Any change in sale price may affect customers, distributers and sellers. Anticipating future prices is one of the best ways to face appropriately such these price changes in the market. Time series have wide range of application in various fields such as economy, management and marketing. Time series is a very important tool to analyze a collection of observations which are recorded as daily, weekly, monthly and annually reports. In this paper, the world price of each ounce of gold during 338 continuous months are considered (Average per month) and the target is to assess the behavior of data and to release a suitable model for this data to anticipate world price of each ounce of gold during upcoming months by means of analysis of time series. The first step to analyze time series is to draw data. Next step is to recognize effective parameters on the series (trend, cycle and seasonal) and to remove them from time series and at last to process a static model on time series. We drew autocorrelation function (ACF) and partial autocorrelation function (PACF) for data. Auto-regression model (AR), moving average model (MA) and a combination of AR and MA models (ARIMA, ARMA) were selected as the grade of recognition model and appropriate model. After all stages to analyze time series and creation of remained parameters and after consideration of fitness of represented model, anticipation of world price of gold for each ounce will take place. In this regard, the result of considering the data in this paper produces information for future to make appropriate and profitable decision based on current data. The process is done by means of MINITAB software.
Keywords: ACF, ARIMA, Forecast, PACF, Time Series
