Data Stream Mining Research Articles

A Survey on Multi-Label Data Stream Classification

Nowadays, many real-world applications of our daily life generate massive volume of streaming data at a higher speed than ever before, to name a few, Web clicking data streams, sensor network data and credit transaction streams. Contrary to traditional data mining using static datasets, there are several challenges for data stream mining, for instance, finite memory, one-pass and timely reaction. In this survey, we provide a comprehensive review of existing multi-label streams mining algorithms and categorize these methods based on different perspectives, which mainly focus on the multi-label data stream classification. We first briefly summarize existing multi-label and data stream classification algorithms and discuss their merits and demerits. Secondly, we identify mining constraints on classification for multi-label streaming data, and present a comprehensive study in algorithms for multi-label data stream classification. Finally, several challenges and open issues in multi-label data stream classification are discussed, which are worthwhile to be pursued by the researchers in the future.

Mining big data streams using business analytics tools: a bird's eye view on MOA and SAMOA

Big data evolves as the prominent field in modern computing era. Big data analytics and its impact on extracting business intelligence is becoming indispensable for plethora of applications. The non-proprietary software revolution paved the way for illustrious evolution of tools like Weka, rapid miner, orange and R. Traditional data mining techniques hardly adapts to the requirements of rapid data analysis. The data stream processing algorithms that handle multitude of data endow with greater challenge in real time. Big data mining requires further improvisation in traditional tools to address the challenges of massive data processing. This paper highlights the importance of data stream mining and explores two important open source frameworks, namely massive online analysis (MOA) and scalable advanced massive online analysis (SAMOA). The implications of both the tools augurs well for further deliberations in big data research community. Business information system (BIS) models can reach unprecedented heights with the proliferation of these business analytics tools.

Two-Stage Cost-Sensitive Learning for Data Streams With Concept Drift and Class Imbalance

Most methods for classifying data streams operate under the hypothesis that the distribution of classes is balanced. Unfortunately, the phenomenon of class imbalance widely exists in many real-world applications. In addition, the underlying concept of data stream may change in a certain way over time, and attacks increase the difficulty of data stream mining. Motivated by this challenge, a Two-Stage Cost-Sensitive (TSCS) classification is proposed for addressing the class imbalance issue in non-stationary data streams. We propose a novel two-stage cost-sensitive framework for data stream classification by utilizing cost information in both feature selection stage and classification stage. Moreover, a window adaptation and drift detection mechanism, which guarantees that an ensemble can adapt promptly to concept drift, is embedded in our method. Our algorithm is compared with competitive algorithms on different kinds of datasets. The result demonstrates that TSCS obtains significant improvement in terms of class imbalance data stream metrics.

GEP-based classifier with drift detection for mining imbalanced data streams

Abstract Mining data streams require to cope with time, data size and possible concept drift constraints. Even more challenging is the case where, apart from the above, one has to deal with imbalanced data. Mining non stationary and imbalanced data streams is a relatively new area of research. In this paper, we propose the Gene Expression Programming (GEP) classifier with drift detection and data reuse for mining imbalanced data streams. GEP is used to evolve a complex expression tree returning predictions. Drift detector role is to signal the occurrence of drift which triggers inducing a new learner. Data reuse mechanism allows for improving the balance between minority and majority instances in a subset of data used for evolving the learner. The proposed approach is validated experimentally. The experiment results confirm that our classifier produces high-quality predictions.

An Adaptive Slide Window Security Method for Transaction Updation in Data Stream Mining

Data steam mining has gained large interest in current research domain. Where various information’s are retrieved based on the content of the context, the accuracy of the input stream with respect to its privacy is a major challenge. Windowing technique is used an effective approach in providing security measure in data stream mining. The recent develop windowing approach operates using sliding window, where anonymity is focused by different processing rules. The linear search sliding window has a constraint of search overhead and loss of generality under distributed information. In this paper, a new adaptive window approach for privacy coding in data stream mining is proposed. This presented approach is developed with the concern of minimize the search overhead and accuracy in search mining performance using adaptive window monitoring.

Online Medical Data Stream Mining Based on Adaptive Neuro-Fuzzy Approaches

Online Medical Data Stream Mining Based on Adaptive Neuro-Fuzzy Approaches

Multi-Label Punitive kNN with Self-Adjusting Memory for Drifting Data Streams

In multi-label learning, data may simultaneously belong to more than one class. When multi-label data arrives as a stream, the challenges associated with multi-label learning are joined by those of data stream mining, including the need for algorithms that are fast and flexible, able to match both the speed and evolving nature of the stream. This article presents a punitive k nearest neighbors algorithm with a self-adjusting memory (MLSAMPkNN) for multi-label, drifting data streams. The memory adjusts in size to contain only the current concept and a novel punitive system identifies and penalizes errant data examples early, removing them from the window. By retaining and using only data that are both current and beneficial, MLSAMPkNN is able to adapt quickly and efficiently to changes within the data stream while still maintaining a low computational complexity. Additionally, the punitive removal mechanism offers increased robustness to various data-level difficulties present in data streams, such as class imbalance and noise. The experimental study compares the proposal to 24 algorithms using 30 real-world and 15 artificial multi-label data streams on six multi-label metrics, evaluation time, and memory consumption. The superior performance of the proposed method is validated through non-parametric statistical analysis, proving both high accuracy and low time complexity. MLSAMPkNN is a versatile classifier, capable of returning excellent performance in diverse stream scenarios.

Round table: Online anti-vaccination movements: the role of social media in public health communications

Abstract Digital health has revolutionised healthcare, with implications for understanding public reaction to health emergencies and interventions. Social media provides a space where like-minded people can share interests and concerns in real-time, regardless of their location. This can be a force for good, as platforms like Twitter can spread correct information about outbreaks, for example in the 2009 swine flu pandemic. However, social media can also disseminate incorrect information or deliberately spread misinformation leading to adverse public health sentiment and outcomes. The current issues around trust in vaccines is the best-known example. Vaccine hesitancy, traditionally linked to issues of trust, misinformation and prior beliefs, has been increasingly fueled by influential groups on social media and the Internet. Ultimately, anti-vaccination movements have the potential to lead to outbreaks of vaccine-preventable diseases, especially if refusal is concentrated locally, creating vulnerable populations. For example, 2018-19 saw a large increase in incidence of measles in the US and Europe (where cases tripled from 2017), two regions where the disease was already or almost eliminated. In 2019, the World Health Organisation listed anti-vaccination movements as one of the top 10 threats to global public health. HPV vaccination is another example of the impact of anti-vaccination movements. As viral videos originating on YouTube spread across social networks, uptake has tumbled in a number of countries, with Japan, Denmark, Colombia and Ireland being badly hit. In Japan, the government came under sufficient pressure that they de-recommended HPV vaccine, seeing an 80% uptake rate fall below 1% in 2014. There have been reports of successful interventions by national governments. A recent campaign run by the HPV Alliance (a coalition of some 35 private companies, charities and public institutions) in Ireland has seen rates below 40% back up to a national average of 75%. A combination of hard-hitting personal testimonials, social media and traditional media promoted the HPV vaccine. Despite this, systematic engagement and supranational strategies are still in the early stages of being formulated. As misleading information spread through social media and digital networks has undesirable impact on attitudes to vaccination (and uptake rates), urgent actions are required. Analysis and visualisation techniques mining data streams from social media platforms, such as Twitter, Youtube enable real-time understanding of vaccine sentiments and information flows. Through identification of key influencers and flashpoints in articles about vaccination going viral, targeted public health responses could be developed. This roundtable discussion will showcase different ways in which media and social networks, accessible in real-time provide an opportunity for detecting a change in public confidence in vaccines, for identifying users and rumors and for assessing potential impact in order to know how to best respond. Key messages Social media has significantly enhanced our understanding of anti-vaccination movements and potential impact on public health attitudes and behaviors regarding vaccination. Innovative methods of analysing social media data, from digital health, data science and computer science, have an important role in developing health promotions to counter anti-vaccination movements.

A stratified reservoir sampling algorithm in streams and large datasets

In data stream mining, a stream is a dataset of unknown size with continuously incoming elements, which is typically large enough so that a computer processing it does not have enough memory to hold it in its entirety and each element can be read only once and only in order. Classical sampling methods such as simple random sampling (SRS), stratified sampling and cluster sampling cannot be used on the stream data since the entire set is not available all at once and data cannot be reread. Vitter’s (1985) Algorithm R is a reservoir sampling method which can be used to select an SRS from a data stream. In this article, we propose Algorithm SR which extends Algorithm R to a stratified reservoir sampling method with optimal allocation. We prove that the proposed method is asymptotically equivalent to classical stratified random sampling with optimal allocation. Implementation results show that the proposed method is efficient and can outperform Algorithm R.

Kappa Updated Ensemble for drifting data stream mining

Learning from data streams in the presence of concept drift is among the biggest challenges of contemporary machine learning. Algorithms designed for such scenarios must take into an account the potentially unbounded size of data, its constantly changing nature, and the requirement for real-time processing. Ensemble approaches for data stream mining have gained significant popularity, due to their high predictive capabilities and effective mechanisms for alleviating concept drift. In this paper, we propose a new ensemble method named Kappa Updated Ensemble (KUE). It is a combination of online and block-based ensemble approaches that uses Kappa statistic for dynamic weighting and selection of base classifiers. In order to achieve a higher diversity among base learners, each of them is trained using a different subset of features and updated with new instances with given probability following a Poisson distribution. Furthermore, we update the ensemble with new classifiers only when they contribute positively to the improvement of the quality of the ensemble. Finally, each base classifier in KUE is capable of abstaining itself for taking a part in voting, thus increasing the overall robustness of KUE. An extensive experimental study shows that KUE is capable of outperforming state-of-the-art ensembles on standard and imbalanced drifting data streams while having a low computational complexity. Moreover, we analyze the use of Kappa versus accuracy to drive the criterion to select and update the classifiers, the contribution of the abstaining mechanism, the contribution of the diversification of classifiers, and the contribution of the hybrid architecture to update the classifiers in an online manner.

SCLUSTREAM: AN EFFICIENT ALGORITHM FOR TRACKING CLUSTERS OVER SLIDING WINDOW IN BIG DATA STREAMING

Mining in data streams has been a hot research topic in the recent time. A main challenge in data stream mining lies in extracting knowledge in real time from a massive, dynamic data stream in only a single scan. Data stream clustering presents an important role in data stream processing. This paper proposes SCluStream an algorithm for tracking clusters over a sliding window to handle such challenges. The algorithm is an enhancement over CluStream which does not involve this sliding window concept. In the sliding window model, only the most recent data is used while the old data is eliminated, which allows for faster execution. A better clustering technique is also involved which managed to contribute to accuracy enhancement. The proposed algorithm has been tested on a dataset for Intrusion detection and the results showed that comparing SCluStream to CluStream has proven that the former algorithm is more efficient for online clusters generation for big data streaming in regard of the accuracy as well as the utilized time and memory resources.

A regression-based algorithm for frequent itemsets mining

PurposeFrequent itemset mining (FIM) is a basic topic in data mining. Most FIM methods build itemset database containing all possible itemsets, and use predefined thresholds to determine whether an itemset is frequent. However, the algorithm has some deficiencies. It is more fit for discrete data rather than ordinal/continuous data, which may result in computational redundancy, and some of the results are difficult to be interpreted. The purpose of this paper is to shed light on this gap by proposing a new data mining method.Design/methodology/approachRegression pattern (RP) model will be introduced, in which the regression model and FIM method will be combined to solve the existing problems. Using a survey data of computer technology and software professional qualification examination, the multiple linear regression model is selected to mine associations between items.FindingsSome interesting associations mined by the proposed algorithm and the results show that the proposed method can be applied in ordinal/continuous data mining area. The experiment of RP model shows that, compared to FIM, the computational redundancy decreased and the results contain more information.Research limitations/implicationsThe proposed algorithm is designed for ordinal/continuous data and is expected to provide inspiration for data stream mining and unstructured data mining.Practical implicationsCompared to FIM, which mines associations between discrete items, RP model could mine associations between ordinal/continuous data sets. Importantly, RP model performs well in saving computational resource and mining meaningful associations.Originality/valueThe proposed algorithms provide a novelty view to define and mine association.

A Comprehensive Survey on Cloud Data Mining (CDM) Frameworks and Algorithms

Data mining is used for finding meaningful information out of a vast expanse of data. With the advent of Big Data concept, data mining has come to much more prominence. Discovering knowledge out of a gigantic volume of data efficiently is a major concern as the resources are limited. Cloud computing plays a major role in such a situation. Cloud data mining fuses the applicability of classical data mining with the promises of cloud computing. This allows it to perform knowledge discovery out of huge volumes of data with efficiency. This article presents the existing frameworks, services, platforms, and algorithms for cloud data mining. The frameworks and platforms are compared among each other based on similarity, data mining task support, parallelism, distribution, streaming data processing support, fault tolerance, security, memory types, storage systems, and others. Similarly, the algorithms are grouped on the basis of parallelism type, scalability, streaming data mining support, and types of data managed. We have also provided taxonomies on the basis of data mining techniques such as clustering, classification, and association rule mining. We also have attempted to discuss and identify the major applications of cloud data mining. The various taxonomies for cloud data mining frameworks, platforms, and algorithms have been identified. This article aims at gaining better insight into the present research realm and directing the future research toward efficient cloud data mining in future cloud systems.

An Insight on Social Media Stream Mining

Data stream mining is one of the realms gaining upper hand over traditional data mining methods. Transfinite volumes of data termed as Data Streams are often generated by Internet traffic, Communication networks, On-line bank or ATM transactions etc. The streams are dynamic and ever-shifting and need to be analysed online as they are obtained. Social media is one of the notable sources of such data streams. While social media streaming has received a lot of attention over the past decade, the ever-expanding streams of data presents huge challenges for learning and maintaining control. Dealing with billions of user’s data measured in pet bytes is a demanding task in itself. It is indeed a challenge to mine such dynamic data from social networks in an uninterrupted and competent way. This paper is purposed to introduce social data streams and the mining techniques involved in processing them. We analyse the most recent trends in social media data stream mining to translate to the detailed study of the matter. We also review innovative implementations of social media stream mining that are currently prevalent.

Handling Concept Drift in Data Stream Classification.

Data Streams are having huge volume and it can-not be stored permanently in the memory for processing. In this paper we would be mainly focusing on issues in data stream, the major factors which are affecting the accuracy of classifier like imbalance class and Concept Drift. The drift in Data Stream mining refers to the change in data. Such as Class imbalance problem notifies that the samples are in the classes are not equal. In our research work we are trying to identify the change (Drift) in data, we are trying to detect Imbalance class and noise from changed data. And According to the type of drift we are applying the algorithms and trying to make the stream more balance and noise free to improve classifier’s accuracy.

An outlier detection approach in large-scale data stream using rough set

Outlier detection has become an important research area in the field of stream data mining due to its vast applications. In the literature, many methods have been proposed, but they work well for simple and positive regions of outliers, where boundary regions are not given much importance. Moreover, an algorithm which processes stream data must be effective and able to compute infinite data in one pass or limited number of passes. These problems have motivated us to propose an outlier detection approach for large-scale data stream. The proposed algorithm employs the concept of relative cardinality, entropy outlier factor theory of information-based system, and size-variant sliding window in stream data. In addition, we propose a new methodology for concept drift adaptation on evolving data streams. The proposed method is executed on nine benchmark datasets and compared with six existing methods that are EXPoSE, iForest, OC-SVM, LOF, KDE, and FastAbod. Experimental results show that the proposed method outperforms six existing methods in terms of receiver operating characteristic curve, precision recall, and computational time for positive regions as well as for boundary regions.

Online Learning Model for Handling Different Concept Drifts Using Diverse Ensemble Classifiers on Evolving Data Streams

The rapid growth of the information technology accelerates organizations to generate vast volumes of high-velocity data streams. The concept drift is a crucial issue, and discovering the sequential patterns over data streams are more challenging. The ensemble classifiers incrementally learn the data for providing quick reaction to the concept drifts. The ensemble classifiers have to process both the gradual and sudden concept drifts that happen in the real-time data streams. Thus, a novel ensemble classifier is essential that significantly reacting to various types of concept drifts quickly and maintaining the classification accuracy. This work proposes the stream data mining on the fly using an adaptive online learning rule (SOAR) model to handle both the gradual and sudden pattern changes and improves mining accuracy. Adding the number of classifiers fails because the ensemble tends to include redundant classifiers instead of high-quality ones. Thus, the SOAR includes different diversity levels of classifiers in the ensemble to provide fast recovery from both the concept drifts. Moreover, the SOAR synthesizes the essential features of the block and online-based ensemble and updates the weight of each classifier, regarding its quality. It facilitates adaptive windowing to handle both gradual and sudden concept drifts. To reduce the computational cost and analyze the data stream quickly, the SOAR caches the occurred primitive patterns into a bitmap with the internal relationship. Finally, the experimental results show that the SOAR performs better classification and accuracy over data streams.

Resource-aware Mining of Data Streams

Mining data streams has raised a number of research challenges for the data mining community. These challenges include the limitations of computational resources, especially because mining streams of data most likely be done on a mobile device with limited resources. Also due to the continuality of data streams, the algorithm should have only one pass or less over the incoming data records. In this article, our Algorithm Output Granularity (AOG) approach in mining data streams is discussed. AOG is a novel adaptable approach that can cope with the challenging inherent features of data streams. We also show the results for AOG based clustering in a resource constrained environment.

Data Stream Mining Applied to Maximum Wind Forecasting in the Canary Islands.

The Canary Islands are a well known tourist destination with generally stable and clement weather conditions. However, occasionally extreme weather conditions occur, which although very unusual, may cause severe damage to the local economy. The ViMetRi-MAC EU funded project has among its goals, managing climate-change-associated risks. The Spanish National Meteorology Agency (AEMET) has a network of weather stations across the eight Canary Islands. Using data from those stations, we propose a novel methodology for the prediction of maximum wind speed in order to trigger an early alert for extreme weather conditions. The methodology proposed has the added value of using an innovative kind of machine learning that is based on the data stream mining paradigm. This type of machine learning system relies on two important features: models are learned incrementally and adaptively. That means the learner tunes the models gradually and endlessly as new observations are received and also modifies it when there is concept drift (statistical instability), in the modeled phenomenon. The results presented seem to prove that this data stream mining approach is a good fit for this kind of problem, clearly improving the results obtained with the accumulative non-adaptive version of the methodology.

On the application of sequential pattern mining primitives to process discovery: Overview, outlook and opportunity identification

AbstractSequential pattern mining (SPM) is a well‐studied theme in data mining, in which one aims to discover common sequences of item sets in a large corpus of temporal itemset data. Due to the sequential nature of data streams, supporting SPM in streaming environments is commonly studied in the area of data stream mining as well. On the other hand, stream‐based process discovery (PD), originating from the field of process mining, focusses on learning process models on the basis of online event data. In particular, the main goal of the models discovered is to describe the underlying generating process in an end‐to‐end fashion. As both SPM and PD use data that are comparable in nature, that is, both involve time‐stamped instances, one expects that techniques from the SPM domain are (partly) transferable to the PD domain. However, thus far, little work has been done in the intersection of the two fields. In this focus article, we therefore study the possible application of SPM techniques in the context of PD. We provide an overview of the two fields, covering their commonalities and differences, highlight the challenges of applying them, and, present an outlook and several avenues for future work.This article is categorized under: Algorithmic Development > Spatial and Temporal Data Mining Fundamental Concepts of Data and Knowledge > Key Design Issues in Data Mining Fundamental Concepts of Data and Knowledge > Big Data Mining