Concept Tracking and Adaptation for Drifting Data Streams under Extreme Verification Latency

Arostegi, Maria; Torre-Bastida, Ana I.; Lobo, Jesús L.; Bilbao, Miren Nekane; Ser, Javier Del

doi:10.1007/978-3-319-99626-4_2

Cited by 3 publications

(1 citation statement)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also note that two other algorithms are also proposed to work in the EVL setting more recently; these are called TRACE [22], which tracks the trajectory of the clusters over time using some trajectory prediction algorithm for instance Kalman filter, instead of tracking clusters using unsupervised learning algorithms as done in COMPOSE and SCARGC, and Affinity-based COMPOSE [23], which is based on COMPOSE with a slight modification in the core support extraction module. Affinity based COMPOSE uses only those samples from the previous timestep as the labeled information which has the highest similarity scores with the unlabeled samples at current time step, computed from the affinity matrix.…”

Section: Algorithm 5: Mclassificationmentioning

confidence: 99%

Comparative Analysis of Extreme Verification Latency Learning Algorithms

Umer,

Polikar

2020

Preprint

View full text Add to dashboard Cite

One of the more challenging real-world problems in computational intelligence is to learn from non-stationary streaming data, also known as concept drift. Perhaps even a more challenging version of this scenario is when -following a small set of initial labeled data -the data stream consists of unlabeled data only. Such a scenario is typically referred to as learning in initially labeled nonstationary environment, or simply as extreme verification latency (EVL). Because of the very challenging nature of the problem, very few algorithms have been proposed in the literature up to date. This work is a very first effort to provide a review of some of the existing algorithms (important/prominent) in this field to the research community. More specifically, this paper is a comprehensive survey and comparative analysis of some of the EVL algorithms to point out the weaknesses and strengths of different approaches from three different perspectives: classification accuracy, computational complexity and parameter sensitivity using several synthetic and real world datasets.

show abstract

Section: Algorithm 5: Mclassificationmentioning

confidence: 99%