ELKI

ELKI

Elki is a suite of algorithms used for data clustering that includes classical partitioning algorithms, EM-based probabilistic algorithms, density-based algorithms, and subspace clustering algorithms. It is also a data mining tool that includes the KNNDistancesSampler function used to estimate the ϵ value by calculating the distances of each point to its closest neighbor.From: Data Clustering [2018], Identification and characterisation of Facebook user profiles considering interaction aspects [2019]

Educational and Software Resources for Data Clustering

View Chapter

Purchase Book

Published in Charu C. Aggarwal, Chandan K. Reddy, Data Clustering, 2018

Charu C. Aggarwal, Chandan K. Reddy

Another comprehensive set of data clustering packages is provided in the ELKI [1] suite of algorithms that include many classical partitioning algorithms, EM-based probabilistic algorithms, density-based algorithms, and subspace clustering algorithms. The KDnuggets web site [42] also provides access to a significant number of open-source software sites for clustering and segmentation. This can be considered a meta-repository, in that it provides pointers to other relevant sites.

Identification and characterisation of Facebook user profiles considering interaction aspects

View Article

Journal Information

Published in Behaviour & Information Technology, 2019

Pedro Henrique B. Ruas, Alan D. Machado, Michelle C. Silva, Magali R. G. Meireles, Ana Maria P. Cardoso, Luis E. Zárate, Cristiane N. Nobre

We performed experiments (using the ELKI data mining tool and SPSS) for each pair of MinPts and ϵ values, with normalised data indexed by the tool. The values were normalised in the interval, and the indexation was made through a R* tree structure (Beckmann et al. 1990), which is recommended in Ester and Kriegel (1996). The noisy data found were treated according to the approach utilised in Moulavi et al. (2014). Therefore, all the experiments ignored noise when calculating evaluation metrics, for those represent inconsistent data regarding cluster evaluation.

Explore chapters and articles related to this topic

Educational and Software Resources for Data Clustering

Identification and characterisation of Facebook user profiles considering interaction aspects