A support vector machine based method for predicting final prices of online auctions

作者：Liu Yang^*; Feng Yu Qiang; Shao Zhen

来源：Journal of Harbin Engineering University, 2010, 31(3): 400-404.

DOI：10.3969/j.issn.1006-7043.2010.03.024

• Calculation time
• Experimental data
• Online auction
• Online auctions
• Price prediction
• Real-time decision support systems
• Training time

A new method was proposed for predicting the final price in online auctions. By analysing bidders' behavior, the problem of long calculation time was solved. Instead of predicting the final price directly, the method used a support vector machine to analyse the time of each bid and then used that data to calculate an end price. The authors collected data on 3310 transactions and their corresponding 8275 bids from Taobao and used them as experimental data. The experiment proved that this model substantially outperforms the naive method of predicting by using the mean category price. The training time was only a few seconds, yet 22.1% of the predicted results were identical to the real ones. This research can provide a foundation for the development of real-time decision support systems.

• 出版日期2010