Využití Kohonenovy mapy generalizačního gradientu klasifikační umělé neuronové sítě jako uživatelsky vstřícného nástroje výběru, umisťování a rozvoje
Utilization of Kohonen’s self-organizing feature map of ANN‘s generalization gradient as a user-friendly tool for employee recruitment, placing (...)
Author(s): Luděk StehlíkSubject(s): Psychology
Published by: Univerzita Karlova v Praze, Nakladatelství Karolinum
Keywords: human resources; benchmark; artificial neural networks; Kohonen self-organizing maps; generalization gradient; multidimensional data visualization
Summary/Abstract: The article refers to a proposal of user friendly and easily adaptable tool which can facilitate and enhance HR managers’ decisions in domain of employee recruitment and development. Proposed tool has three main parts: (a) benchmark study of the employee ideal profile, (b) three-layered feedforward artificial neural network (ANN) used for processing benchmark study data and for classification of employees and applicants into two different performance groups and (c) Kohonen Self-Organizing maps enabling visualization of ANN’s generalization gradient. User interface of proposed tool is a set of maps with identical topographical organization but with different vertical profile according to displayed variable. HR managers can read very easily and intuitively from these maps important and useful information which can be directly used in the HR related decision making process.
Journal: Psychologie pro praxi
- Issue Year: 2010
- Issue No: 1-2
- Page Range: 9-27
- Page Count: 19
- Language: Czech