| Summary: | 碩士 === 東吳大學 === 經濟學系 === 101 === To live in peace and enjoy one’s work is a common goal of our life. However, house prices fluctuate naturally affected by many external factors with the limited supply of land resources. When irrational prices change, as previously the U.S. subprime mortgage crisis or the real estate bubble worldwide, the impact of the events is so huge and accidental. Therefore to understand the future trend of prices has become the continuing concern issues for governmental policies and people's life.
From the existing research literature study shows that the prediction model using artificial intelligence, compared with the traditional method alone, can be easier achieved better prediction results. At the same time, the paper also introduces a new evolutionary computation method – fruit-fly optimization algorithm (FOA). We hope that hybridize grey neural network and generalized regression neural network with FOA could construct effective prediction models.
The empirical results show that Grey Prediction and Grey Model Neural Network forecasts in a small number of samples data have better predictive abilities in the used-residential house prices of Taipei City. Through FOA to fine tune the parameters of grey neural network or generalized regression neural network, the hybrid model can significantly to improve the performance, i.e. the accuracy of prediction.
Five variables are selected from stepwise regression. They are one-year floating rate of post office savings, M2, consumer price index, domiciled households and stock index. Through these variables training, our models in this paper can get very well predictive value. It also shows that housing prices change is dominated by the overall impact of macroeconomic indicators. To assess the recent easing of monetary policy and low interest rate, coupled with inflation concerns, produce more big impact to the transaction price of housing within the region in recent years.
|
|---|
