| Summary: | 碩士 === 實踐大學 === 資訊科技與管理學系碩士班 === 106 === The Ministry of the Interior (MOI) of the Republic of China (Taiwan) has promoted the actual price registration policy since 2012. There have been a large number of estate transactions aggregated and published on the actual price registration platform since 2012 in order to achieve information transparency in real estate transactions and reduce information asymmetry. People who is going to buy or sell the house can acquire the information on real estate transactions from the actual price registration platform to estimate a reasonable range of housing price.
In the past studies, many scholars established housing price estimation models by statistical or mathematical methods to avoid subjectively estimating housing prices and reduce the cost of individual housing prices estimation. We can establish an effective housing price estimation model by learning and testing. A reference value of the housing price can be estimated through the housing price model.
The mortgage burden rate and the house price-to-income ratio in Taipei are the highest in Taiwan, so it’s very difficult for people to buy a house in Taipei. In this study, the housing prices in Taipei will be analyzed and estimated. Real estate transactions data were collected from 2012 to 2017 on the actual price registration platform in MOI. First, special status transactions data were excluded from the dataset. And surrounding housing living variable data were added to the dataset. Then, the back-propagation neural network (BPNN) was used to develop the housing price estimation model. Different combination of input variable and neural network structures are tested to achieve the best estimation model. The housing price estimation model developed in this research provides a good reference price for people who want to buy a house.
|
|---|
