rhyde67 / cedas

CEDAS: Clustering of Evolving Data-streams into Arbitrary Shapes

CEDAS

CEDAS: Clustering of Evolving Data-streams into Arbitrary Shapes

Clustering of Evolving Data-streams into Arbitrary Shapes

This branch contains two versions of CEDAS for Matlab.

CEDAS.m

This is the full algorithm implementation and requires Matlab 2015b or later as it uses Matlab's built in 'graph' function.

The file CEDAS_IS_Demo.m creates a rando data stream and applies the CEDAS algorithm in an online manner, producing complete cluster results with each data sample. By using the Matlab 'graph' function the full graph structure is available and also displayed.

CEDAS_Pre2015b.m

This version does not use the Matlab 'graph' function included with Matlab 2015b and on. The code runs faster, but does not easily enable visualization of the graph structure.

The data file M-G_3Dx2.csv contains a data stream consisting of 2 Mackey-Glass data streams in 3D - 2 data co-ords and a time parameter.

Run_CEDAS_MG.m reads the Mackey-Glass data and applies the CEDAS algorithm on every sample. The micro-clusters are displayed as sphere's, coloured by their macro-clusters. The macro-clusters can be seen to marge and separate.

Other files

distinguishable_colours is not the work of the CEDAS author. For full details see the headers in the file. It is available for downlaod from Matlab's file exchange.

cedas's Issues

Data processing methods and parameter threshold and other related issues

Dear,
I am a student and interested in your work about the CEDAS, I want to get your KDDCup99 dataset and ask for the threshold of the Radius and MinimumClusterSize.Because I failed to reproduce your achievements may be used the wrong data preprocessing method of KDDCup99 dataset, I make many individual data points clustered on their own, so I can not get the right cluster quality.
Thanks for your help, I look forward to your reply.