隨著資訊技術以及資料庫的迅速發展,許多大型企業累積越來越 多的客戶輪廓、消費行為以及交易紀錄等不同的資料,因此從龐大資料庫中尋找潛藏商業價值的資料採礦技術,也越來越受這些企業的重視,同時也促使資料採礦的相關研究逐漸興盛。資料採礦意指尋找隱藏在大量資料中的訊息,如趨勢 (Trend)、特徵 (Pattern)及相關性等的一個過程;也就是從資料中發掘有價值的資訊或知識的過程。 房屋仲介業的主要經營模式,在於提供一個房屋買賣的交易平台,並透過促成交易以收取服務費用。因而在整個服務過程中,如何在最短的時間內,將合適的物件資訊推薦給需要的客戶,以促成交易的圓滿成交,自然成為房仲業者營運上的關鍵課題;不僅是爭取顧客滿意的利器,同時也是同業競爭的關鍵。因此本研究以台灣地區某房仲業的顧客資料庫為研究對象,應用關聯法則(Association rule)中的Apriori演算法,建立購屋者與購買房屋之間的關聯法則,藉此找出購屋客戶特性與購買房屋之間的關係,以利往後建構一個完整的房屋銷售推薦模型;並針對模型結果進行管理分析,例如研究發現當買方客戶現居地區在中和市,且是透過行銷募集管道而來時,有12.50%機率會購買中和市,大小在21-30坪且總價介於601-700萬間的房屋,最後依據分析結果提出相關的管理意涵,以為公司經營者在決策制定上的參考。
Many large enterprises nowadays store a lot of data of customer profile, consumer behavior and transaction records since the information technology and database have developed rapidly. Data mining techniques that look for information with business potential are thus getting more importance in those enterprises. Research on data-mining also flourishes. Data mining refers to the process in which valuable information or knowledge such as trend, pattern and relevance is discovered from large data sets. The major business model of (current) real estate industry is to provide a platform that mediates buyers and sellers of properties and to profit from the service charges upon successful mediation. As a result, promoting a higher rate of successful mediations, that is, matching the right properties to the right clients within the shortest time possible becomes essentially critical. It is the key not only to a higher customer satisfaction but also to outperform competitors. This research thus aims to employ data mining techniques to construct a recommendation model that predicts potential properties transaction of a house selling through the information of customers’ characteristics. This model is built upon the customer database of a specific real estate agency in Taiwan by applying the Apriori algorithm in Association Rule. Predictions made from this model provide managerial implications. For example, it is found that if a house buyer residing in Jhonghe City contacts an agency from its marketing activities, there is a 12.5% probability that the customer will buy a house located in Jhonghe City and costs six to seven million NT dollars with a size of 21 to 30 pings. This information will improve the decisions making of managers in real estate industry.