在證券市場預測的應用上,以往關於股價的趨勢預測研究,大多是利用時間數列的分析方法,如ARIMA當作預測的主要方法。但是,由於證券市場充斥著許多不明確性,且所獲得的資料有曖昧性之可能,因此本文提供新發展獨特的可能性灰色預測模型與模糊迴歸兩種預測方法,作為投資者在進行預測時的輔助決策。本研究除針對新創的可能性灰預測做說明外,亦希望對於因果分析法與時間數列法分析其中的優劣,藉以為投資者找出一個在目前狀況下的最佳投資預測模型。本研究之主要目的除比較分析可能性灰色預測與模糊迴歸兩種預測方法之特色外,尚特別針對研究中所必須取決的樣本數進行相關探討,並且以台灣股市加權指數作為本研究之實例驗證。在實例驗證所顯示的結果中,發現可能性灰色預測與模糊迴歸兩種方法在預測能力上均屬佳,只是在運用的範圍上略有不同;至於在樣本數上,依本例結果發現,十個樣本數較能兼具六個與二十個樣本之優點,達到有效又即時之最佳預測結果。
Causality and time series model are the most effective methods adopted in forecasting practices. Time series model, such as ARIMA, is applied for most researchers in stock prices prediction. However, the financial environment and information around the stock market are mostly vague. It is therefore, this paper is to present two forecasting methods: Possibility grey forecasting and fuzzy regression. These two models help investors make decisions in stock market. For promoting the performance, the differences between these models and the scenarios of implementing are also analyzed in this paper. These will assist investors to formulate their investing strategy while facing various conditions. The main purpose of this paper is to analyze and compare the different characteristics between possibility grey forecasting model and fuzzy regression, especially about the numbers of samples in possibility grey forecasting model. The results have showed that the possibility grey forecasting model and fuzzy regression's ability of forecasting are proper, but in different application. In possibility grey forecasting model, the ten samples outperform the six and twenty ones. Six and twenty samples can play good roles in forecasting, but ten samples have better performance.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。