A data stream is a continuously arriving sequence of data and clustering data streams requires additional considerations to traditional clustering. A stream is potentially unbounded, data points arrive online and each data point can be examined only once. This imposes limitations on available memory and processing time. Furthermore, streams can be noisy and the number of clusters in the data and their statistical properties can change over time. This paper presents an online, bio-inspired approach to clustering dynamic data streams. The proposed ant colony stream clustering (ACSC) algorithm is a density-based clustering algorithm, whereby clusters are identified as high-density areas of the feature space separated by low-density areas. ACSC identifies clusters as groups of micro-clusters. The tumbling window model is used to read a stream and rough clusters are incrementally formed during a single pass of a window. A stochastic method is employed to find these rough clusters, this is shown to significantly speeding up the algorithm with only a minor cost to performance, as compared to a deterministic approach. The rough clusters are then refined using a method inspired by the observed sorting behavior of ants. Ants pick-up and drop items based on the similarity with the surrounding items. Artificial ants sort clusters by probabilistically picking and dropping micro-clusters based on local density and local similarity. Clusters are summarized using their constituent micro-clusters and these summary statistics are stored offline. Experimental results show that the clustering quality of ACSC is scalable, robust to noise and favorable to leading ant clustering and stream-clustering algorithms. It also requires fewer parameters and less computational time.

Change is one of the biggest challenges in dynamic stream mining. From a data-mining perspective, adapting and tracking change is desirable in order to understand how and why change has occurred. Clustering, a form of unsupervised learning, can be used to identify the underlying patterns in a stream. Density-based clustering identifies clusters as areas of high density separated by areas of low density. This paper proposes a Multi-Density Stream Clustering (MDSC) algorithm to address these two problems; the multi-density problem and the problem of discovering and tracking changes in a dynamic stream. MDSC consists of two on-line components; discovered, labelled clusters and an outlier buffer. Incoming points are assigned to a live cluster or passed to the outlier buffer. New clusters are discovered in the buffer using an ant-inspired swarm intelligence approach. The newly discovered cluster is uniquely labelled and added to the set of live clusters. Processed data is subject to an ageing function and will disappear when it is no longer relevant. MDSC is shown to perform favourably to state-of-the-art peer stream-clustering algorithms on a range of real and synthetic data-streams. Experimental results suggest that MDSC can discover qualitatively useful patterns while being scalable and robust to noise.

Change in a data stream can occur at the concept level and at the feature level. Change at the feature level can occur if new, additional features appear in the stream or if the importance and relevance of a feature changes as the stream progresses. This type of change has not received as much attention as concept-level change. Furthermore, a lot of the methods proposed for clustering streams (density-based, graph-based, and grid-based) rely on some form of distance as a similarity metric and this is problematic in high-dimensional data where the curse of dimensionality renders distance measurements and any concept of ''density'' difficult. To address these two challenges we propose combining them and framing the problem as a feature selection problem, specifically a dynamic feature selection problem. We propose a dynamic feature mask for clustering high dimensional data streams. Redundant features are masked and clustering is performed along unmasked, relevant features. If a feature's perceived importance changes, the mask is updated accordingly; previously unimportant features are unmasked and features which lose relevance become masked. The proposed method is algorithm-independent and can be used with any of the existing density-based clustering algorithms which typically do not have a mechanism for dealing with feature drift and struggle with high-dimensional data. We evaluate the proposed method on four density-based clustering algorithms across four high-dimensional streams; two text streams and two image streams. In each case, the proposed dynamic feature mask improves clustering performance and reduces the processing time required by the underlying algorithm. Furthermore, change at the feature level can be observed and tracked.INDEX TERMS Data stream clustering, dynamic feature selection, feature drift, feature evolution, unsupervised feature selection.

Density based methods have been shown to be an effective approach for clustering non-stationary data streams. The number of clusters does not need to be known a priori and density methods are robust to noise and changes in the statistical properties of the data. However, most density approaches require sensitive, data dependent parameters. These parameters greatly affect the clustering performance and in a dynamic stream a good set of parameters at time t are not necessarily the best at time t+1. Furthermore, these parameters are global and so restrict the algorithm to finding clusters of the same density. In this paper, we propose a density based algorithm with adaptive parameters which are local to each discovered cluster. The algorithm, denoted Ant Colony Multi-Density Clustering (ACMDC), uses artificial ants to form nests in dense areas of the data. As the ants move between nests, their collective memory is stored in the form of pheromone trails. Clusters are identified as groups of similar nests. The proposed algorithm is evaluated across a number of synthetic data streams containing overlapping and embedded multi-density clusters. The performance of the algorithm is shown to be favourable to a leading density based stream-clustering algorithm despite requiring no tunable parameters.

In a dynamic stream there is an assumption that the underlying process generating the stream is non-stationary and that concepts within the stream will drift and change as the stream progresses. Concepts learned by a classification model are prone to change and non-adaptive models are likely to deteriorate and become ineffective over time. The challenge of recognising and reacting to change in a stream is compounded by the scarcity of labels problem. This refers to the very realistic situation in which the true class label of an incoming point is not immediately available (or might never be available) or in situations where manually annotating data points are prohibitively expensive. In a high-velocity stream, it is perhaps impossible to manually label every incoming point and pursue a fully supervised approach. In this article, we formally describe the types of change, which can occur in a data-stream and then catalogue the methods for dealing with change when there is limited access to labels. We present an overview of the most influential ideas in the field along with recent advancements and we highlight trends, research gaps, and future research directions.