Concept drift is a change in the joint probability distribution of the problem. This term can be subdivided into two types: real drifts that affect the conditional probabilities p(y|x) or virtual drifts that affect the unconditional probability distribution p(x). Most existing work focuses on dealing with real concept drifts. However, virtual drifts can also cause degradation in predictive performance, requiring mechanisms to be tackled. Moreover, as virtual drifts frequently mean that part of the old knowledge remains useful, they require different strategies from real drifts to be effectively tackled. Motivated on this, we propose an approach called Gaussian Mixture Model for Dealing With Virtual and Real Concept Drifts (GMM-VRD), which updates and creates Gaussians to tackle virtual drifts and resets the system to deal with real drifts. The main results show that the proposed approach obtained the best results, in terms of average accuracy, in relation to the literature methods, which propose to solve that same problem. In terms of accuracy over time, the proposed approach showed lower degradation on concept drifts, which indicates that the proposed approach was efficient.
I thank my parents, husband, and family for all support and encouragement for the realization of this dream. I thank also my friend, Richard McGill, for the support and review of papers related to this Thesis."Obstacles are those frightful things you see when you take your eyes off your goal."
O artigo desenvolve uma discussão teórico-metodológica sobre as possibilidades e condições para a articulação de abordagens pós-estruturalistas, especialmente da Teoria do Discurso de Laclau e Mouffe, na realização de pesquisas empíricas em educação. Inicialmente, é delineada uma breve revisão conceitual sobre a ontologia pós-estruturalista e sobre suas consequências quanto à rejeição do positivismo e da epistemologia moderna. A seguir, são apresentadas e analisadas quatro proposições formuladas por Glynos e Howarth para a realização de pesquisas referenciadas na teoria do discurso. Por fim, é resgatado o debate sobre a exigência de rigor analítico e sobre o potencial crítico dessa teoria na realização de pesquisas no campo da Educação.
Introduction Immobilisation of biomaterials has been widely used in many fields, such as industry, biochemistry and immunology (Cheetham, 1985; Hermanson et al., 1992). This technology employs a large number of natural and artificial materials as matrices (Kennedy and Cabral, 1987). In our laboratory, we have proposed immobilisation procedures using dacron (Carneiro Leão et al., 1994), polyaniline (Nadruz et al., 1996) and glyptal (Jordão et al., 1996). Here, polyvinyl alcoholglutaraldehyde (PVA-glutaraldehyde) network is described for antigen and enzyme immobilisation. Previously, this material was used for the diagnosis of plague as a solid-phase in ELISA (Araujo et al., 1996) and also in laser induced fluorescence (Carvalho et al., 1996). Materials and methods Synthesis of PVA-glutaraldehyde beads Polyvinyl alcohol (2.0 g; Reagen ®) was dissolved in deionised water (20 mL) containing sodium dodecylsulfate (10 mg; Sigma) under heating and stirring. Glutaraldehyde (25%, w/v; 4.12 mL; Sigma) and H 2 SO 4 (0.3 M; 1 mL) were successively added to the dissolved PVA and then the solution was slowly poured into a beaker containing mineral oil/water mixture (45:45; v/v) with stirring. After about 15 min, PVA-glutaraldehyde beads were synthesised. In order to remove oil and sodium dodecylsulfate the beads were exhaustively washed with ethanol 95% v/v (500 mL) and deionised water (1 L). Synthesis of PVA-glutaraldehyde discs According to Araujo et al. (1996). Enzyme immobilisation ␣-Amylase (20 mL of 65 U/mg protein; Termamyl, Novo Nordisk, Paraná, Brazil) prepared in 50 mM citrate-phosphate buffer, pH 4.5, was incubated with PVA-glutaraldehyde beads (1 g) under stirring overnight at 4°C. After immobilisation, the water insoluble enzymatic derivative was washed with 1 M NaCl (100 mL) in buffer. Amyloglucosidase (20 mL of 6.5 U/mg protein; AMG, Novo Nordisk, Paraná, Brazil), prepared in 50 mM citrate-phosphate buffer, pH 6.5, was incubated with PVA-glutaraldehyde as described above. Xanthine oxidase (10 mL; Nutritional Biochemichals Corporation) prepared in 0.1 M phosphate buffer, pH 8.0, was incubated with PVAglutaraldehyde beads (0.6 g) under mild stirring overnight at 4°C. Then, the immobilised enzymatic derivative was exhaustively washed with deionised water and 1 M NaCl, successively. Afterwards, overnight incubation with 1 M glycine was carried out to inactivate free carbonyl groups and washings were carried out again.
Real-world applications have been dealing with large amounts of data that arrive over time and generally present changes in their underlying joint probability distribution, i.e., concept drift. Concept drift can be subdivided into two types: virtual drift, which affects the unconditional probability distribution p(x), and real drift, which affects the conditional probability distribution p(y|x). Existing works focuses on real drift. However, strategies to cope with real drift may not be the best suited for dealing with virtual drift, since the real class boundaries remain unchanged. We provide the first in depth analysis of the differences between the impact of virtual and real drifts on classifiers' suitability. We propose an approach to handle both drifts called On-line Gaussian Mixture Model With Noise Filter For Handling Virtual and Real Concept Drifts (OGMMF-VRD). Experiments with seven synthetics and seven real-world datasets show that OGMMF-VRD outperforms other approaches with separate mechanisms to deal with virtual and real drifts. It also has more stable rankings and smaller drops in performance during drifting periods than existing ensemble approaches, thus being more reliable for adoption in practice.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.