<![CDATA[Various fuzzy and non-fuzzy models have been presented to forecast the stock market with multiple inputs data or variables. In other words, some of the researchers have overlooked the key success in financial time series forecasting which is minimizing number of inputs. Moreover, most of the existing time series models have been focused on data consisting of single values, or fuzzy numbers without randomness into consideration. In real situations, there exists a genuine need to cope with data that involves the factors of fuzziness and probability. To address the drawbacks, we propose an enhanced fuzzy random auto-regression model for better stock market forecasting using the low-high procedure. This procedure is able to represent the daily prices variations in stocks. The daily stock markets of Indonesia-Malaysia are used as numerical examples and efficiency of the proposed procedure is compared with baselines models.]]>
Copyrights © 2017
