Fourth-order Fuzzy Time Series Based on Multi-Period Adaptation Models for Kuala Lumpur Composite Index Forecasting

the length o ti time series with adaption period was proposed by Chen et al as one of the mechanisms in improving forecasting accuracy. First, second, third and fourth order fuzzy times series based on multi-period adaptation models were successfully tested with stock exchange indexes in Taiwan and Hong Kong. However feasibility of the fourth order model especially the effect of multi-period adaptations to local indexes remains unknown. Therefore the present paper tests the fourth-order based on one, two, three and four-period adaptation model with datasets of Kuala Lumpur Composite Index. With the fourth-order of the adaptation model, it is found that the forecasted index based on two-period adaptation performed better than the other adaptation periods. The fourth-order based on two-period adaption seems perfectly worked with datasets of Kuala Lumpur Composite Index.