A Fuzzy Time Series Model for Forecasting the Volume of Cargo

Abstract In this article, we intend to apply the fuzzy timeseries model to forecast the volume of cargo handledby international commercial ports in Taiwan area. Theroot mean square error is one criteria to evaluate theforecast performance. Empirical results show that theproposed fuzzy time series model is suitable for theprediction of volume of cargo. Keywords ：Fuzzy time series, fuzzy theory, forecast 1. Introduction Forecasting of the volume of cargo is central to theplanning and the operation of seaport organizationsand government transportation departments at bothmicro and macro levels. At the seaport organizationlevel, forecasts of cargo volumes are needed as theessential inputs to many decision activities in variousfunctional areas such as building new terminals,operation plans, marketing strategies, as well asfinance and accounting. Whether new terminalsshould be built is a controversial issue in Taiwan. Onereason for this is the huge and irreversible investmentin new terminals and related infrastructure. Anotherreason is the different ways in forecasting the volumeof cargo, which lead to quite different conclusions. Atthe government transportation department level,forecasts of cargo volume also provide basis forregional and national transportation plans.The concept of fuzzy sets, which was introduced byZadeh in 1965, led to the definition of the fuzzynumber and its implementation in fuzzy control,approximate reasoning and fuzzy forecast problems.Fuzzy time series model have been developed andapplied to solving many forecast problems in the realworld [1-10].In this article, we intend to apply the fuzzy timeseries model to forecast the volume of cargo handledby international commercial ports in Taiwan area. Therest of the article is structured as follows. Section 2presents the fuzzy time series definition. In Section 3,we present the fuzzy time series steps based onChou’s method [2-5], the method is also evaluated insection 4. Finally, conclusions are made in section 5.