基于影响域的新型众包定价算法模型构建
2018-08-15周春樵肖昌昊刘扬姚安琪黄君扬
周春樵 肖昌昊 刘扬 姚安琪 黄君扬
摘 要:“众包”已成为时下新兴的一种基于互联网进行信息检查和搜集的商业模式,其成功率取决于诸多因素的影响,其中最大的影响因素为任务发布者的出价。针对此问题,本文提出了一种基于“影响域”的新型众包定价策略,该策略以经济学中的供求关系模型为建模方法,利用任务与劳动者的地理位置分布规律动态定价,同时,对新数据与原始数据进行相似性分析,通过机器学习模拟任务的完成概率,从而评价定价策略的优劣。本文以“拍照赚钱”自助式服务模式作为研究样本;在利用影响域定价模型重新定价后,经济效用较原始方案增长80.65%,效果良好。
关键词:众包;影响域;供求关系;标准化欧氏距离;机器学习
中图分类号:TP301 文献标识码:A
Abstract:Crowdsourcing has become a new business model nowadays.It not only makes human knowledge and wisdom improved and disseminated infinitely,but also creates amazing social wealth.However,the success rate of crowdsourcing missions depends on a number of factors,among which the most important one is the bid given by mission publishers.In this paper,a new pricing strategy based on domain-of-influence is proposed,which uses the geographical distribution of missions and the employees to price dynamically,then readjusts the size of the affected domain for an iterative calculation until the pricing result is stable.In addition,this paper establishes a mathematical model to simulate the probability of completion of a mission,which is used to test the merits of the pricing strategy based on domain-of-influence.This paper takes the self-service model of Photographing for Money as the study subject and the financial rewards have increased by 80.65% after repricing by means of domain-of-influence compared with the previous pricing method.
Keywords:crowdsourcing;domain-of-influence;supply-demand relationship;standardized Euclidean distance;machine
learning
1 引言(Introduction)
2006年6月份的《连线》杂志中,记者Jeff Howe[1]在《众包的崛起》一文中首次提出了“众包”的概念。众包,指公司或机构把工作任务通过网络外包给非特定的大众,是“网络大众”与“外包”的合成词汇。它弥补了公司或机构自身的资源缺陷问题,提高了企业的工作效率,同时给完成任务的网络大众带来一定的经济收入,可谓双赢的经营模式。时下,已有学者对众包做出了详尽的经济学解析[2-4],也对任务定价的影响因素做了具体研究[5]。众包任务的定价策略是该模式中举足轻重的部分——过低的定价可能会导致任务无法顺利完成;过高的定价则为企业带来沉重的负担。本文提出的基于影响域的众包定价策略,通过在每个任务点的周围划定了“影响域”,统计该任务点附近的任务密度和劳动力密度,根据供求关系模型[6,7]确定该任务点的定价。……
