The Application of Adaptive Network-based Fuzzy Inference System and Grey Relational Analysis for Taiwanese Government Bonds Yield Prediction

• Main Authors: Kun-Lung Hsieh, 謝坤龍
• Other Authors: Wei-Yuan Lin
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/80878986695980404301

碩士 === 東吳大學 === 經濟學系 === 94 === The stock market and bond market are the major components of domestic financial market. The bond market has been getting more attention because of its expansion for years. Interest rate is the main variable of the operation of economic system according to economic theory. Its variation demonstrates the various development of economic and influences economic activities. The aggregate behavior of whole society can be observed through economic statistics and indices basing on the concepts of macroeconomics. Most participants of financial market also make anticipation by their changes . Therefore, this paper believes they can demonstrate the variation of government bond yield. This paper forecasts the trend of government bond yield by Multiple Regression, Grey Prediction, and Adaptive Network-based Fuzzy Inference System (ANFIS). There are lots of domestic literatures studying how to forecast the trend of government bond yield using the Neural Networks, but rare literatures adopt ANFIS. ANFIS has the advantage of both Fuzzy Logic and Artificial Neural Network. Fuzzy Logic works via the corresponding relationship of input variables and output variables. Artificial Neural Network finds the best prediction model through the training and learning of history. This paper gets the anticipation using four points rolling model and huge data which are processed by AGO adopting traditional GM(1,1) model in Grey Prediction. This paper evaluates the prediction values with RMSE and Wilcoxon Sign Test. Grey Prediction performs best in prediction according to the results.