Approximation and Prediction of Wages Based on Granular Neural Network
This article offers a detailed computational algorithm used in that type of neural networks, extends their applications to fit and predict the data of wages time series, conducts experiments and indicates the gain of granular neural networks, specifically
- PDF / 220,031 Bytes
- 8 Pages / 430 x 660 pts Page_size
- 77 Downloads / 280 Views
aculty of Philosophy and Science, Silesian University, 746 01 0pava, Czech Republic & MEDIS Nitra, Ltd., Pri Dobrotke 659/81, 949 01 Nitra-Dražovce, Slovak Republic [email protected] 2 Faculty of Philosophy and Science, Silesian University, 746 01 0pava, Czech Republic & Faculty of Management Science and Informatics, University of Zilina 010 26 Zilina, Slovak Republic [email protected], [email protected]
Abstract. This article offers a detailed computational algorithm used in that type of neural networks, extends their applications to fit and predict the data of wages time series, conducts experiments and indicates the gain of granular neural networks, specifically conducting experimentation using the classical (statistical) or econometric methods and conventional/soft RBF neural networks. Results are analysed and opportunities for future research are suggested. Keywords: Probabilistic time-series models, Fuzzy system, Classic and soft RBF network, Cloud models, Granular computing.
1 Introduction As mentioned in Liao et al. [5] and Zhang et al. [11] neural networks document competitive performance on a larger number of time series, indicating the use of increased computational power to automate NN forecasting on a scale suitable for automatic forecasting. The scope of the paper is confined to some statistical forecasting methods and methods based on granular computing. According to Zadeh [10] granulation plays an essential role in human cognition and has a position of centrality in both granular computing and rough set theory. The exploitation of granular concept for forecasting purposes can be found in many works. In Yu et al. [9] a method is proposed based on information granulation model and the granular discretization method to form fuzzy rules from granular time series for a fuzzy forecasting system. In Yang et al. [8] methods of time series prediction based on cloud methods and on the different time (short term and long term granularities) were presented and described respectively. Lastly in Marcek et al. [6] a new approach of function estimation is shown for time series model of daily sales by means of a granular RBF neural network. In comparison with [6], this paper extends the application of granular network to fit and predict the quarterly data of wages time series, gives new calculating algorithm for the specific granular network and compares obtained results with those obtained G. Wang et al. (Eds.): RSKT 2008, LNAI 5009, pp. 556–563, 2008. © Springer-Verlag Berlin Heidelberg 2008
Approximation and Prediction of Wages Based on Granular Neural Network
557
using statistical procedures. The organization of the paper involves 5 sections. Section 2 outlines necessary prerequisites for predictors based on the fuzzy system and RBF neural network approach. In Section 3, we briefly offer a granular extension of RBF neural networks. In Section 4 we give the complete algorithm for weights updating in the granular RBF network and for calculating of the output values and the statistical summary me
Data Loading...