- 期刊
以模糊類神經綱路設定電子拍賣市集競標參數
A fuzzy-neural-network Model for Bidding Parameters of auction-based e-market
摘要
隨著電子拍賣市集的蓬勃發展，網路購物已成爲現代人的另一種生活方式。在廣大的電子拍賣市集中，專家在對物品進行拍賣時，最爲關心的，莫過於競標參數之設定，如投標期間、起標價等，因它涉及專家是否能夠獲利的關鍵因素。本研究藉由倒傳遞類神經網路及模糊類神經網路建構一套對競標結果有精凖預測之網路模型，讓專家在拍賣物品之前可先行評估其獲利。此外，尚可根據已建構出之網路模型建立出知識庫，供賣家在決定競標參數設定時的參考。本研究以eBay拍賣網站爲例，挑選出兩種競標之參數作爲主要之研究對象，一個屬於數值型（結標金額），一個屬於類別型（投標期間），建立預測模式。本研究發展之預測可擴展到一般之電子商務市場。
並列摘要
Due to the development of e-market, online shopping has become a king of living things for people. In the vast auction-based e-market, sellers take care of designing of proper auction-based parameters as selling things; for example, the period for bidding, starting bid and so on, so that profit can be maximized. In the study, we use back-propagation neural network and fuzzy-neural-network respectively to build forecast models resulting in an accurate auction design. We built two forecast models for auction parameters. One is a numerical type (the deal price), and the other a category type (the duration for bidding). The models estimate the profits and suggest a best bid duration before selling things on-line, Beside, by using the knowledge base built by the proposed network model, one can decide when to adjust the auction-based parameters. The cellular phone trading of eBay.com is illustrated as an example in the study.