE 32

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al. 2020-05-04 11

THA K Y O U N !

slide-33
SLIDE 33

CB-DBSCAN: A Novel Clustering Algorithm for Adjacent Clusters with Different Densities | Ghazizadeh et al.

References

  • Ester, M., Kriegel, H. P., Sander, J., & Xu, X. (1996, August). Density-based spatial clustering of

applications with noise. In Int. Conf. Knowledge Discovery and Data Mining (Vol. 240, p. 6).

  • Kim, J. H., Choi, J. H., Yoo, K. H., & Nasridinov, A. (2019). AA-DBSCAN: an approximate adaptive

DBSCAN for finding clusters with varying densities. The Journal of Supercomputing, 75(1), 142-169.

  • Jiang, H., Li, J., Yi, S., Wang, X., & Hu, X. (2011). A new hybrid method based on partitioning-based

DBSCAN and ant clustering. Expert Systems with Applications, 38(8), 9373-9381.

  • Ankerst, M., Breunig, M. M., Kriegel, H. P., & Sander, J. (1999). OPTICS: ordering points to identify

the clustering structure. ACM Sigmod record, 28(2), 49-60.

2020-05-04 12