High-order fuzzy time series forecasting model for advance prediction of temperature

Forecasting future outcomes has always been a subject of keen interest in real world. To construct such a model, fuzzy time series provide tremendous flexibility to choose the probabilistic measures of input parameters. A new high-order time series model based on concept of fuzzy set is presented in this paper. With the help of a novel defuzzification method, temperature values are predicted. Experiments have been performed using temperature time series data set of Taipei. Result of proposed model is compared with contemporary advanced forecasting model to predict temperature. Statistical metrics Mean, Standard Deviation, Root Mean Square Error (RMSE) are used to evaluate the performance of the model. Average RMSE of proposed model is 1.21 in contrast to 1.24 (model based upon Artificial Neural Network) and 1.91 (Multivariate Markov Chain model). Empirical results show that the proposed method is well suitable and efficient for forecasting compared to other existing advanced techniques.