Amineh Amini

On Density-Based Data Streams Clustering Algorithms: A Survey

Clustering data streams has drawn lots of attention in the last few years due to their ever-growing presence. Data streams put additional challenges on clustering such as limited time and memory and one pass clustering. Furthermore, discovering clusters with arbitrary shapes is very important in data stream applications. Data streams are infinite and evolving over time, and we do not have any knowledge about the number of clusters. In a data stream environment due to various factors, some noise appears occasionally. Density-based method is a remarkable class in clustering data streams, which has the ability to discover arbitrary shape clusters and to detect noise. Furthermore, it does not need the number of clusters in advance. Due to data stream characteristics, the traditional density-based clustering is not applicable. Recently, a lot of density-based clustering algorithms are extended for data streams. The main idea in these algorithms is using density-based methods in the clustering process and at the same time overcoming the constraints, which are put out by data stream’s nature. The purpose of this paper is to shed light on some algorithms in the literature on density-based clustering over data streams. We not only summarize the main density-based clustering algorithms on data streams, discuss their uniqueness and limitations, but also explain how they address the challenges in clustering data streams. Moreover, we investigate the evaluation metrics used in validating cluster quality and measuring algorithms’ performance. It is hoped that this survey will serve as a steppingstone for researchers studying data streams clustering, particularly density-based algorithms.

A study of density-grid based clustering algorithms on data streams

Clustering data streams attracted many researchers since the applications that generate data streams have become more popular. Several clustering algorithms have been introduced for data streams based on distance which are incompetent to find clusters of arbitrary shapes and cannot handle the outliers. Density-based clustering algorithms are remarkable not only to find arbitrarily shaped clusters but also to deal with noise in data. In density-based clustering algorithms, dense areas of objects in the data space are considered as clusters which are segregated by low-density area. Another group of the clustering methods for data streams is grid-based clustering where the data space is quantized into finite number of cells which form the grid structure and perform clustering on the grids. Grid-based clustering maps the infinite number of data records in data streams to finite numbers of grids. In this paper we review the grid based clustering algorithms that use density-based algorithms or density concept for the clustering. We called them density-grid clustering algorithms. We explore the algorithms in details and the merits and limitations of them. The algorithms are also summarized in a table based on the important features. Besides that, we discuss about how well the algorithms address the challenging issues in the clustering data streams.

MuDi-Stream

Density-based method has emerged as a worthwhile class for clustering data streams. Recently, a number of density-based algorithms have been developed for clustering data streams. However, existing density-based data stream clustering algorithms are not without problem. There is a dramatic decrease in the quality of clustering when there is a range in density of data. In this paper, a new method, called the MuDi-Stream, is developed. It is an online-offline algorithm with four main components. In the online phase, it keeps summary information about evolving multi-density data stream in the form of core mini-clusters. The offline phase generates the final clusters using an adapted density-based clustering algorithm. The grid-based method is used as an outlier buffer to handle both noises and multi-density data and yet is used to reduce the merging time of clustering. The algorithm is evaluated on various synthetic and real-world datasets using different quality metrics and further, scalability results are compared. The experimental results show that the proposed method in this study improves clustering quality in multi-density environments.

A Multi Density-Based Clustering Algorithm for Data Stream with Noise

Density-based clustering can detect arbitrary shape clusters, handle outliers and do not need the number of clusters in advance. However, they cannot work properly in multi density environments. The existing multi density clustering algorithms have some problems in order to be applicable for data streams such as the need of whole data to perform clustering, two-pass clustering and high execution time. Data stream arrives continuously and they have to be processed in limited time and memory. Therefore, we need an algorithm to cluster data stream with different densities as well as to overcome the challenges in clustering data streams. In this paper, we introduce a Multi-Density clustering algorithm for data stream called MuDi-Stream. MuDi-Stream is an online-offline clustering algorithm, in which the online phase forms core-mini-clusters using a new proposed core distance and offline phase clusters the core-mini-clusters based on a density-based method. The new core distance called mini core distance is calculated based on the number of neighboring data points around the core. Therefore, the algorithm has different core distances for different clusters that leads to cover multi density environments. A novel pruning strategy is also used to filter out the real data from the noise by mapping the outliers in the grid. The grid structure keeps the neighbors of the data point to determine mini-core distance and remove noise effectively. Our performance study over synthetic data sets demonstrates effectiveness of our method.

A Comparative Study of Density-based Clustering Algorithms on Data Streams: Micro-clustering Approaches

Clustering data streams is a challenging problem in mining data streams. Data streams need to be read by a clustering algorithm in a single pass with limited time, and memory whereas they may change over time. Different clustering algorithms have been developed for data streams. Density-based algorithms are a remarkable group in clustering data that can find arbitrary shape clusters, and handle the outliers as well. In recent years, density-based clustering algorithms are adopted for data streams. However, in clustering data streams, it is impossible to record all data streams. Micro-clustering is a summarization method used to record synopsis information about data streams. Various algorithms apply micro-clustering methods for clustering data streams. In this paper, we will concentrate on the density-based clustering algorithms that use micro-clustering methods for clustering and we refer them as density-micro clustering algorithms. We review the algorithms in details and compare them based on different characteristics.

DMM-Stream: A Density Mini-Micro Clustering Algorithm for Evolving Data Streams

Clustering real-time stream data is an important and challenging problem. The existing algorithms have not considered the distribution of data inside micro cluster, specifically when data points are non uniformly distributed inside micro cluster. In this situation, a large radius of micro cluster has to be considered which leads to lower quality. In this paper, we present a density-based clustering algorithm, DMM-Stream, for evolving data streams. It is an online-offline algorithm which considers the distribution of data inside micro cluster. In DMM-Stream, we introduce mini-micro cluster for keeping summary information of data points inside micro cluster. In our method, based on the distribution of the dense areas inside the micro cluster at least one representative point, either micro cluster itself or its mini-micro clusters’ centers, are sent to the offline phase. By choosing a proper mini-micro and micro center, we increase cluster quality while maintaining the time complexity. A pruning strategy is also used to filter out the real data from noise by introducing dense and sparse mini-micro and micro cluster. Our performance study over real and synthetic data sets demonstrates effectiveness of our method.

A Fast Density-Based Clustering Algorithm for Real-Time Internet of Things Stream

Data streams are continuously generated over time from Internet of Things (IoT) devices. The faster all of this data is analyzed, its hidden trends and patterns discovered, and new strategies created, the faster action can be taken, creating greater value for organizations. Density-based method is a prominent class in clustering data streams. It has the ability to detect arbitrary shape clusters, to handle outlier, and it does not need the number of clusters in advance. Therefore, density-based clustering algorithm is a proper choice for clustering IoT streams. Recently, several density-based algorithms have been proposed for clustering data streams. However, density-based clustering in limited time is still a challenging issue. In this paper, we propose a density-based clustering algorithm for IoT streams. The method has fast processing time to be applicable in real-time application of IoT devices. Experimental results show that the proposed approach obtains high quality results with low computation time on real and synthetic datasets.

Failure Recovery of Composite Semantic Services using Expiration Times

Composite services are examples of volatile processes, which are prone to failures due to several problems that may occur during the executions. The recovery of their failure at execution time must be done efficiently to survive the system from deviation of its quality of service. Replacing a failed service with another similar service is not always reliable. Substituting a subgraph of the directed graph which represents the composite service with another sequence of services shows a major step in increasing the likelihood of the system’s failure recovery. In this paper, we propose the use of expiration times of provided services to lower the time complexity of subdigraph identifications as well as other steps of a recovery approach. We evaluated our work, which shows a significant improvement to the similar approaches.

On Density-Based Clustering Algorithms over Evolving Data Streams: A Summarization Paradigm

Clustering is one of the prominent classes in the mining data streams. Among various clustering algorithms that have been developed, density-based method has the ability to discover arbitrary shape clusters, and to detect the outliers. Recently, various algorithms adopted density-based methods for clustering data streams. In this paper, we look into three remarkable algorithms in two groups of micro-clustering and grid-based including DenStream, D-Stream, and MR-Stream. We compare the algorithms based on evaluating algorithm performance and clustering quality metrics.

Anomaly detection and prediction of sensors faults in a refinery using data mining techniques and fuzzy logic

Like all manufacturing companies, refineries use many sensors to monitor and control the process of refining, therefore it is very crucial to detect any sensor faults or anomalies as early as possible, and to be able to replace or repair a sensor well in advance of any fault. Objective of this paper is to present a method for detecting anomalies in a sensor data, as well as to predict next occurance of a sensor failure. Data mining techniques to detect anomaly in sensor data and predict the occurrence of next faulty event were introduced. For anomaly detection, this research used MATLAB’s fuzzy logic toolbox tools to find clusters which uses subtractive fuzzy clustering algorithm and generates a model, a Sugeno-type fuzzy inference system. The same toolbox was used to evaluate the model with a promising result. To predict sensor fault, the original time series were used to create a new ‘derived time series’. Two prediction models known as ‘auto regressive integrated moving average’ and ‘autoregressive tree models’ were used against the new time series to predict next occurrence of sensor failure. The results of these models were compared. The model developed and introduced in this paper serves as an additional tool, which helps not only engineers and operators of oil refineries, but also other engineers of other disciplines to work more efficiently.