基於模糊神經網路之多變量高階加權模糊時間序列模式

Multivariate High-Order Weighted Fuzzy Time Series Based on Fuzzy Neural Networks

黃煜傑(Yu-Jie Huang)

指導教授 ： 張景榮

朝陽科技大學/資訊學院/資訊管理系/碩士(2012年)

摘要

人類社會中存在許多不確定性的問題，如經濟成長率、財務危機的預測等，自從Song 與Chissom 兩位學者在1993 年提出所謂的模糊時間序列(Fuzzy Time Series)的觀念後，許多學者先後提出不同的模式來處理這些問題，然而先前的研究通常未考慮相關變數選取以及在模糊化過程中僅依據主觀意見進行模糊語意的離散化，因此不能夠客觀地反應資料集的特性，此外在進行預測時往往將模糊規則視為同等重要，未能考慮每一條模糊規則的重要性。基於上述原因，本研究透過自組織映射圖網路(Self-Organizing Map, SOM)進行變數選取(Factor Selection)並提出多變量高階加權模糊時間序列模式基於模糊神經網路(Fuzzy-BPN)，且採用循序加權平均運算子(Ordered Weighted Averaging operator, OWA)進行加權預測。因此，為了驗證提出的方法，採用台灣證券交易所(Taiwan Stock Exchange Corporation)提供的台灣股價加權指數(Taiwan Stock Exchange Capitalization Weighted Stock Index, TAIEX)為實驗預測目標，以此篩選適當相關變數，在實驗最後納入近年其他的研究模式與本研究一併比較，結果顯示本研究之預測能力有進一步的改善。

關鍵字

循序加權平均運算子 ； 倒傳遞神經網路 ； 自組織映射圖網路 ； 模糊時間序列

並列摘要

There are many uncertainty problems in the Human society, such as the forecasting of economic growth rate, financial crisis, etc. Since Song and Chissom proposed the concept of fuzzy time series in 1993, many scholars have proposed different models to deal with these problems. However, previous studies usually do not consider the factor selection and transfer original data to the fuzzy linguistic value by the subjective opinions in fuzzy process, which cannot objectively show the characteristics of the data. In addition, the fuzzy rules usually assign equal weight in the forecasting process, and it failed to consider the importance of each fuzzy rule. Based on above concepts, this study adopt the self-organizing map network (SOM) for the purpose of factor selection and proposed a multivariate high-order weighted fuzzy time series model based on fuzzy neural network (Fuzzy-BPN) and ordered weighted averaging operator (OWA) to make forecasts. In order to verify the proposed method, the Taiwan Stock Exchange Capitalization Weighted Stock Index (TAIEX) from the Taiwan Stock Exchange Corporation are used in the experiment and filter the appropriate factors, and the experiment results are compared with other methods in with this study. The forecasting performance shows that the proposed method has better forecasting ability.

並列關鍵字

ordered weighted averaging operator ； back-propagation neural network ； self-organizing map network ； fuzzy time series

參考文獻

[1] 林延臻(2008)，隸屬度函數及區間長度改良對模糊時間序列預測之影響探討，碩士論文，大同大學應用數學研究所，台北。

[13] Chang, J. R., Wei, L. Y., and Cheng, C. H. (2011), “A Hybrid ANFIS model based on AR and volatility for TAIEX forecasting,” Applied Soft Computing, Vol. 11, No. 1, pp.1388-1395.

[14] Chang, J. R., Lee, Y. T., Liao, S. Y., and Cheng, C. H. (2007), “Cardinality-based Fuzzy Time Series for Forecasting Enrollments,” (IEA/AIE 2007, Japan)H.G. Okuno and M. Ali(Eds.), Lecture Notes in Artificial Intelligence, Vol. 4570, pp. 735-744.

[15] Chen, S. M. (1996), “Forecasting enrollments based on fuzzy time series,” Fuzzy sets and systems, Vol. 81, No. 3, pp. 311-319.

[17] Chen, S. M., and Tanuwijaya, K. (2011), “Multivariate fuzzy forecasting based on fuzzy time series and automatic clustering techniques,” Expert Systems with Applications, Vol. 38, No. 8, pp. 10594-10605.