POSSIBLE METHODS FOR PRICE FORECASTING Szilágyi Roland 1 , Varga Beatrix 2 , Géczi-Papp Renáta 3 1 Ph.D, Associate professor, 2 Ph.D, Associate professor, 3 Ph.D student 1,2,3 University of Miskolc, Faculty of Economics, Department of Business Statistics and Economic Forecasting ABSTRACT The study gives an overview about the different time series analytical methods, which can be used for price forecasting. The comparison of the decomposition and stochastic methods will help to navigate through the many option of analytical techniques. To find a reliable forecasting method, the study introduces an empirical time series price analysis with a hybrid method: forecasting based on creeping trend with harmonic weights. The result show that this method provides an accurate and reliable prediction of future prices. INTRODUCTION The determination of prices is important in the economy. Prices are one of the major factors of a firm’s competitiveness, because it has a huge influence on the future of the company. Acquiring sophisticated knowledge of the past and current internal rules of development is of utmost importance in making a reliable forecasting of the future processes. In order to predict the future, the past and the present have to be carefully analysed since the future is the further development of the present. To make the right decisions it is essential an accurate, reliable, scientific based forecast. There are several analytical methods for forecasting, and this study aims at presenting the most important ones through the example of price forecast. The first to chapter are presented the most important forecasting methods in the case of decomposition and stochastic relations too. This part contains a comparison too, which includes the advantages and disadvantages of the models. In the third chapter are presented the results of an empirical analysis, with forecasting based on creeping trend with harmonic weights, using methanol price data. This is a mixed method, using difference techniques and gives reliable results. The study concludes with a summary. 1. FORECASTING ON THE BASIS OF TIME SERIES TECHNIQUE “A time series is defined as a set of quantitative observations arranged in chronological order. We generally assume that time is a discrete variable.” [1] During an analysis special attention should be paid to the ways the components’ linkage. Four types of links can be distinguished in the decomposition models [2]: -additive model: Y=T+S+I, (1) -multiplicative model: Y=T*S*I, (2) -log additive model: logY=logT+logS+logI, (3) -pseudo additive model: Y=T*(S+I-1), (4) MultiScience - XXX. microCAD International Multidisciplinary Scientific Conference University of Miskolc, Hungary, 21-22 April 2016, ISBN 978-963-358-113-1 DOI: 10.26649/musci.2016.135