A novel hybrid system for feature selection based on an improved gravitational search algorithm and k-NN method

Xiang, Jie; Han, Xiaoqing; Fu, Duan; Qiang, Yan; Xiong, Xuejian; Yuan, Liang; Chai, Haishui

doi:10.1016/j.asoc.2015.01.043

Cited by 62 publications

(21 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Gravitational search algorithm (GSA) belongs to a class of population‐based heuristic optimization algorithms that operate using Newton principles of motion . GSA involves the interaction of masses under the influence of gravitational force in multiple dimensional search spaces and, ultimately, results in fast convergence due to sluggishness of the massive agent with approximately zero acceleration . The global solution corresponds the position of the most massive agent with a high tendency of attracting agents of lower masses .…”

Section: Formulation Of the Proposed Hybrid Intelligent Modelmentioning

confidence: 99%

Hybrid Intelligent Modelling of the Viscoelastic Moduli of Coal Fly Ash Based Polymer Gel System for Water Shutoff Treatment in Oil and Gas Wells

et al. 2019

View full text Add to dashboard Cite

Qualitative polymer gels exhibiting considerable viscoelastic properties are of immense significance in the petroleum industry as they curtail excessive water production during the tertiary stage of crude oil production. The incorporation of suitable organic or inorganic solid substances into the polymer gel system strengthens its viscoelastic properties. This research work presents hybrid gravitational search algorithm (GSA) and support vector regression (SVR) as novel modelling tools for viscoelastic properties estimation of polymer gel synthesized from polyacrylamide (PAM)/polyethyleneimine (PEI). The coal fly ash (CFA) was incorporated in PAM/PEI gel in order to enhance its viscoelastic characteristics. The hybrid GSA‐SVR models were developed using real‐life experimental data. The results of the models validated using unseen data show a high degree of correlation coefficients (87.3 % for elastic modulus and 97.2 % for viscous modulus), with experimental data generating a small root mean square error (coefficients 669.5Pa for elastic modulus and 38.5Pa for viscous modulus). The developed hybrid models were further used to scrutinize the effect of the addition of various amounts of CFA on the PAM/PEI viscoelastic properties and observed behaviours were presented and discussed. The predictive capability of the developed hybrid model coupled with its strength in modelling the impact of CFA in the PAM/PEI gel are highly meritorious and the outcomes from this investigation would be relevant during the design and formulation of polymer gels employed for water shutoff treatment during hydrocarbon recovery.

show abstract

Section: Formulation Of the Proposed Hybrid Intelligent Modelmentioning

confidence: 99%

Hybrid Intelligent Modelling of the Viscoelastic Moduli of Coal Fly Ash Based Polymer Gel System for Water Shutoff Treatment in Oil and Gas Wells

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Actualmente se implementan algoritmos híbridos [14,15] combinando diferentes técnicas de filtrado de datos y algoritmos wrapper que trabajan de forma paralela, aplicados en la selección y extracción de genes relevantes para el diagnóstico de una enfermedad. A pesar del número de técnicas implementadas para abordar el problema de selección de genes, no se ha llegado a una solución concreta, cada modelo o método presentado selecciona genes que no se han reportado y abre el panorama a nuevos estudios, debido a esto surgen más trabajos con nuevas propuestas dando un estudio más confiable de los genes seleccionados.…”

Section: Estado Del Arteunclassified

Selección y clasificación de genes cancerígenos utilizando un método híbrido filtro/wrapper

Montiel¹,

Pérez²,

Ruiz³

2017

RCS

View full text Add to dashboard Cite

En este artículo, se presenta un método híbrido aplicado en la selección y clasificación de genes obtenidos de microarreglos de ADN. El método realiza una limpieza del microarreglo utilizando un preprocesamiento basado en técnicas de filtrado de datos, después, se realiza una selección de genes dentro del subconjunto obtenido por el filtro, utilizando una búsqueda gravitacional combinada con un clasificador KNN. El método se ha construido para obtener un subconjunto de genes relevantes de alto desempeño, los resultados obtenidos se comparan con diferentes métodos reportados en la literatura, este método es implementado en cinco microarreglos de dominio público.

show abstract

“…The proposed method is compared with well-known feature selection algorithms such as chi-square (CHI), information gain (IG) and correlation feature selection (CFS), and some recent feature selection methods that involve metaheuristics including harmony search and stochastic search algorithms for feature selection (Nekkaa and Boughaci, 2015), a binary ABC with DE operators (Hancer et. al, 2015), a novel system for feature selection based on gravitational search (Xiang et. al, 2015), an unsupervised feature selection method based on improved version of DE (Bhadra and Bandyopadhyay, 2015), the relevance-redundancy filter method based on ACO for feature selection (Tabakhi and Moradi, 2015), as well as standard ABC and DE algorithms.…”

Section: Introductionmentioning

confidence: 99%

A hybrid approach of differential evolution and artificial bee colony for feature selection

Zorarpacı

Özel

2016

Expert Systems with Applications

305

117

View full text Add to dashboard Cite

A novel hybrid system for feature selection based on an improved gravitational search algorithm and k-NN method

Cited by 62 publications

References 43 publications

Hybrid Intelligent Modelling of the Viscoelastic Moduli of Coal Fly Ash Based Polymer Gel System for Water Shutoff Treatment in Oil and Gas Wells

Hybrid Intelligent Modelling of the Viscoelastic Moduli of Coal Fly Ash Based Polymer Gel System for Water Shutoff Treatment in Oil and Gas Wells

Selección y clasificación de genes cancerígenos utilizando un método híbrido filtro/wrapper

A hybrid approach of differential evolution and artificial bee colony for feature selection

Contact Info

Product

Resources

About