Multi-view clustering using mixture models in subspace projections

Abstract

Detecting multiple clustering solutions is an emerging research field. While data is often multi-faceted in its very nature, traditional clustering methods are restricted to find just a single grouping. To overcome this limitation, methods aiming at the detection of alternative and multiple clustering solutions have been proposed. In this work, we present a Bayesian framework to tackle the problem of multi-view clustering. We provide multiple generalizations of the data by using multiple mixture models. Each mixture describes a specific view on the data by using a mixture of Beta distributions in subspace projections. Since a mixture summarizes the clusters located in similar subspace projections, each view highlights specific aspects of the data. In addition, our model handles overlapping views, where the mixture components compete against each other in the data generation process. For efficiently learning the distributions, we propose the algorithm MVGen that exploits the ICM principle and uses Bayesian model selection to trade-off the cluster model's complexity against its goodness of fit. With experiments on various real-world data sets, we demonstrate the high potential of MVGen to detect multiple, overlapping clustering views in subspace projections of the data.

SMVC

“…That is, starting with low weights, we linearly increase the values wi,j until they reach the user specified scores. For initializing our method, we exploit the same principle as described in [21]. The random variable C/q3 is initialized randomly based on its prior distribution.…”

“…[28,23,22,21]). Here, the inevitable connection of multi-view clustering and subspace clustering has been observed first [28,22,21], which later also influenced sequentially working approaches like [15]. Subspace clustering assumes each cluster to have an individual set of relevant data attributes, which corresponds well with the motivation of multi-view clustering that different views on the data (i.e.…”

Stochastic subspace search for top-k multi-view clustering

“…According to [11], they can briefly be categorized into methods operating on the original (full-dimensional) dataspace [6], methods performing space transformations [5,14], and methods analyzing (axis-parallel) subspace projections [9,8]. In this work-in-progress paper, we describe a novel method belonging to the last category.…”

“…Similar to [8], we also allow views to overlap, i.e. individual dimensions might belong to multiple views.…”

“…individual dimensions might belong to multiple views. While [8] proposes an extension of mixture models where the number of clusters needs to be specified for each view, we exploit the paradigm of kernel density estimation allowing us to automatically determine the number of clusters per view.…”