Building a multi-output hybrid model for interval-valued time series forecasting
Abstract
Time series analysis and forecasting is an attractive research area over the last few years, especialy in interval-valued time series. Financial market is an example; a useful model may be of great interest to home brokers who do not possess sufficient knowledge to invest in such companies. In this paper, multi-input multi-output least square support vector regression (MIMO-LSSVR) is an improved algorithm based on support vector machine (SVM), with the combination of the sliding-window algorithm is proposed for interval-valued time series forecasting, a new branch in time series analysis field. The experiment shows MIMO-S-LSSVR positive outcomes than previous results. A retest using twotime series data sets in three years demonstrates that the proposed model is a promising alternative for interval-valued time series forecasting.
Moraref