本研究係探討影響中古屋成交價格的變動因素。本文中的中古屋價格以高雄市區九個行政區域爲主、由房仲業者自2007年至2008年所提供的實際房屋成交價格資料。房屋成交價格資料爲依變數，建物總坪數（坪）、屋齡（年）、車位、房屋類型與行政區域爲自變數。利用自動互動檢視法(CHAID)的分析工具，逐次搜索各種元素。使降低模型中因子之水準數目組合，並適切掌握變數間的相關與互動模型，即時有效回應、母體現況的結構。與住宅供需相關議題之研究均以總體經濟因素與房價之探討爲主，但由相關研究可知總體經濟變數對高雄市房價的解釋能力皆不高。本研究乃以高雄都會區房仲業者所提供之實際交易價格爲基礎。以「CHAID」呈現現況之房屋成交價格的架構，利用交叉分析探討影響因素；更進一步利用「對應分析」指出兩成對變數的各類別問之對應關係，且由對數線性模型探測主要影響交易次數變動的因素。將這些因素組合系統化，更能呈現影響成交價格之因素其所提供的資訊。

This research is to study the influential factors of the housing sold prices. The house sold prices are provided by the real estate agents of houses sold from year 2007 to 2008 and primarily based on the 9 administrative areas in Kaohsiung city. The house sold prices are dependence variation; and the home size, the house age, the garage, the house types and the area are self-variation. By using CHAID as the analysis tool to inspect each factor, it can reduce figure association of parameter in models and well manage the relevance of the variations, and effectively reflect current main structure. This research takes the sold prices as the basis. By using decision tree to show the structure of current housing prices; and using alternate analysis to analyze the influence; and further using corresponding analysis as to get the correspondence. Moreover, it takes loglinear model as the main influential factor. Combining all these factors systematically, it will reflect the data of these influential factors of the house sold prices.