Rainfall time series of 18 stations in Kerala, India, starting from 1981 and ending in 2013, is used. Along with this, the effectiveness of Yeo-Johnson transformation (YJT) in improving the forecast accuracy of the models is assessed. It enables an assessment of the significant difference in the rainfall characteristics at various locations that influence the relative forecasting accuracies of the models. The techniques are applied to forecast the rainfall time series of the stations located in Kerala. In this article, the performance evaluation of four univariate time-series forecasting techniques, namely Hyndman Khandakar-Seasonal Autoregressive Integrated Moving Average (HK-SARIMA), Non-Stationary Thomas-Fiering (NSTF), Yeo-Johnson Transformed Non-Stationary Thomas-Fiering (YJNSTF) and Seasonal Naïve (SN) method, is carried out. Forecasting of monthly, weekly, daily monsoon time series for 14 years, i.e., from 2014 to 2027 was done using the developed SARIMA models. Nash-Sutcliffe coefficient also indicated a high degree of model fitness to the observed data. The mean and standard deviation of the predicted data were found close to the observed data. Performance and validation of the SARIMA models were evaluated based on various statistical measures. As per Ljung-Box Q statistics, residuals were random in nature and there was no need for further modelling. The best SARIMA models were selected based on the autocorrelation function (ACF) and partial autocorrelation function (PACF), and the minimum values of Akaike Information Criterion (AIC) and Schwarz Bayseian Information (SBC). Using Box-Cox transformation and subsequently differencing, monthly, weekly and daily monsoon rainfall time series were made stationary. Seasonal Auto Regressive Integrative Moving Average models (SARIMA) were developed for monthly, weekly and daily monsoon rainfall time series.
0 kommentar(er)
