摘要
近年来,如何从个人角度分析竞买者出价行为成为研究难点。本文运用基于函数性稀疏数据聚类方法从竞买者个人角度定义了基于条件期望的距离矩阵,并结合多维尺度分析方法(MDS)和相平面图分析法对竞买者出价行为特征进行聚类分析。结果表明,按出价时间、金额、出价能量以及获胜概率可将竞买者分为显著的四类。其中,经验丰富型竞买者获胜概率最高,势在必得型竞买者获胜意愿最强烈,缺乏经验型竞买者退出竞拍时间较早导致获胜概率较低,围观型竞买者由于竞拍意愿不强烈所以获胜概率最低。
In recent years,how to analyze bidding behavior from the perspective of single bidder has always attracted many attentions.By using of the clustering methods of sparsely observed functional data,we define the distance matrix and carry out a cluster analysis to bidders' bidding trajectories with the help of Multidimensional Scaling(MDS)and phase planes.The results indicate that there are four distinct clusters of Chinese on-line bidders' bidding patterns,namely Experienced Bidder,Green Hand,Strong Minded Bidder and Participants.They vary in bidding time,bidding amount and winning probability.
出处
《当代经济科学》
CSSCI
北大核心
2017年第6期115-121,共7页
Modern Economic Science
关键词
稀疏数据
函数性聚类分析
出价行为特征
网上拍卖
Sparse data
Clustering method on functional data
Bidding behavioral characteristics
Online auction
