Mining time-changing data streams

By Geoff Hulten, Laurie Spencer, Pedro M. Domingos. 2001

1,940 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Systems, Evaluation & Benchmarks

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Evaluation & Benchmarks

Citations in the corpus

Towards the Science of Security and Privacy in Machine Learning (2016, background): “In Section VI-A, we present a line of work that builds on that observation to propose learning strategies robust to distribution drifts [43].”
Memory Efficient Experience Replay for Streaming Learning (2018, background): “Many are based on Hoeffding Decision Trees [6], [18], [25], [36],”
Lifelong Machine Learning with Deep Streaming Linear Discriminant Analysis (2019, background): “Hoeffding Decision Trees [4, 16, 21, 27, 28] incrementally grow decision trees over time under the Hoeffding bound theoretical guarantees.”
SoK: Security and Privacy in Machine Learning (2018, background): “1, we present a line of work that builds on that observation to propose learning strategies robust to distribution drifts [36].”
Learning under Concept Drift: A Review (2019, background): “[58], [59], [64], [72], [73], [74], [79], [82], [85],”
A Survey on Ensemble Learning for Data Stream Classification (2017, background): “Speciﬁcally Hoeffding Trees 6 [Domingos and Hulten 2000] or some of its variations that explicitly deals with concept drift as Adaptive Hoeffding Trees [Bifet and Gavald`a 2009] and Concept-Adaptive Very Fast Decision Trees (CVFDT) [Hul-ten et al. 2001], or that replaces majority class predictions b”
Concept drift detection via competence models (2014, methodology): “Otherwise, as concept drifts, the induced pattern may not be relevant to the new data [5, 6], which may result in an increasing number of errors [7].”
Learning in Nonstationary Environments: A Survey (2015, methodology): “using an adaptive sliding window for training [76].”
Efficient elastic burst detection in data streams (2003, background): “The work [13] studies the problem of learning models from timechanging streams without explicitly applying the sliding window model.”
Accurate decision trees for mining high-speed data streams (2003, background): “In this paper, we do not discuss the problem of timechanging concepts [7].”
A survey on data preprocessing for data stream mining: Current status and future directions (2017, background): “• Sliding windows assume that we keep a buffer of fixed size containing most recent examples [26].”
Incremental Clustering for the Classification of Concept-Drifting Data Streams (2008, methodology): “An additional group of methods could be the category of Quickly adapting incremental algorithms: This group consists of methods like CVFDT [8] that try to adapt rapidly to the newest batch of data in order to cope with the drift.”
Classifier evaluation and attribute selection against active adversaries (2010, background): “As discussed in Dalvi et al. (2004), the challenges created by the malicious adversaries are quite different from those in concept drift (Hulten et al. 2001), because the concept is maliciously changed based on the actions of the classifier.”
Efficient decision tree construction on streaming data (2003, background): “It may be possible for a decision tree construction algorithm to adjust the tree to changes in the distribution of data instances in the stream [21], but we do not consider this possibility here.”
Quantifying counts and costs via classification (2008, methodology): “As a side note, we point out a potential connection between tracking trends over time and the growing field of learning from data streams (e.g., Hulten et al. 2001 ).”
Models and issues in data stream systems (2002, background): “[28, 29] have studied the problem of maintaining decision trees over datastreams.”
Online Machine Learning in Big Data Streams (2018, methodology): “out fast semantics. Large databases available for mining today have been gathered over months or years, and the underlying processes generating them have changed during this time, sometimes radically [119]. In data analysis tasks, fundamental properties of the data may change Online Machine Learning”
Estimation of energy consumption in machine learning (2019, background): “The state-of-the-art algorithms in this area are able to handle changes in the input data distribution, known as concept drift [41].”
A Survey of Clustering Data Mining Techniques (2006, methodology): “These bounds were used in the clustering algorithm CURE [Guha et al. 1998] and in the development of scalable decision trees in predictive mining [Hulten et al. 2001].”