On Partitional Clustering of Malware
View/ Open
Author
Cordeiro De Amorim, Renato
Komisarczuk, Peter
Attention
2299/16819
Abstract
In this paper we fully describe a novel clustering method for malware, from the transformation of data into a manipulable standardised data matrix, finding the number of clusters until the clustering itself including visualisation of the high-dimensional data. Our clustering method deals well with categorical data and clusters the behavioural data of 17,000 websites, acquired with Capture-HPC, in less than 2 minutes