1 January 2020 in Theoretical Computer Science
30 August 2017 in Advances in Intelligent Systems and Computing
1 January 2022 in SSRN Electronic Journal
Clustering with high-dimensional data is a well-known challenge faced by many data scientists, especially when dealing with distance based methods that are often affected by increasing dimensionality: many dimensions may be irrelevant and can mask existing clusters in noisy data. Therefore, Subspace Clustering is used to find clusters within different subspaces of dimensions and its soft category of algorithms assigns a weight to each feature in different clusters without transforming or reducing its feature space. In this work, we have investigated whether or not Soft Subspace Clustering (SSC), by means of the Fuzzy Cluster Validity Indices (FCVIs), with the appropriate parameters, performs better than Fuzzy c-Means when dealing with high-dimensional data. Our main result is a novel protocol for evaluating FCVIs performances concerning SSC. To provide firm conclusions about our proposal, we compared the performance of four clustering algorithms over thirty-seven high-dimensional datasets validated by nineteen FCVIS.