Advances on Concept Drift Detection in Regression Tasks Using Social Networks Theory

Advances on Concept Drift Detection in Regression Tasks Using Social Networks Theory

Jean Paul Barddal (Programa de Pós-Graduação em Informática, Pontifícia Universidade Católica do Paraná, Curitiba, Brazil), Heitor Murilo Gomes (Programa de Pós-Graduação em Informática, Pontifícia Universidade Católica do Paraná, Curitiba, Brazil) and Fabrício Enembreck (Programa de Pós-Graduação em Informática, Pontifícia Universidade Católica do Paraná, Curitiba, Brazil)
Copyright: © 2015 |Pages: 16
DOI: 10.4018/ijncr.2015010102

Abstract

Mining data streams is one of the main studies in machine learning area due to its application in many knowledge areas. One of the major challenges on mining data streams is concept drift, which requires the learner to discard the current concept and adapt to a new one. Ensemble-based drift detection algorithms have been used successfully to the classification task but usually maintain a fixed size ensemble of learners running the risk of needlessly spending processing time and memory. In this paper the authors present improvements to the Scale-free Network Regressor (SFNR), a dynamic ensemble-based method for regression that employs social networks theory. In order to detect concept drifts SFNR uses the Adaptive Window (ADWIN) algorithm. Results show improvements in accuracy, especially in concept drift situations and better performance compared to other state-of-the-art algorithms in both real and synthetic data.
Article Preview

Most of the existing works on ensembles rely on developing algorithms to improve overall accuracy coping with concept drift explicitly (Bifet, Holmes, Pfahringer, Kirkby, & Gavaldà, 2009) or implicitly (Kolter & Maloof, 2005; Widmer & Kubate, 1996). Authors in (Kuncheva, 2004) shows that an ensemble can surpass an individual expert's accuracy if its component experts are diverse. An ensemble is said to be diverse if its members misclassify different instances (in regression tasks if they predict instances with different values). Another important trait of an ensemble refers to how it combines individual decisions. If the combination strategy fails to highlight correct and obfuscate incorrect decisions then the method is jeopardized. In the remainder of this section we present the state-of-the-art algorithms for data stream regression, including single classifier and ensemble methods.

Moving Average

The Moving Average is one of the oldest indicators of technical analysis for stock market forecasting (Brockwell & Davis, 2002). Its computation is based on a weighted average of historic stock values. We chose the Exponential Moving Average (EMA) since a conventional Moving Average takes too long to predict market tendencies. Equation 1 presents the Exponential Moving Average computation where stands for the price of a given stock in a time and is the algorithm's window size.

(1)

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 8: 4 Issues (2019): Forthcoming, Available for Pre-Order
Volume 7: 4 Issues (2018)
Volume 6: 2 Issues (2017)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing