摘要：This work presents the first the analysis of the seasonal time series by decomposition in 3: trend, seasonal oscillation and residuals. This is made by statistical filtering (STL linear filtering and Holt-Winters filtering) which results in the graphic plotting of the components. As an example we have chosen the monthly time series of alive newborn in Bacau county between 2000-2005 on the whole and in the two media. The last phase shows forecasts on those time series using ARIMA method. It is an analysis of the seasonality of time series and of the differences between the two media. It was used the R environment that allows the user fast access and suggestive charts. The method of filtering shows hidden aspects of the components making them visible on the chart. Both the filtering and ARIMA methods allow an appreciation of future trends and of objectively regarding the quality of the existing time series. The analysis of the components leads to the fact that the seasonal scope is higher compared to the general level of the time series
