Title: Easing knowledge management in the power sector by means of a neuro-genetic system

Authors: Lourdes Sáiz-Bárcena; Álvaro Herrero; Miguel Ángel Manzanedo del Campo; Ricardo Del Olmo Martínez

Addresses: Departamento de Ingeniería Civil, Escuela Politécnica Superior, Universidad de Burgos, Avda. de Cantabria S/N, 09006, Burgos, Spain ' Departamento de Ingeniería Civil, Escuela Politécnica Superior, Universidad de Burgos, Avda. de Cantabria S/N, 09006, Burgos, Spain ' Departamento de Ingeniería Civil, Escuela Politécnica Superior, Universidad de Burgos, Avda. de Cantabria S/N, 09006, Burgos, Spain ' Departamento de Ingeniería Civil, Escuela Politécnica Superior, Universidad de Burgos, Avda. de Cantabria S/N, 09006, Burgos, Spain

Abstract: A hybrid intelligent system that combines neural networks and genetic algorithms for feature selection is proposed in this paper. In the world of companies, it is widely acknowledged the importance of knowledge management (KM) in terms of organisation survival and maintenance of competitive strength. Furthermore, in the energy sector, it is even more critical as energy is a strategic national industry for most countries. However, there are not enough tools with which to systematise the work of KM and to provide support for decision-making. The present study proposes a tool that addresses the problem of selecting the features of KM (from the immense amount of available data) that would provide managers with most valuable information. The developed system has been applied to real-world data from a power plant, to mine deep knowledge on that KM data, and to select the most informative features. Experimental results justify a meaningful subset of features in the opinion of KM experts, and validate the proposed hybrid system.

Keywords: knowledge management; decision support; power generation; wrapper feature selection; genetic algorithms; support vector machines; SVM; neural projection techniques; neural networks; data mining.

DOI: 10.1504/IJBIC.2015.069556

International Journal of Bio-Inspired Computation, 2015 Vol.7 No.3, pp.170 - 175

Received: 27 Sep 2014
Accepted: 07 Jan 2015
Published online: 26 May 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article