估價於房地產市場實扮演著一不可或缺的角色，精確的估價不僅可提供消費者正確極充分的購屋資訊，亦為政府擬定政策方針之基礎。由於台灣房地產市場為一不完全市場，消費者在購屋的同時更常因資訊的不健全而遭受不必要之損失，因此精確及流通之估價資訊實為健全台灣房地產市場之首務。 鑑於過去估價技術仍未成熟，所佔之房價常無法令人信服。本研究欲以類神經網路之功能，將其原理應用於房地產估價上，試圖解決過去估價方法本身之缺失，並作為估價人員輔助之工具。本研究主要以倒傳遞及理解倒傳遞類神經網路與特徵價格法進行公證比較分析，並以特徵價理論為基礎，利用類神經網路得出影率房地產價格更具代表性之因素，以做為未來建立房地產估價輔助系統之參考。 為了解不同的資料型態是否會使類神經網路有不同的學習效果，本研究將資料分為四組實驗設計，分別對不同的資料型態進行測試，研究結果顯示類神經網路對於資料型態較為敏感，其中又以理解倒傳遞類神經網路為最，使得其在預測能力上易受異常點或極端值的影響，而有好壞差異較大的情況。即類神經網路之學習效果端視資料是否具代表性而定。

The objective of the thesis is to determine the economic rent amount for a family house located in Brno. The concept “economic rent” refers to the amount of payment that brings adequate profit to the landlord. First, all the factors are identified that compose cost rent, i.e. the rent of mere providing and maintaining a property. The main variable is represented by the reproduction value of a dwelling. For the purposes of this thesis, the reproduction value is determined using the cost method (in agreement with the evaluation regulation). Second, the capitalization rate is determined so that an adequate profit could be calculated. The following two methods are employed: the surcharge method and the method of comparison with the financial market. The thesis presents seven alternatives of an economic rent, depending on the interest rates. Lastly, the calculation of the rate of return is performed, along with a correctness test of the calculation of the economic rent amount.