A comparative study on forecast analysis of growth-type time series with ‘Gap’

Usually time series curve of the overall uptrend shows abnormal deviation phenomenon, called ‘gap’ phenomena. To solve the forecast problem for the time series data with ‘gap’, the output of aquatic products of Liaoning Province from 1985 to 2008 is analyzed as analysis object. The Logistic model, autoregressive moving average model and fuzzy time series model are used for the output of aquatic products of Liaoning Province analysis respectively. The comparative study of the forecast results and the error analysis results of the three kinds of models shows that proper fuzzy time series analysis methods are more effective for time series analysis. Two new time series data are formed through adjusting some data of the output of aquatic products of Liaoning Province. Each of them has two ‘gaps’ in the same side or in the opposite side. In this case the image analysis and error estimation still show that the fuzzy time series model has established good prediction effect.