A Precise Statistical approach for concept change detection in unlabeled data streams

Mozafari, Niloofar; Hashemi, Sattar; Hamzeh, Ali

doi:10.1016/j.camwa.2011.06.001

Cited by 13 publications

(8 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…or by a strangeness measure [14,26]. After reducing the dimensionality, online change detection boils down to monitoring a univariate datastream, for instance by Martingale-based permutation tests [14].…”

Section: Related Workmentioning

confidence: 99%

Change Detection in Multivariate Datastreams Controlling False Alarms

Frittoli

Carrera

Boracchi

2021

Machine Learning and Knowledge Discovery in Databases. Research Track

View full text Add to dashboard Cite

We introduce QuantTree Exponentially Weighted Moving Average (QT-EWMA), a novel change-detection algorithm for multivariate datastreams that can operate in a nonparametric and online manner. QT-EWMA can be configured to yield a target Average Run Length (ARL0), thus controlling the expected time before a false alarm. Control over false alarms has many practical implications and is rarely guaranteed by online change-detection algorithms that can monitor multivariate datastreams whose distribution is unknown. Our experiments, performed on synthetic and real-world datasets, demonstrate that QT-EWMA controls the ARL0 and the false alarm rate better than state-of-the-art methods operating in similar conditions, achieving comparable detection delays.

show abstract

Section: Related Workmentioning

confidence: 99%

Change Detection in Multivariate Datastreams Controlling False Alarms

Frittoli

Carrera

Boracchi

2021

Machine Learning and Knowledge Discovery in Databases. Research Track

View full text Add to dashboard Cite

show abstract

“…With this information, the tweet is then run through a hypothesis test to determine if its features represent a prominent change in the data stream. More formally, the test is as follows: H0, there is no change in the data stream (i.e, no marked emotional change), and H1 otherwise [38].…”

Section: Design Of a Martingale-based Approach For Emotion Change Detmentioning

confidence: 99%

“…In order to detect the movements of the behavioral features towards or away from their representative we average two martingale sequences [38]. The first martingale is as defined above and the second one is calculated using the complement of the -value at each observation (i.e., 1 −̂).…”

Section: Unified Strangeness Measure (Usm)mentioning

confidence: 99%

Detection of suicide-related posts in Twitter data streams

Vioulès¹,

Moulahi²,

Azé³

et al. 2018

IBM J. Res. & Dev.

120

View full text Add to dashboard Cite

Suicidal ideation detection in online social networks is an emerging research area with major challenges. Recent research has shown that the publicly available information spread across social media platforms holds valuable indicators to effectively detecting individuals with suicidal intentions. The key challenge of suicide prevention is understanding and detecting the complex risk factors and warning signs that may precipitate the event. In this paper, we present a new approach that uses the social media platform Twitter to quantify suicide-warning signs for individuals and to detect posts containing suicide-related content. The main originality of this approach is the automatic identification of sudden changes in a user's online behavior. To detect such changes, we combine natural language processing techniques to aggregate behavioral and textual features and pass these features through a martingale framework, which is widely used for change detection in data streams. Experiments show that our text-scoring approach effectively captures warning signs in text compared to traditional machine learning classifiers. Additionally, the application of the martingale framework highlights changes in online behavior and shows promise for detecting behavioral changes in at-risk individuals.

show abstract

“…[41] Although Ho's method detects changes points accurately, it can only detect some types of changes to be more detailed in. [42]…”

Section: Data Driven Concept Change Detectionmentioning

confidence: 99%

“…Thus, we can detect changes when martingale value is greater than λ. [42] When a change is occurred in any cluster, all the previous information is removed. To be more illustrative, we presented the outline of our method for clustering data stream as follow: StatisStreamClustStatisStreamClust Partitions instances after arriving 50 instances using kmeans algorithm Loop A new unlabeled data stream z i is received.…”

Section: Statisstreamclustmentioning

confidence: 99%

A statistical approach for clustering in streaming data

Mozafari

Hashemi

Hamzeh³

2014

AIR

Self Cite

View full text Add to dashboard Cite

Recently data stream has been extensively explored due to its emergence in large deal of applications such as sensor networks, web click streams and network flows. Vast majority of researches in the context of data stream mining are devoted to supervise learning, whereas, in real word human practice label of data are rarely available to the learning algorithms. Hence, clustering as the most important unsupervised learning has been in the gravity of focus of quite a lot number of the researchers in data stream community. Clustering paradigms basically place the similar objects together and separate the dissimilar ones into different clusters. In this paper, we propose a Statistical framework for data Stream Clustering, which abbreviated as StatisStreamClust that makes use of two components to find clusters in data stream. The first component especially designed to detect concept change where data underlying distributions change from time to time. Upon detection of concept change by the first component, the second component is triggered to update the whole clustering model. StatisStreamClust brings great benefits to data stream clustering including no sensitivity to the number of clusters and dimensions, reasonable complexity and in the meantime desirable performance, and finally no need to determine window size a priori. To explore the advantages of our approach, quite a lot of experiments with different settings and specifications are conducted. The obtained results are very promising.

show abstract

A Precise Statistical approach for concept change detection in unlabeled data streams

Cited by 13 publications

References 12 publications

Change Detection in Multivariate Datastreams Controlling False Alarms

Change Detection in Multivariate Datastreams Controlling False Alarms

Detection of suicide-related posts in Twitter data streams

A statistical approach for clustering in streaming data

Contact Info

Product

Resources

About