Loading...
|
Please use this identifier to cite or link to this item:
https://nccur.lib.nccu.edu.tw/handle/140.119/157609
|
| Title: | 從大數據預測不動產市場
Predicting the Real Estate Markety through the Big Data |
| Authors: | 林左裕 |
| Contributors: | 地政系 |
| Date: | 2017-09 |
| Issue Date: | 2025-06-30 13:54:58 (UTC+8) |
| Abstract: | 以往不動產市場相關預測研究多藉官方或私部門整理過之制式資料，如經濟成長率、股價指 數、通貨膨脹率及利率等傳統總體經濟或金融指標，應用這些指標的各種預測模型視情況有不同的 預測能力，但共通的現象是，這些指標是經過一段期間後(如每季或每月）經整理後再發布的數據，
因此即使計量預測模型完美，但仍無法達到即時預測之目的，這也是傳統經濟預測模型的限制，在 實務應用有即時需求的業者或是政策而言，有時即有緩不濟急之感。
近年來網路使用頻繁，使用者常藉搜尋引擎主動搜尋所需之資訊，這些未經整理的搜尋紀錄，
在不動產市場上常見的字串，如不動產仲介（或經紀）、買屋、賣屋及銀行貸款等，即透漏出搜尋者 之需求，因此若能應用這些未經機構整理的”大數據”進行不動產市場之預測，且能得到有效之結 果，未來的預測模型將可能有巨幅之轉變或預測效力之提升，也可提供給具需求之交易者及政策即 時且有效之參考。
本研究擬以搜尋引擎所記錄之相關字串預估住宅市場之供需、價格及交易量，研究方法實證方法 擬依資料取得之方便性採複迴歸或(及)時間序列模型進行探討，除了可藉供需間之關係預估住宅價格、 提高對於住宅價格之預估能力外，更期望能藉搜尋引擎探究對於住宅貸款違約率之預估，未來政府可 檢測搜尋引擎指數以監測不動產金融市場，提高金融市場之穩定性，並提供政府制定政策時之參考。
Traditional forecasting models on the real estate market mostly employ macro-economic indicators, such as economic growth rate, inflation rate, interest rate or stock price, for empirical analysis. Different models usually have different levels of forecasting validity, depending on the variables or sample period. The common feature of these traditional models is that these variables are collected and announced by government or private institutions. These variables usually are released quarterly or monthly. Therefore, although some models are plausible, the lagged variables are the major limitations for the application of these models.
The usage of internet has become increasingly popular in recent years. Users are getting used to search what they need through the searching engine. These searching records, or “big data”, may provide valid implication for the real estate market, e.g., buy/sell houses, mortgages, brokerage, ... , and so on. If these searching keywords can be collected for analysis, researchers may be able to forecast real estate market immediately with higher validity. The results may provide immediate implications to market traders or policy makers.
This study intends to collect the searching records in the searching engine for empirical analysis. Research method will be either the multiple regression or time series analysis, depending on the availability of the data. It is expected that the forecasting model through big data may enhance the predicting power for the traditional approach. Further, should the mortgage default behavior can also be predicted, government may be able to launch policies in time to prevent from the occurrence of financial crisis. |
| Relation: | 科技部, MOST105-2410-H004-168, 105.08-106.07 |
| Data Type: | report |
| Appears in Collections: | [地政學系] 國科會研究計畫
|
Files in This Item:
| File |
Description |
Size | Format | |
|---|
| index.html | | 0Kb | HTML | 1 | View/Open |
|
All items in 政大典藏 are protected by copyright, with all rights reserved.
|
